From ad82bb2c48564a6f23e005e01c89e4c12318510c Mon Sep 17 00:00:00 2001
From: peng_windows <2686728826@qq.com>
Date: Tue, 26 Mar 2024 12:26:17 +0800
Subject: [PATCH 1/6] llm-introduction-attention-transformer
---
.../_config.yml | 1 +
open-machine-learning-jupyter-book/_toc.yml | 9 +
.../llm/basic/transformer-architecture.ipynb | 730 +
.../llm/basic/attention.ipynb | 367 +
.../llm/basic/basic.ipynb | 64 +
.../llm/basic/transformer.ipynb | 20020 ++++++++++++++++
.../llm/image/attention_example.svg | 9628 ++++++++
.../llm/image/cifar100_example_anomaly.png | Bin 0 -> 155392 bytes
.../llm/image/comparison_conv_rnn.svg | 1809 ++
.../llm/image/implicit-order.png | Bin 0 -> 34807 bytes
.../llm/image/llm.png | Bin 0 -> 166613 bytes
.../llm/image/multihead_attention.svg | 288 +
.../llm/image/scaled_dot_product_attn.svg | 351 +
.../llm/image/scaling-laws.png | Bin 0 -> 86847 bytes
.../llm/image/transformer_architecture.svg | 118 +
.../llm/image/warmup_loss_plot.svg | 1579 ++
.../llm/introduction.ipynb | 155 +
17 files changed, 35119 insertions(+)
create mode 100644 open-machine-learning-jupyter-book/assignments/llm/basic/transformer-architecture.ipynb
create mode 100644 open-machine-learning-jupyter-book/llm/basic/attention.ipynb
create mode 100644 open-machine-learning-jupyter-book/llm/basic/basic.ipynb
create mode 100644 open-machine-learning-jupyter-book/llm/basic/transformer.ipynb
create mode 100644 open-machine-learning-jupyter-book/llm/image/attention_example.svg
create mode 100644 open-machine-learning-jupyter-book/llm/image/cifar100_example_anomaly.png
create mode 100644 open-machine-learning-jupyter-book/llm/image/comparison_conv_rnn.svg
create mode 100644 open-machine-learning-jupyter-book/llm/image/implicit-order.png
create mode 100644 open-machine-learning-jupyter-book/llm/image/llm.png
create mode 100644 open-machine-learning-jupyter-book/llm/image/multihead_attention.svg
create mode 100644 open-machine-learning-jupyter-book/llm/image/scaled_dot_product_attn.svg
create mode 100644 open-machine-learning-jupyter-book/llm/image/scaling-laws.png
create mode 100644 open-machine-learning-jupyter-book/llm/image/transformer_architecture.svg
create mode 100644 open-machine-learning-jupyter-book/llm/image/warmup_loss_plot.svg
create mode 100644 open-machine-learning-jupyter-book/llm/introduction.ipynb
diff --git a/open-machine-learning-jupyter-book/_config.yml b/open-machine-learning-jupyter-book/_config.yml
index e7464e4cd..a0f9b925d 100644
--- a/open-machine-learning-jupyter-book/_config.yml
+++ b/open-machine-learning-jupyter-book/_config.yml
@@ -23,6 +23,7 @@ execute:
- 'ml-advanced/unsupervised-learning-pca-and-clustering.ipynb'
- 'ml-advanced/unsupervised-learning.ipynb'
- 'data-science/data-science-in-the-cloud/the-azure-ml-sdk-way.ipynb'
+ - 'llm/basic/transformer.ipynb'
parse:
myst_enable_extensions:
diff --git a/open-machine-learning-jupyter-book/_toc.yml b/open-machine-learning-jupyter-book/_toc.yml
index a09d32cd1..5c9b8c367 100644
--- a/open-machine-learning-jupyter-book/_toc.yml
+++ b/open-machine-learning-jupyter-book/_toc.yml
@@ -122,6 +122,14 @@ parts:
- file: machine-learning-productionization/data-engineering
- file: machine-learning-productionization/model-training-and-evaluation
- file: machine-learning-productionization/model-deployment
+- caption: Large Language Models
+ numbered: True
+ chapters:
+ - file: llm/introduction
+ - file: llm/basic/basic
+ sections:
+ - file: llm/basic/attention
+ - file: llm/basic/transformer
- caption: OTHERS
numbered: True
maxdepth: 1
@@ -237,6 +245,7 @@ parts:
- file: assignments/deep-learning/nlp/getting-start-nlp-with-classification-task
- file: assignments/deep-learning/nlp/beginner-guide-to-text-preprocessing
- file: assignments/deep-learning/nlp/news-topic-classification-tasks
+ - file: assignments/llm/basic/transformer-architecture
- file: slides/introduction
sections:
- file: slides/python-programming/python-programming-introduction
diff --git a/open-machine-learning-jupyter-book/assignments/llm/basic/transformer-architecture.ipynb b/open-machine-learning-jupyter-book/assignments/llm/basic/transformer-architecture.ipynb
new file mode 100644
index 000000000..274f60583
--- /dev/null
+++ b/open-machine-learning-jupyter-book/assignments/llm/basic/transformer-architecture.ipynb
@@ -0,0 +1,730 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "# Complete the transformer architecture"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "# set up the env\n",
+ "\n",
+ "import pytest\n",
+ "import ipytest\n",
+ "import unittest\n",
+ "\n",
+ "ipytest.autoconfig()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Transformer Model\n",
+ "\n",
+ "The encoder-decoder architecture based on the Transformer structure is illustrated in figure below. The left and right sides correspond to the encoder and decoder structures, respectively. They consist of several basic Transformer blocks (represented by the gray boxes in the figure), stacked N times. Each component comprises multiple Transformer blocks, which are stacked N times.\n",
+ "\n",
+ "Here's an overview of the key components and processes involved in the semantic abstraction process from input to output:\n",
+ "\n",
+ "Encoder:\n",
+ "\n",
+ "The encoder takes an input sequence {xi}ti=1, where each xi represents the representation of a word in the text sequence.\n",
+ "It consists of stacked Transformer blocks. Each block includes:\n",
+ "Attention Layer: Utilizes multi-head attention mechanisms to capture dependencies between words in the input sequence, facilitating the modeling of long-range dependencies without traditional recurrent structures.\n",
+ "Position-wise Feedforward Layer: Applies complex transformations to the representations of each word in the input sequence.\n",
+ "Residual Connections: Directly connect the input and output of the attention and feedforward layers, aiding in efficient information flow and model optimization.\n",
+ "Layer Normalization: Normalizes the output representations of the attention and feedforward layers, stabilizing optimization.\n",
+ "Decoder:\n",
+ "\n",
+ "The decoder generates an output sequence {yi}ti=1 based on the representations learned by the encoder.\n",
+ "Similar to the encoder, it consists of stacked Transformer blocks, each including the same components as described above.\n",
+ "In addition, the decoder includes an additional attention mechanism that focuses on the encoder's output to incorporate context information during sequence generation.\n",
+ "Overall, the encoder-decoder architecture based on the Transformer structure allows for effective semantic abstraction by leveraging attention mechanisms, position-wise feedforward layers, residual connections, and layer normalization. This architecture enables the model to capture complex dependencies between words in the input sequence and generate meaningful outputs for various sequence-to-sequence tasks.\n",
+ "\n",
+ ":::{figure} https://media.geeksforgeeks.org/wp-content/uploads/20230531140926/Transformer-python-(1).png\n",
+ "---\n",
+ "\n",
+ "width: 90%\n",
+ "---\n",
+ "Transformer-based encoder and decoder Architecture\n",
+ ":::\n",
+ "\n",
+ "Next, we'll discuss the specific functionalities and implementation methods of each module in detail."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Embedding Layer\n",
+ "\n",
+ "The Embedding Layer in the Transformer model is responsible for converting discrete token indices into continuous vector representations. Each token index is mapped to a high-dimensional vector, which is learned during the training process. These embeddings capture semantic and syntactic information about the tokens.\n",
+ "\n",
+ "Implementation in PyTorch:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import torch\n",
+ "import torch.nn as nn\n",
+ "import math\n",
+ "\n",
+ "class PositionalEncoder(nn.Module):\n",
+ " def __init__(self, d_model, max_seq_len=80):\n",
+ " super().__init__()\n",
+ " self.d_model = d_model\n",
+ " ## Create a constant PE matrix based on pos and i\n",
+ " pe = torch.zeros(max_seq_len, d_model)\n",
+ " for pos in range(max_seq_len):\n",
+ " for i in range(0, d_model, 2):\n",
+ " pe[pos, i] = math.sin(pos / (10000 ** ((2 * i) / d_model)))\n",
+ " pe[pos, i + 1] = math.cos(pos / (10000 ** ((2 * (i + 1)) / d_model)))\n",
+ " pe = pe.unsqueeze(0)\n",
+ " self.register_buffer('pe', pe)\n",
+ "\n",
+ " def forward(self, x):\n",
+ " ## Scale word embedding representations\n",
+ " x = x * math.sqrt(self.d_model)\n",
+ " ## Add positional constants to word embedding representations\n",
+ " seq_len = x.size(1)\n",
+ " pe = torch.autograd.Variable(self.pe[:, :seq_len], requires_grad=False).cuda()\n",
+ " x = x + pe\n",
+ " return x"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "
Check result by executing below... 📝 "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "jupyter": {
+ "source_hidden": true
+ },
+ "tags": [
+ "hide-input"
+ ]
+ },
+ "outputs": [],
+ "source": [
+ "%%ipytest -qq\n",
+ "\n",
+ "class TestPositionalEncoder(unittest.TestCase):\n",
+ " def setUp(self):\n",
+ " self.d_model = 512\n",
+ " self.max_seq_len = 10 # Maximum sequence length for testing\n",
+ " self.positional_encoder = PositionalEncoder(self.d_model, self.max_seq_len)\n",
+ "\n",
+ " def test_forward(self):\n",
+ " # Create a sample input tensor representing word embeddings\n",
+ " batch_size = 2\n",
+ " seq_length = 5\n",
+ " word_embeddings = torch.randn(batch_size, seq_length, self.d_model)\n",
+ "\n",
+ " # Forward pass through the PositionalEncoder module\n",
+ " output = self.positional_encoder(word_embeddings)\n",
+ "\n",
+ " # Check if the output shape matches the input shape\n",
+ " assert output.shape == (batch_size, seq_length, self.d_model)\n",
+ "\n",
+ " # Check if positional encoding is correctly applied\n",
+ " # Example: Verify if the first element of the first embedding vector matches the expected value\n",
+ " expected_first_element = torch.sin(torch.tensor([0.0])) * math.sqrt(self.d_model)\n",
+ " assert math.isclose(output[0, 0, 0].item(), expected_first_element.item(), rel_tol=1e-6)\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "In this code:\n",
+ "\n",
+ "We define a PositionalEncoder class that inherits from nn.Module.\n",
+ "The constructor initializes the positional encoding matrix (pe) based on the given d_model (dimension of the model) and max_seq_len (maximum sequence length).\n",
+ "The forward method scales the input embeddings (x) by the square root of the model dimension and adds the positional encoding matrix (pe) to the input embeddings.\n",
+ "Note that we're using PyTorch's Variable and autograd to ensure that the positional encoding is compatible with the autograd mechanism for backpropagation.\n",
+ "Finally, the PositionalEncoder class can be used within a larger PyTorch model to incorporate positional information into word embeddings."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Attention Layer\n",
+ "The Attention Layer in the Transformer model enables the model to focus on different parts of the input sequence when processing each token. It computes attention scores between each pair of tokens in the input sequence and generates a context vector for each token based on the importance of other tokens. This mechanism allows the model to capture long-range dependencies in the input sequence effectively.\n",
+ "\n",
+ "Implementation in PyTorch:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import torch\n",
+ "import torch.nn as nn\n",
+ "import torch.nn.functional as F\n",
+ "import math\n",
+ "\n",
+ "class MultiHeadAttention(nn.Module):\n",
+ " def __init__(self, heads, d_model, dropout=0.1):\n",
+ " super().__init__()\n",
+ " self.d_model = d_model\n",
+ " self.d_k = d_model // heads\n",
+ " self.h = heads\n",
+ " self.q_linear = nn.Linear(d_model, d_model)\n",
+ " self.v_linear = nn.Linear(d_model, d_model)\n",
+ " self.k_linear = nn.Linear(d_model, d_model)\n",
+ " self.dropout = nn.Dropout(dropout)\n",
+ " self.out = nn.Linear(d_model, d_model)\n",
+ " \n",
+ " def attention(self, q, k, v, d_k, mask=None, dropout=None):\n",
+ " scores = torch.matmul(q, k.transpose(-2, -1)) / math.sqrt(d_k)\n",
+ " if mask is not None:\n",
+ " mask = mask.unsqueeze(1)\n",
+ " scores = scores.masked_fill(mask == 0, -1e9)\n",
+ " scores = F.softmax(scores, dim=-1)\n",
+ " if dropout is not None:\n",
+ " scores = dropout(scores)\n",
+ " output = torch.matmul(scores, v)\n",
+ " return output\n",
+ " \n",
+ " def forward(self, q, k, v, mask=None):\n",
+ " bs = q.size(0)\n",
+ " k = self.k_linear(k).view(bs, -1, self.h, self.d_k)\n",
+ " q = self.q_linear(q).view(bs, -1, self.h, self.d_k)\n",
+ " v = self.v_linear(v).view(bs, -1, self.h, self.d_k)\n",
+ " k = k.transpose(1, 2)\n",
+ " q = q.transpose(1, 2)\n",
+ " v = v.transpose(1, 2)\n",
+ " scores = self.attention(q, k, v, self.d_k, mask, self.dropout)\n",
+ " concat = scores.transpose(1, 2).contiguous().view(bs, -1, self.d_model)\n",
+ " output = self.out(concat)\n",
+ " return output"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Check result by executing below... 📝 "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "jupyter": {
+ "source_hidden": true
+ },
+ "tags": [
+ "hide-input"
+ ]
+ },
+ "outputs": [],
+ "source": [
+ "%%ipytest -qq\n",
+ "\n",
+ "class TestMultiHeadAttention(unittest.TestCase):\n",
+ " def test_forward(self):\n",
+ " # Instantiate MultiHeadAttention module\n",
+ " heads = 4\n",
+ " d_model = 64\n",
+ " dropout = 0.1\n",
+ " multihead_attn = MultiHeadAttention(heads, d_model, dropout)\n",
+ "\n",
+ " # Create sample input tensors\n",
+ " batch_size = 2\n",
+ " seq_length = 5\n",
+ " q = torch.randn(batch_size, seq_length, d_model)\n",
+ " k = torch.randn(batch_size, seq_length, d_model)\n",
+ " v = torch.randn(batch_size, seq_length, d_model)\n",
+ " mask = torch.randint(0, 2, (batch_size, 1, seq_length)) # Example mask tensor\n",
+ "\n",
+ " # Forward pass through the MultiHeadAttention module\n",
+ " output = multihead_attn(q, k, v, mask)\n",
+ "\n",
+ " # Check output shape\n",
+ " self.assertEqual(output.shape, (batch_size, seq_length, d_model))\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "In this implementation:\n",
+ "\n",
+ "The MultiHeadAttention class defines a multi-head self-attention layer.\n",
+ "The forward method performs linear operations to divide inputs into multiple heads, computes attention scores, and aggregates the outputs of multiple heads."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Feedforward Layer\n",
+ "\n",
+ "The Position-wise Feedforward Layer in the Transformer model applies a simple feedforward neural network independently to each position in the sequence. It consists of two linear transformations with a non-linear activation function (commonly ReLU) applied in between. This layer helps capture complex interactions between different dimensions of the input embeddings.\n",
+ "\n",
+ "Implementation in PyTorch:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import torch\n",
+ "import torch.nn as nn\n",
+ "import torch.nn.functional as F\n",
+ "\n",
+ "class FeedForward(nn.Module):\n",
+ " def __init__(self, d_model, d_ff=2048, dropout=0.1):\n",
+ " super().__init__()\n",
+ " ## Set d_ff default to 2048\n",
+ " self.linear_1 = nn.Linear(d_model, d_ff)\n",
+ " self.dropout = nn.Dropout(dropout)\n",
+ " self.linear_2 = nn.Linear(d_ff, d_model)\n",
+ "\n",
+ " def forward(self, x):\n",
+ " x = self.dropout(F.relu(self.linear_1(x)))\n",
+ " x = self.linear_2(x)\n",
+ " return x"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Check result by executing below... 📝 "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "jupyter": {
+ "source_hidden": true
+ },
+ "tags": [
+ "hide-input"
+ ]
+ },
+ "outputs": [],
+ "source": [
+ "%%ipytest -qq\n",
+ "\n",
+ "class TestFeedForward(unittest.TestCase):\n",
+ " def test_forward(self):\n",
+ " # Instantiate FeedForward module\n",
+ " d_model = 512\n",
+ " d_ff = 2048\n",
+ " dropout = 0.1\n",
+ " feed_forward = FeedForward(d_model, d_ff, dropout)\n",
+ "\n",
+ " # Create sample input tensor\n",
+ " batch_size = 2\n",
+ " seq_length = 5\n",
+ " input_tensor = torch.randn(batch_size, seq_length, d_model)\n",
+ "\n",
+ " # Forward pass through the FeedForward module\n",
+ " output = feed_forward(input_tensor)\n",
+ "\n",
+ " # Check output shape\n",
+ " self.assertEqual(output.shape, (batch_size, seq_length, d_model))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "In this implementation:\n",
+ "\n",
+ "The FeedForward class defines a feedforward layer.\n",
+ "The forward method applies ReLU activation to the output of the first linear transformation, followed by dropout, and then performs the second linear transformation to produce the final output."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Residual Connection and Layer Normalization\n",
+ "\n",
+ "Residual Connection:\n",
+ "The Residual Connection, also known as skip connection, is a technique used in deep neural networks to mitigate the vanishing gradient problem and facilitate the flow of information through the network. In the context of the Transformer model, residual connections are added around each sub-layer (such as attention and feedforward layers) before applying layer normalization. This allows the model to learn residual representations and thus ease the optimization process.\n",
+ "\n",
+ "Layer Normalization:\n",
+ "Layer Normalization is a technique used to stabilize the training of deep neural networks by normalizing the activations of each layer. In the Transformer model, layer normalization is applied after each sub-layer (such as attention and feedforward layers) and before the residual connection. It normalizes the activations along the feature dimension, allowing the model to learn more robust representations and accelerate convergence during training.\n",
+ "\n",
+ "Implementation in PyTorch:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import torch\n",
+ "import torch.nn as nn\n",
+ "\n",
+ "class NormLayer(nn.Module):\n",
+ " def __init__(self, d_model, eps=1e-6):\n",
+ " super().__init__()\n",
+ " self.size = d_model\n",
+ " ## Layer normalization includes two learnable parameters\n",
+ " self.alpha = nn.Parameter(torch.ones(self.size))\n",
+ " self.bias = nn.Parameter(torch.zeros(self.size))\n",
+ " self.eps = eps\n",
+ " \n",
+ " def forward(self, x):\n",
+ " norm = self.alpha * (x - x.mean(dim=-1, keepdim=True)) / (x.std(dim=-1, keepdim=True) + self.eps) + self.bias\n",
+ " return norm"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Check result by executing below... 📝 "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "jupyter": {
+ "source_hidden": true
+ },
+ "tags": [
+ "hide-input"
+ ]
+ },
+ "outputs": [],
+ "source": [
+ "%%ipytest -qq\n",
+ "\n",
+ "class TestNormLayer(unittest.TestCase):\n",
+ " def test_forward(self):\n",
+ " # Instantiate NormLayer module\n",
+ " d_model = 512\n",
+ " eps = 1e-6\n",
+ " norm_layer = NormLayer(d_model, eps)\n",
+ "\n",
+ " # Create sample input tensor\n",
+ " batch_size = 2\n",
+ " seq_length = 5\n",
+ " input_tensor = torch.randn(batch_size, seq_length, d_model)\n",
+ "\n",
+ " # Forward pass through the NormLayer module\n",
+ " output = norm_layer(input_tensor)\n",
+ "\n",
+ " # Check output shape\n",
+ " self.assertEqual(output.shape, (batch_size, seq_length, d_model))"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "In this implementation:\n",
+ "\n",
+ "The NormLayer class defines a layer normalization layer.\n",
+ "The forward method computes the layer normalization using the given input tensor x."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Encoder and Decoder Structure\n",
+ "Encoder Structure:\n",
+ "The Encoder in the Transformer model consists of multiple stacked Encoder layers. Each Encoder layer typically contains a Multi-Head Attention sub-layer followed by a FeedForward sub-layer, each with Residual Connection and Layer Normalization.\n",
+ "\n",
+ "Decoder Structure:\n",
+ "Similarly, the Decoder in the Transformer model also consists of multiple stacked Decoder layers. Each Decoder layer contains three sub-layers:\n",
+ "\n",
+ "Masked Multi-Head Attention sub-layer to attend to previous tokens in the output sequence.\n",
+ "Multi-Head Attention sub-layer that attends to the encoder's output.\n",
+ "FeedForward sub-layer. Again, each sub-layer is followed by Residual Connection and Layer Normalization.\n",
+ "\n",
+ "Below are the Python implementations for the Encoder and Decoder structures:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class EncoderLayer(nn.Module):\n",
+ " def __init__(self, d_model, heads, dropout=0.1):\n",
+ " super().__init__()\n",
+ " self.norm_1 = Norm(d_model)\n",
+ " self.norm_2 = Norm(d_model)\n",
+ " self.attn = MultiHeadAttention(heads, d_model, dropout=dropout)\n",
+ " self.ff = FeedForward(d_model, dropout=dropout)\n",
+ " self.dropout_1 = nn.Dropout(dropout)\n",
+ " self.dropout_2 = nn.Dropout(dropout)\n",
+ "\n",
+ " def forward(self, x, mask):\n",
+ " x2 = self.norm_1(x)\n",
+ " x = x + self.dropout_1(self.attn(x2, x2, x2, mask))\n",
+ " x2 = self.norm_2(x)\n",
+ " x = x + self.dropout_2(self.ff(x2))\n",
+ " return x\n",
+ "\n",
+ "class Encoder(nn.Module):\n",
+ " def __init__(self, vocab_size, d_model, N, heads):\n",
+ " super().__init__()\n",
+ " self.N = N\n",
+ " self.embed = Embedder(vocab_size, d_model)\n",
+ " self.pe = PositionalEncoder(d_model)\n",
+ " self.layers = get_clones(EncoderLayer(d_model, heads), N)\n",
+ " self.norm = Norm(d_model)\n",
+ "\n",
+ " def forward(self, src, mask):\n",
+ " x = self.embed(src)\n",
+ " x = self.pe(x)\n",
+ " for i in range(self.N):\n",
+ " x = self.layers[i](x, mask)\n",
+ " return self.norm(x)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class DecoderLayer(nn.Module):\n",
+ " def __init__(self, d_model, heads, dropout=0.1):\n",
+ " super().__init__()\n",
+ " self.norm_1 = Norm(d_model)\n",
+ " self.norm_2 = Norm(d_model)\n",
+ " self.norm_3 = Norm(d_model)\n",
+ " self.dropout_1 = nn.Dropout(dropout)\n",
+ " self.dropout_2 = nn.Dropout(dropout)\n",
+ " self.dropout_3 = nn.Dropout(dropout)\n",
+ " self.attn_1 = MultiHeadAttention(heads, d_model, dropout=dropout)\n",
+ " self.attn_2 = MultiHeadAttention(heads, d_model, dropout=dropout)\n",
+ " self.ff = FeedForward(d_model, dropout=dropout)\n",
+ "\n",
+ " def forward(self, x, e_outputs, src_mask, trg_mask):\n",
+ " x2 = self.norm_1(x)\n",
+ " x = x + self.dropout_1(self.attn_1(x2, x2, x2, trg_mask))\n",
+ " x2 = self.norm_2(x)\n",
+ " x = x + self.dropout_2(self.attn_2(x2, e_outputs, e_outputs, src_mask))\n",
+ " x2 = self.norm_3(x)\n",
+ " x = x + self.dropout_3(self.ff(x2))\n",
+ " return x\n",
+ "\n",
+ "class Decoder(nn.Module):\n",
+ " def __init__(self, vocab_size, d_model, N, heads, dropout):\n",
+ " super().__init__()\n",
+ " self.N = N\n",
+ " self.embed = Embedder(vocab_size, d_model)\n",
+ " self.pe = PositionalEncoder(d_model, dropout=dropout)\n",
+ " self.layers = get_clones(DecoderLayer(d_model, heads, dropout), N)\n",
+ " self.norm = Norm(d_model)\n",
+ "\n",
+ " def forward(self, trg, e_outputs, src_mask, trg_mask):\n",
+ " x = self.embed(trg)\n",
+ " x = self.pe(x)\n",
+ " for i in range(self.N):\n",
+ " x = self.layers[i](x, e_outputs, src_mask, trg_mask)\n",
+ " return self.norm(x)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "In these implementations:\n",
+ "\n",
+ "The EncoderLayer and DecoderLayer classes define encoder and decoder layers, respectively.\n",
+ "The Encoder and Decoder classes define encoder and decoder modules, respectively, composed of multiple layers of encoder or decoder layers.\n",
+ "These classes follow the architecture described in the text, including the use of multi-head attention, feedforward layers, residual connections, and layer normalization."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "The overall implementation of the Transformer encoder and decoder structure:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import torch\n",
+ "import torch.nn as nn\n",
+ "import torch.nn.functional as F\n",
+ "import time\n",
+ "import numpy as np\n",
+ "\n",
+ "class Transformer(nn.Module):\n",
+ " def __init__(self, src_vocab, trg_vocab, d_model, N, heads, dropout):\n",
+ " super().__init__()\n",
+ " self.encoder = Encoder(src_vocab, d_model, N, heads, dropout)\n",
+ " self.decoder = Decoder(trg_vocab, d_model, N, heads, dropout)\n",
+ " self.out = nn.Linear(d_model, trg_vocab)\n",
+ "\n",
+ " def forward(self, src, trg, src_mask, trg_mask):\n",
+ " e_outputs = self.encoder(src, src_mask)\n",
+ " d_output = self.decoder(trg, e_outputs, src_mask, trg_mask)\n",
+ " output = self.out(d_output)\n",
+ " return output\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "The training process for the Transformer model:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "## Model parameters\n",
+ "d_model = 512\n",
+ "heads = 8\n",
+ "N = 6\n",
+ "src_vocab = len(EN_TEXT.vocab)\n",
+ "trg_vocab = len(FR_TEXT.vocab)\n",
+ "\n",
+ "## Initialize the model\n",
+ "model = Transformer(src_vocab, trg_vocab, d_model, N, heads)\n",
+ "\n",
+ "## Initialize optimizer\n",
+ "optim = torch.optim.Adam(model.parameters(), lr=0.0001, betas=(0.9, 0.98), eps=1e-9)\n",
+ "\n",
+ "## Training function\n",
+ "def train_model(epochs, print_every=100):\n",
+ " model.train()\n",
+ " start = time.time()\n",
+ " temp = start\n",
+ " total_loss = 0\n",
+ "\n",
+ " for epoch in range(epochs):\n",
+ " for i, batch in enumerate(train_iter):\n",
+ " src = batch.English.transpose(0, 1)\n",
+ " trg = batch.French.transpose(0, 1)\n",
+ " trg_input = trg[:, :-1]\n",
+ " targets = trg[:, 1:].contiguous().view(-1)\n",
+ " src_mask, trg_mask = create_masks(src, trg_input)\n",
+ "\n",
+ " preds = model(src, trg_input, src_mask, trg_mask)\n",
+ " optim.zero_grad()\n",
+ " loss = F.cross_entropy(preds.view(-1, preds.size(-1)), targets, ignore_index=target_pad)\n",
+ " loss.backward()\n",
+ " optim.step()\n",
+ " total_loss += loss.data[0]\n",
+ "\n",
+ " if (i + 1) % print_every == 0:\n",
+ " loss_avg = total_loss / print_every\n",
+ " print(\"time = %dm, epoch %d, iter = %d, loss = %.3f, %ds per %d iters\" % (\n",
+ " (time.time() - start) // 60, epoch + 1, i + 1, loss_avg, time.time() - temp, print_every))\n",
+ " total_loss = 0\n",
+ " temp = time.time()\n",
+ "\n",
+ "## Train the model\n",
+ "train_model(epochs=10)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Test the trained model:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def translate(model, src, max_len=80, custom_string=False):\n",
+ " model.eval()\n",
+ " if custom_string == True:\n",
+ " src = tokenize_en(src)\n",
+ " sentence = Variable(torch.LongTensor([[EN_TEXT.vocab.stoi[tok] for tok in sentence]])).cuda()\n",
+ " src_mask = (src != input_pad).unsqueeze(-2)\n",
+ " e_outputs = model.encoder(src, src_mask)\n",
+ " outputs = torch.zeros(max_len).type_as(src.data)\n",
+ " outputs[0] = torch.LongTensor([FR_TEXT.vocab.stoi['']])\n",
+ "\n",
+ " for i in range(1, max_len):\n",
+ " trg_mask = np.triu(np.ones((1, i, i), k=1).astype('uint8'))\n",
+ " trg_mask = Variable(torch.from_numpy(trg_mask) == 0).cuda()\n",
+ " out = model.out(model.decoder(outputs[:i].unsqueeze(0), e_outputs, src_mask, trg_mask))\n",
+ "\n",
+ " out = F.softmax(out, dim=-1)\n",
+ " val, ix = out[:, -1].data.topk(1)\n",
+ " outputs[i] = ix[0][0]\n",
+ "\n",
+ " if ix[0][0] == FR_TEXT.vocab.stoi['']:\n",
+ " break\n",
+ "\n",
+ " return ' '.join([FR_TEXT.vocab.itos[ix] for ix in outputs[:i]])\n"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Acknowledgments\n",
+ "\n",
+ "Thanks to the awesome open source project for Transformer learning, which inspire this chapter.\n",
+ "\n",
+ "- [chatgpt](https://openai.com/product/chatgpt)"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "open-machine-learning-jupyter-book",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.9.18"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git a/open-machine-learning-jupyter-book/llm/basic/attention.ipynb b/open-machine-learning-jupyter-book/llm/basic/attention.ipynb
new file mode 100644
index 000000000..2c4b9c098
--- /dev/null
+++ b/open-machine-learning-jupyter-book/llm/basic/attention.ipynb
@@ -0,0 +1,367 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "tags": [
+ "remove-cell"
+ ]
+ },
+ "source": [
+ "---\n",
+ "license:\n",
+ " code: MIT\n",
+ " content: CC-BY-4.0\n",
+ "github: https://github.com/ocademy-ai/machine-learning\n",
+ "venue: By Ocademy\n",
+ "open_access: true\n",
+ "bibliography:\n",
+ " - https://raw.githubusercontent.com/ocademy-ai/machine-learning/main/open-machine-learning-jupyter-book/references.bib\n",
+ "---"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "# Attention\n",
+ "## What is Attention?\n",
+ "\n",
+ "The attention mechanism describes a recent new group of layers in neural networks that has attracted a lot of interest in the past few years, especially in sequence tasks. There are a lot of different possible definitions of \"attention\" in the literature, but the one we will use here is the following: _the attention mechanism describes a weighted average of (sequence) elements with the weights dynamically computed based on an input query and elements' keys_. So what does this exactly mean? The goal is to take an average over the features of multiple elements. However, instead of weighting each element equally, we want to weight them depending on their actual values. In other words, we want to dynamically decide on which inputs we want to \"attend\" more than others. In particular, an attention mechanism has usually four parts we need to specify:\n",
+ "\n",
+ "* **Query**: The query is a feature vector that describes what we are looking for in the sequence, i.e. what would we maybe want to pay attention to.\n",
+ "* **Keys**: For each input element, we have a key which is again a feature vector. This feature vector roughly describes what the element is \"offering\", or when it might be important. The keys should be designed such that we can identify the elements we want to pay attention to based on the query.\n",
+ "* **Values**: For each input element, we also have a value vector. This feature vector is the one we want to average over.\n",
+ "* **Score function**: To rate which elements we want to pay attention to, we need to specify a score function $f_{attn}$. The score function takes the query and a key as input, and output the score/attention weight of the query-key pair. It is usually implemented by simple similarity metrics like a dot product, or a small MLP.\n",
+ "\n",
+ "\n",
+ "The weights of the average are calculated by a softmax over all score function outputs. Hence, we assign those value vectors a higher weight whose corresponding key is most similar to the query. If we try to describe it with pseudo-math, we can write: \n",
+ "\n",
+ "$$\n",
+ "\\alpha_i = \\frac{\\exp\\left(f_{attn}\\left(\\text{key}_i, \\text{query}\\right)\\right)}{\\sum_j \\exp\\left(f_{attn}\\left(\\text{key}_j, \\text{query}\\right)\\right)}, \\hspace{5mm} \\text{out} = \\sum_i \\alpha_i \\cdot \\text{value}_i\n",
+ "$$\n",
+ "\n",
+ "Visually, we can show the attention over a sequence of words as follows:\n",
+ "\n",
+ ":::{figure} ../image/attention_example.svg\n",
+ ":::\n",
+ "\n",
+ "For every word, we have one key and one value vector. The query is compared to all keys with a score function (in this case the dot product) to determine the weights. The softmax is not visualized for simplicity. Finally, the value vectors of all words are averaged using the attention weights.\n",
+ "\n",
+ "Most attention mechanisms differ in terms of what queries they use, how the key and value vectors are defined, and what score function is used. The attention applied inside the Transformer architecture is called **self-attention**. In self-attention, each sequence element provides a key, value, and query. For each element, we perform an attention layer where based on its query, we check the similarity of the all sequence elements' keys, and returned a different, averaged value vector for each element. We will now go into a bit more detail by first looking at the specific implementation of the attention mechanism which is in the Transformer case the scaled dot product attention."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Scaled Dot Product Attention\n",
+ "\n",
+ "The core concept behind self-attention is the scaled dot product attention. Our goal is to have an attention mechanism with which any element in a sequence can attend to any other while still being efficient to compute. The dot product attention takes as input a set of queries $Q\\in\\mathbb{R}^{T\\times d_k}$, keys $K\\in\\mathbb{R}^{T\\times d_k}$ and values $V\\in\\mathbb{R}^{T\\times d_v}$ where $T$ is the sequence length, and $d_k$ and $d_v$ are the hidden dimensionality for queries/keys and values respectively. For simplicity, we neglect the batch dimension for now. The attention value from element $i$ to $j$ is based on its similarity of the query $Q_i$ and key $K_j$, using the dot product as the similarity metric. In math, we calculate the dot product attention as follows:\n",
+ "\n",
+ "$$\\text{Attention}(Q,K,V)=\\text{softmax}\\left(\\frac{QK^T}{\\sqrt{d_k}}\\right)V$$\n",
+ "\n",
+ "The matrix multiplication $QK^T$ performs the dot product for every possible pair of queries and keys, resulting in a matrix of the shape $T\\times T$. Each row represents the attention logits for a specific element $i$ to all other elements in the sequence. On these, we apply a softmax and multiply with the value vector to obtain a weighted mean (the weights being determined by the attention). Another perspective on this attention mechanism offers the computation graph which is visualized below (figure credit - [Vaswani et al., 2017](https://arxiv.org/abs/1706.03762)).\n",
+ "\n",
+ ":::{figure} ../image/scaled_dot_product_attn.svg\n",
+ ":::\n",
+ "\n",
+ "One aspect we haven't discussed yet is the scaling factor of $1/\\sqrt{d_k}$. This scaling factor is crucial to maintain an appropriate variance of attention values after initialization. Remember that we intialize our layers with the intention of having equal variance throughout the model, and hence, $Q$ and $K$ might also have a variance close to $1$. However, performing a dot product over two vectors with a variance $\\sigma^2$ results in a scalar having $d_k$-times higher variance: \n",
+ "\n",
+ "$$q_i \\sim \\mathcal{N}(0,\\sigma^2), k_i \\sim \\mathcal{N}(0,\\sigma^2) \\to \\text{Var}\\left(\\sum_{i=1}^{d_k} q_i\\cdot k_i\\right) = \\sigma^4\\cdot d_k$$\n",
+ "\n",
+ "\n",
+ "If we do not scale down the variance back to $\\sim\\sigma^2$, the softmax over the logits will already saturate to $1$ for one random element and $0$ for all others. The gradients through the softmax will be close to zero so that we can't learn the parameters appropriately. Note that the extra factor of $\\sigma^2$, i.e., having $\\sigma^4$ instead of $\\sigma^2$, is usually not an issue, since we keep the original variance $\\sigma^2$ close to $1$ anyways.\n",
+ "\n",
+ "The block `Mask (opt.)` in the diagram above represents the optional masking of specific entries in the attention matrix. This is for instance used if we stack multiple sequences with different lengths into a batch. To still benefit from parallelization in PyTorch, we pad the sentences to the same length and mask out the padding tokens during the calculation of the attention values. This is usually done by setting the respective attention logits to a very low value. \n",
+ "\n",
+ "After we have discussed the details of the scaled dot product attention block, we can write a function below which computes the output features given the triple of queries, keys, and values:"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Below, we import the standard libraries."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 16,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Device: cpu\n"
+ ]
+ }
+ ],
+ "source": [
+ "## Standard libraries\n",
+ "import os\n",
+ "import numpy as np\n",
+ "import random\n",
+ "import math\n",
+ "import json\n",
+ "from functools import partial\n",
+ "\n",
+ "## Imports for plotting\n",
+ "import matplotlib.pyplot as plt\n",
+ "plt.set_cmap('cividis')\n",
+ "%matplotlib inline\n",
+ "from matplotlib.colors import to_rgb\n",
+ "import matplotlib\n",
+ "matplotlib.rcParams['lines.linewidth'] = 2.0\n",
+ "import seaborn as sns\n",
+ "sns.reset_orig()\n",
+ "\n",
+ "## tqdm for loading bars\n",
+ "from tqdm.notebook import tqdm\n",
+ "\n",
+ "## PyTorch\n",
+ "import torch\n",
+ "import torch.nn as nn\n",
+ "import torch.nn.functional as F\n",
+ "import torch.utils.data as data\n",
+ "import torch.optim as optim\n",
+ "\n",
+ "## Torchvision\n",
+ "import torchvision\n",
+ "from torchvision.datasets import CIFAR100\n",
+ "from torchvision import transforms\n",
+ "\n",
+ "# PyTorch Lightning\n",
+ "try:\n",
+ " import pytorch_lightning as pl\n",
+ "except ModuleNotFoundError: # Google Colab does not have PyTorch Lightning installed by default. Hence, we do it here if necessary\n",
+ " !pip install --quiet pytorch-lightning>=1.4\n",
+ " import pytorch_lightning as pl\n",
+ "from pytorch_lightning.callbacks import LearningRateMonitor, ModelCheckpoint\n",
+ "\n",
+ "# Ensure that all operations are deterministic on GPU (if used) for reproducibility\n",
+ "torch.backends.cudnn.deterministic = True\n",
+ "torch.backends.cudnn.benchmark = False\n",
+ "\n",
+ "device = torch.device(\"cuda:0\") if torch.cuda.is_available() else torch.device(\"cpu\")\n",
+ "print(\"Device:\", device)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 17,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def scaled_dot_product(q, k, v, mask=None):\n",
+ " d_k = q.size()[-1]\n",
+ " attn_logits = torch.matmul(q, k.transpose(-2, -1))\n",
+ " attn_logits = attn_logits / math.sqrt(d_k)\n",
+ " if mask is not None:\n",
+ " attn_logits = attn_logits.masked_fill(mask == 0, -9e15)\n",
+ " attention = F.softmax(attn_logits, dim=-1)\n",
+ " values = torch.matmul(attention, v)\n",
+ " return values, attention"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Note that our code above supports any additional dimensionality in front of the sequence length so that we can also use it for batches. However, for a better understanding, let's generate a few random queries, keys, and value vectors, and calculate the attention outputs:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 18,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "Seed set to 42\n"
+ ]
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Q\n",
+ " tensor([[ 0.3367, 0.1288],\n",
+ " [ 0.2345, 0.2303],\n",
+ " [-1.1229, -0.1863]])\n",
+ "K\n",
+ " tensor([[ 2.2082, -0.6380],\n",
+ " [ 0.4617, 0.2674],\n",
+ " [ 0.5349, 0.8094]])\n",
+ "V\n",
+ " tensor([[ 1.1103, -1.6898],\n",
+ " [-0.9890, 0.9580],\n",
+ " [ 1.3221, 0.8172]])\n",
+ "Values\n",
+ " tensor([[ 0.5698, -0.1520],\n",
+ " [ 0.5379, -0.0265],\n",
+ " [ 0.2246, 0.5556]])\n",
+ "Attention\n",
+ " tensor([[0.4028, 0.2886, 0.3086],\n",
+ " [0.3538, 0.3069, 0.3393],\n",
+ " [0.1303, 0.4630, 0.4067]])\n"
+ ]
+ }
+ ],
+ "source": [
+ "seq_len, d_k = 3, 2\n",
+ "pl.seed_everything(42)\n",
+ "q = torch.randn(seq_len, d_k)\n",
+ "k = torch.randn(seq_len, d_k)\n",
+ "v = torch.randn(seq_len, d_k)\n",
+ "values, attention = scaled_dot_product(q, k, v)\n",
+ "print(\"Q\\n\", q)\n",
+ "print(\"K\\n\", k)\n",
+ "print(\"V\\n\", v)\n",
+ "print(\"Values\\n\", values)\n",
+ "print(\"Attention\\n\", attention)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Before continuing, make sure you can follow the calculation of the specific values here, and also check it by hand. It is important to fully understand how the scaled dot product attention is calculated."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Multi-Head Attention\n",
+ "\n",
+ "The scaled dot product attention allows a network to attend over a sequence. However, often there are multiple different aspects a sequence element wants to attend to, and a single weighted average is not a good option for it. This is why we extend the attention mechanisms to multiple heads, i.e. multiple different query-key-value triplets on the same features. Specifically, given a query, key, and value matrix, we transform those into $h$ sub-queries, sub-keys, and sub-values, which we pass through the scaled dot product attention independently. Afterward, we concatenate the heads and combine them with a final weight matrix. Mathematically, we can express this operation as:\n",
+ "\n",
+ "$$\n",
+ "\\begin{split}\n",
+ " \\text{Multihead}(Q,K,V) & = \\text{Concat}(\\text{head}_1,...,\\text{head}_h)W^{O}\\\\\n",
+ " \\text{where } \\text{head}_i & = \\text{Attention}(QW_i^Q,KW_i^K, VW_i^V)\n",
+ "\\end{split}\n",
+ "$$\n",
+ "\n",
+ "We refer to this as Multi-Head Attention layer with the learnable parameters $W_{1...h}^{Q}\\in\\mathbb{R}^{D\\times d_k}$, $W_{1...h}^{K}\\in\\mathbb{R}^{D\\times d_k}$, $W_{1...h}^{V}\\in\\mathbb{R}^{D\\times d_v}$, and $W^{O}\\in\\mathbb{R}^{h\\cdot d_v\\times d_{out}}$ ($D$ being the input dimensionality). Expressed in a computational graph, we can visualize it as below (figure credit - [Vaswani et al., 2017](https://arxiv.org/abs/1706.03762)).\n",
+ "\n",
+ ":::{figure} ../image/multihead_attention.svg\n",
+ ":::\n",
+ "\n",
+ "How are we applying a Multi-Head Attention layer in a neural network, where we don't have an arbitrary query, key, and value vector as input? Looking at the computation graph above, a simple but effective implementation is to set the current feature map in a NN, $X\\in\\mathbb{R}^{B\\times T\\times d_{\\text{model}}}$, as $Q$, $K$ and $V$ ($B$ being the batch size, $T$ the sequence length, $d_{\\text{model}}$ the hidden dimensionality of $X$). The consecutive weight matrices $W^{Q}$, $W^{K}$, and $W^{V}$ can transform $X$ to the corresponding feature vectors that represent the queries, keys, and values of the input. Using this approach, we can implement the Multi-Head Attention module below."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 19,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "# Helper function to support different mask shapes.\n",
+ "# Output shape supports (batch_size, number of heads, seq length, seq length)\n",
+ "# If 2D: broadcasted over batch size and number of heads\n",
+ "# If 3D: broadcasted over number of heads\n",
+ "# If 4D: leave as is\n",
+ "def expand_mask(mask):\n",
+ " assert mask.ndim >= 2, \"Mask must be at least 2-dimensional with seq_length x seq_length\"\n",
+ " if mask.ndim == 3:\n",
+ " mask = mask.unsqueeze(1)\n",
+ " while mask.ndim < 4:\n",
+ " mask = mask.unsqueeze(0)\n",
+ " return mask"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 20,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class MultiheadAttention(nn.Module):\n",
+ " \n",
+ " def __init__(self, input_dim, embed_dim, num_heads):\n",
+ " super().__init__()\n",
+ " assert embed_dim % num_heads == 0, \"Embedding dimension must be 0 modulo number of heads.\"\n",
+ " \n",
+ " self.embed_dim = embed_dim\n",
+ " self.num_heads = num_heads\n",
+ " self.head_dim = embed_dim // num_heads\n",
+ " \n",
+ " # Stack all weight matrices 1...h together for efficiency\n",
+ " # Note that in many implementations you see \"bias=False\" which is optional\n",
+ " self.qkv_proj = nn.Linear(input_dim, 3*embed_dim)\n",
+ " self.o_proj = nn.Linear(embed_dim, embed_dim)\n",
+ " \n",
+ " self._reset_parameters()\n",
+ "\n",
+ " def _reset_parameters(self):\n",
+ " # Original Transformer initialization, see PyTorch documentation\n",
+ " nn.init.xavier_uniform_(self.qkv_proj.weight)\n",
+ " self.qkv_proj.bias.data.fill_(0)\n",
+ " nn.init.xavier_uniform_(self.o_proj.weight)\n",
+ " self.o_proj.bias.data.fill_(0)\n",
+ "\n",
+ " def forward(self, x, mask=None, return_attention=False):\n",
+ " batch_size, seq_length, _ = x.size()\n",
+ " if mask is not None:\n",
+ " mask = expand_mask(mask)\n",
+ " qkv = self.qkv_proj(x)\n",
+ " \n",
+ " # Separate Q, K, V from linear output\n",
+ " qkv = qkv.reshape(batch_size, seq_length, self.num_heads, 3*self.head_dim)\n",
+ " qkv = qkv.permute(0, 2, 1, 3) # [Batch, Head, SeqLen, Dims]\n",
+ " q, k, v = qkv.chunk(3, dim=-1)\n",
+ " \n",
+ " # Determine value outputs\n",
+ " values, attention = scaled_dot_product(q, k, v, mask=mask)\n",
+ " values = values.permute(0, 2, 1, 3) # [Batch, SeqLen, Head, Dims]\n",
+ " values = values.reshape(batch_size, seq_length, self.embed_dim)\n",
+ " o = self.o_proj(values)\n",
+ " \n",
+ " if return_attention:\n",
+ " return o, attention\n",
+ " else:\n",
+ " return o"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "One crucial characteristic of the multi-head attention is that it is permutation-equivariant with respect to its inputs. This means that if we switch two input elements in the sequence, e.g. $X_1\\leftrightarrow X_2$ (neglecting the batch dimension for now), the output is exactly the same besides the elements 1 and 2 switched. Hence, the multi-head attention is actually looking at the input not as a sequence, but as a set of elements. This property makes the multi-head attention block and the Transformer architecture so powerful and widely applicable! But what if the order of the input is actually important for solving the task, like language modeling? The answer is to encode the position in the input features, which we will take a closer look at later (topic _Positional encodings_ below).\n",
+ "\n",
+ "Before moving on to creating the Transformer architecture, we can compare the self-attention operation with our other common layer competitors for sequence data: convolutions and recurrent neural networks. Below you can find a table by [Vaswani et al. (2017)](https://arxiv.org/abs/1706.03762) on the complexity per layer, the number of sequential operations, and maximum path length. The complexity is measured by the upper bound of the number of operations to perform, while the maximum path length represents the maximum number of steps a forward or backward signal has to traverse to reach any other position. The lower this length, the better gradient signals can backpropagate for long-range dependencies. Let's take a look at the table below:\n",
+ "\n",
+ ":::{figure} ../image/comparison_conv_rnn.svg\n",
+ ":::\n",
+ "\n",
+ "$n$ is the sequence length, $d$ is the representation dimension and $k$ is the kernel size of convolutions. In contrast to recurrent networks, the self-attention layer can parallelize all its operations making it much faster to execute for smaller sequence lengths. However, when the sequence length exceeds the hidden dimensionality, self-attention becomes more expensive than RNNs. One way of reducing the computational cost for long sequences is by restricting the self-attention to a neighborhood of inputs to attend over, denoted by $r$. Nevertheless, there has been recently a lot of work on more efficient Transformer architectures that still allow long dependencies, of which you can find an overview in the paper by [Tay et al. (2020)](https://arxiv.org/abs/2009.06732) if interested."
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3 (ipykernel)",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.10.4"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}
diff --git a/open-machine-learning-jupyter-book/llm/basic/basic.ipynb b/open-machine-learning-jupyter-book/llm/basic/basic.ipynb
new file mode 100644
index 000000000..ef8e960b7
--- /dev/null
+++ b/open-machine-learning-jupyter-book/llm/basic/basic.ipynb
@@ -0,0 +1,64 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "tags": [
+ "remove-cell"
+ ]
+ },
+ "source": [
+ "---\n",
+ "license:\n",
+ " code: MIT\n",
+ " content: CC-BY-4.0\n",
+ "github: https://github.com/ocademy-ai/machine-learning\n",
+ "venue: By Ocademy\n",
+ "open_access: true\n",
+ "bibliography:\n",
+ " - https://raw.githubusercontent.com/ocademy-ai/machine-learning/main/open-machine-learning-jupyter-book/references.bib\n",
+ "---"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "# Large Language Models Basic\n",
+ "In these sections, we will explore the attention mechanism, which allows models to focus on specific parts of the input during processing. We will study the Transformer model architecture, which serves as the cornerstone for many state-of-the-art language models, and how it has fundamentally transformed the field of Natural Language Processing (NLP). Additionally, we will introduce generative pre-trained language models like GPT, delve into the network structures of large language models, optimization techniques for attention mechanisms, and practical applications stemming from these foundations.\n",
+ "\n",
+ ":::{figure} ../image/llm.png\n",
+ ":::"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ ":::{tableofcontents}\n",
+ ":::"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "open-machine-learning-jupyter-book",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.9.18"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}
diff --git a/open-machine-learning-jupyter-book/llm/basic/transformer.ipynb b/open-machine-learning-jupyter-book/llm/basic/transformer.ipynb
new file mode 100644
index 000000000..4143ec705
--- /dev/null
+++ b/open-machine-learning-jupyter-book/llm/basic/transformer.ipynb
@@ -0,0 +1,20020 @@
+{
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "metadata": {
+ "tags": [
+ "remove-cell"
+ ]
+ },
+ "source": [
+ "---\n",
+ "license:\n",
+ " code: MIT\n",
+ " content: CC-BY-4.0\n",
+ "github: https://github.com/ocademy-ai/machine-learning\n",
+ "venue: By Ocademy\n",
+ "open_access: true\n",
+ "bibliography:\n",
+ " - https://raw.githubusercontent.com/ocademy-ai/machine-learning/main/open-machine-learning-jupyter-book/references.bib\n",
+ "---"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "# Transformer\n",
+ "In this tutorial, we will discuss one of the most impactful architectures of the last 2 years: the Transformer model. Since the paper [Attention Is All You Need](https://arxiv.org/abs/1706.03762) by Vaswani et al. had been published in 2017, the Transformer architecture has continued to beat benchmarks in many domains, most importantly in Natural Language Processing. Transformers with an incredible amount of parameters can generate long, convincing [essays](https://www.theguardian.com/commentisfree/2020/sep/08/robot-wrote-this-article-gpt-3), and opened up new application fields of AI. As the hype of the Transformer architecture seems not to come to an end in the next years, it is important to understand how it works, and have implemented it yourself, which we will do in this notebook. We focus here on what makes the Transformer and self-attention so powerful in general.\n",
+ "\n",
+ "Below, we import the standard libraries."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 1,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Device: cuda:0\n"
+ ]
+ }
+ ],
+ "source": [
+ "## Standard libraries\n",
+ "import os\n",
+ "import numpy as np \n",
+ "import random\n",
+ "import math\n",
+ "import json\n",
+ "from functools import partial\n",
+ "\n",
+ "## Imports for plotting\n",
+ "import matplotlib.pyplot as plt\n",
+ "plt.set_cmap('cividis')\n",
+ "%matplotlib inline \n",
+ "from matplotlib.colors import to_rgb\n",
+ "import matplotlib\n",
+ "matplotlib.rcParams['lines.linewidth'] = 2.0\n",
+ "import seaborn as sns\n",
+ "sns.reset_orig()\n",
+ "\n",
+ "## tqdm for loading bars\n",
+ "from tqdm.notebook import tqdm\n",
+ "\n",
+ "## PyTorch\n",
+ "import torch\n",
+ "import torch.nn as nn\n",
+ "import torch.nn.functional as F\n",
+ "import torch.utils.data as data\n",
+ "import torch.optim as optim\n",
+ "\n",
+ "## Torchvision\n",
+ "import torchvision\n",
+ "from torchvision.datasets import CIFAR100\n",
+ "from torchvision import transforms\n",
+ "\n",
+ "# PyTorch Lightning\n",
+ "try:\n",
+ " import pytorch_lightning as pl\n",
+ "except ModuleNotFoundError: # Google Colab does not have PyTorch Lightning installed by default. Hence, we do it here if necessary\n",
+ " !pip install --quiet pytorch-lightning>=1.4\n",
+ " import pytorch_lightning as pl\n",
+ "from pytorch_lightning.callbacks import LearningRateMonitor, ModelCheckpoint\n",
+ "\n",
+ "# Path to the folder where the datasets are/should be downloaded (e.g. CIFAR10)\n",
+ "DATASET_PATH = \"./data\"\n",
+ "# Path to the folder where the pretrained models are saved\n",
+ "CHECKPOINT_PATH = \"./saved_models\"\n",
+ "\n",
+ "# Setting the seed\n",
+ "pl.seed_everything(42)\n",
+ "\n",
+ "# Ensure that all operations are deterministic on GPU (if used) for reproducibility\n",
+ "torch.backends.cudnn.deterministic = True\n",
+ "torch.backends.cudnn.benchmark = False\n",
+ "\n",
+ "device = torch.device(\"cuda:0\") if torch.cuda.is_available() else torch.device(\"cpu\")\n",
+ "print(\"Device:\", device)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Two pre-trained models are downloaded below. Make sure to have adjusted your `CHECKPOINT_PATH` before running this code if not already done."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 2,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import urllib.request\n",
+ "from urllib.error import HTTPError\n",
+ "# Github URL where saved models are stored for this tutorial\n",
+ "base_url = \"https://raw.githubusercontent.com/phlippe/saved_models/main/tutorial6/\"\n",
+ "# Files to download\n",
+ "pretrained_files = [\"ReverseTask.ckpt\", \"SetAnomalyTask.ckpt\"]\n",
+ "\n",
+ "# Create checkpoint path if it doesn't exist yet\n",
+ "os.makedirs(CHECKPOINT_PATH, exist_ok=True)\n",
+ "\n",
+ "# For each file, check whether it already exists. If not, try downloading it.\n",
+ "for file_name in pretrained_files:\n",
+ " file_path = os.path.join(CHECKPOINT_PATH, file_name)\n",
+ " if \"/\" in file_name:\n",
+ " os.makedirs(file_path.rsplit(\"/\",1)[0], exist_ok=True)\n",
+ " if not os.path.isfile(file_path):\n",
+ " file_url = base_url + file_name\n",
+ " print(f\"Downloading {file_url}...\")\n",
+ " try:\n",
+ " urllib.request.urlretrieve(file_url, file_path)\n",
+ " except HTTPError as e:\n",
+ " print(\"Something went wrong. Please try to download the file from the GDrive folder, or contact the author with the full output including the following error:\\n\", e)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## The Transformer architecture\n",
+ "\n",
+ "In the first part of this notebook, we will implement the Transformer architecture by hand. As the architecture is so popular, there already exists a Pytorch module `nn.Transformer` ([documentation](https://pytorch.org/docs/stable/generated/torch.nn.Transformer.html)) and a [tutorial](https://pytorch.org/tutorials/beginner/transformer_tutorial.html) on how to use it for next token prediction. However, we will implement it here ourselves, to get through to the smallest details."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Transformer Encoder\n",
+ "\n",
+ "Next, we will look at how to apply the multi-head attention block inside the Transformer architecture. Originally, the Transformer model was designed for machine translation. Hence, it got an encoder-decoder structure where the encoder takes as input the sentence in the original language and generates an attention-based representation. On the other hand, the decoder attends over the encoded information and generates the translated sentence in an autoregressive manner, as in a standard RNN. While this structure is extremely useful for Sequence-to-Sequence tasks with the necessity of autoregressive decoding, we will focus here on the encoder part. Many advances in NLP have been made using pure encoder-based Transformer models (if interested, models include the [BERT](https://arxiv.org/abs/1810.04805)-family, the [Vision Transformer](https://arxiv.org/abs/2010.11929), and more), and in our tutorial, we will also mainly focus on the encoder part. If you have understood the encoder architecture, the decoder is a very small step to implement as well. The full Transformer architecture looks as follows (figure credit - [Vaswani et al., 2017](https://arxiv.org/abs/1706.03762)).:\n",
+ "\n",
+ ":::{figure} ../image/transformer_architecture.svg\n",
+ ":::\n",
+ "\n",
+ "The encoder consists of $N$ identical blocks that are applied in sequence. Taking as input $x$, it is first passed through a Multi-Head Attention block as we have implemented above. The output is added to the original input using a residual connection, and we apply a consecutive Layer Normalization on the sum. Overall, it calculates $\\text{LayerNorm}(x+\\text{Multihead}(x,x,x))$ ($x$ being $Q$, $K$ and $V$ input to the attention layer). The residual connection is crucial in the Transformer architecture for two reasons: \n",
+ "\n",
+ "1. Similar to ResNets, Transformers are designed to be very deep. Some models contain more than 24 blocks in the encoder. Hence, the residual connections are crucial for enabling a smooth gradient flow through the model.\n",
+ "2. Without the residual connection, the information about the original sequence is lost. Remember that the Multi-Head Attention layer ignores the position of elements in a sequence, and can only learn it based on the input features. Removing the residual connections would mean that this information is lost after the first attention layer (after initialization), and with a randomly initialized query and key vector, the output vectors for position $i$ has no relation to its original input. All outputs of the attention are likely to represent similar/same information, and there is no chance for the model to distinguish which information came from which input element. An alternative option to residual connection would be to fix at least one head to focus on its original input, but this is very inefficient and does not have the benefit of the improved gradient flow.\n",
+ "\n",
+ "The Layer Normalization also plays an important role in the Transformer architecture as it enables faster training and provides small regularization. Additionally, it ensures that the features are in a similar magnitude among the elements in the sequence. We are not using Batch Normalization because it depends on the batch size which is often small with Transformers (they require a lot of GPU memory), and BatchNorm has shown to perform particularly bad in language as the features of words tend to have a much higher variance (there are many, very rare words which need to be considered for a good distribution estimate).\n",
+ "\n",
+ "Additionally to the Multi-Head Attention, a small fully connected feed-forward network is added to the model, which is applied to each position separately and identically. Specifically, the model uses a Linear$\\to$ReLU$\\to$Linear MLP. The full transformation including the residual connection can be expressed as: \n",
+ "\n",
+ "$$\n",
+ "\\begin{split}\n",
+ " \\text{FFN}(x) & = \\max(0, xW_1+b_1)W_2 + b_2\\\\\n",
+ " x & = \\text{LayerNorm}(x + \\text{FFN}(x))\n",
+ "\\end{split}\n",
+ "$$\n",
+ "\n",
+ "This MLP adds extra complexity to the model and allows transformations on each sequence element separately. You can imagine as this allows the model to \"post-process\" the new information added by the previous Multi-Head Attention, and prepare it for the next attention block. Usually, the inner dimensionality of the MLP is 2-8$\\times$ larger than $d_{\\text{model}}$, i.e. the dimensionality of the original input $x$. The general advantage of a wider layer instead of a narrow, multi-layer MLP is the faster, parallelizable execution.\n",
+ "\n",
+ "Finally, after looking at all parts of the encoder architecture, we can start implementing it below. We first start by implementing a single encoder block. Additionally to the layers described above, we will add dropout layers in the MLP and on the output of the MLP and Multi-Head Attention for regularization."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 6,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class EncoderBlock(nn.Module):\n",
+ " \n",
+ " def __init__(self, input_dim, num_heads, dim_feedforward, dropout=0.0):\n",
+ " \"\"\"\n",
+ " Inputs:\n",
+ " input_dim - Dimensionality of the input\n",
+ " num_heads - Number of heads to use in the attention block\n",
+ " dim_feedforward - Dimensionality of the hidden layer in the MLP\n",
+ " dropout - Dropout probability to use in the dropout layers\n",
+ " \"\"\"\n",
+ " super().__init__()\n",
+ " \n",
+ " # Attention layer\n",
+ " self.self_attn = MultiheadAttention(input_dim, input_dim, num_heads)\n",
+ " \n",
+ " # Two-layer MLP\n",
+ " self.linear_net = nn.Sequential(\n",
+ " nn.Linear(input_dim, dim_feedforward),\n",
+ " nn.Dropout(dropout),\n",
+ " nn.ReLU(inplace=True),\n",
+ " nn.Linear(dim_feedforward, input_dim)\n",
+ " )\n",
+ " \n",
+ " # Layers to apply in between the main layers\n",
+ " self.norm1 = nn.LayerNorm(input_dim)\n",
+ " self.norm2 = nn.LayerNorm(input_dim)\n",
+ " self.dropout = nn.Dropout(dropout)\n",
+ "\n",
+ " def forward(self, x, mask=None):\n",
+ " # Attention part\n",
+ " attn_out = self.self_attn(x, mask=mask)\n",
+ " x = x + self.dropout(attn_out)\n",
+ " x = self.norm1(x)\n",
+ " \n",
+ " # MLP part\n",
+ " linear_out = self.linear_net(x)\n",
+ " x = x + self.dropout(linear_out)\n",
+ " x = self.norm2(x)\n",
+ " \n",
+ " return x"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Based on this block, we can implement a module for the full Transformer encoder. Additionally to a forward function that iterates through the sequence of encoder blocks, we also provide a function called `get_attention_maps`. The idea of this function is to return the attention probabilities for all Multi-Head Attention blocks in the encoder. This helps us in understanding, and in a sense, explaining the model. However, the attention probabilities should be interpreted with a grain of salt as it does not necessarily reflect the true interpretation of the model (there is a series of papers about this, including [Attention is not Explanation](https://arxiv.org/abs/1902.10186) and [Attention is not not Explanation](https://arxiv.org/abs/1908.04626))."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 7,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class TransformerEncoder(nn.Module):\n",
+ " \n",
+ " def __init__(self, num_layers, **block_args):\n",
+ " super().__init__()\n",
+ " self.layers = nn.ModuleList([EncoderBlock(**block_args) for _ in range(num_layers)])\n",
+ "\n",
+ " def forward(self, x, mask=None):\n",
+ " for l in self.layers:\n",
+ " x = l(x, mask=mask)\n",
+ " return x\n",
+ "\n",
+ " def get_attention_maps(self, x, mask=None):\n",
+ " attention_maps = []\n",
+ " for l in self.layers:\n",
+ " _, attn_map = l.self_attn(x, mask=mask, return_attention=True)\n",
+ " attention_maps.append(attn_map)\n",
+ " x = l(x)\n",
+ " return attention_maps"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Positional encoding\n",
+ "\n",
+ "We have discussed before that the Multi-Head Attention block is permutation-equivariant, and cannot distinguish whether an input comes before another one in the sequence or not. In tasks like language understanding, however, the position is important for interpreting the input words. The position information can therefore be added via the input features. We could learn a embedding for every possible position, but this would not generalize to a dynamical input sequence length. Hence, the better option is to use feature patterns that the network can identify from the features and potentially generalize to larger sequences. The specific pattern chosen by Vaswani et al. are sine and cosine functions of different frequencies, as follows:\n",
+ "\n",
+ "$$\n",
+ "PE_{(pos,i)} = \\begin{cases}\n",
+ " \\sin\\left(\\frac{pos}{10000^{i/d_{\\text{model}}}}\\right) & \\text{if}\\hspace{3mm} i \\text{ mod } 2=0\\\\\n",
+ " \\cos\\left(\\frac{pos}{10000^{(i-1)/d_{\\text{model}}}}\\right) & \\text{otherwise}\\\\\n",
+ "\\end{cases}\n",
+ "$$\n",
+ "\n",
+ "$PE_{(pos,i)}$ represents the position encoding at position $pos$ in the sequence, and hidden dimensionality $i$. These values, concatenated for all hidden dimensions, are added to the original input features (in the Transformer visualization above, see \"Positional encoding\"), and constitute the position information. We distinguish between even ($i \\text{ mod } 2=0$) and uneven ($i \\text{ mod } 2=1$) hidden dimensionalities where we apply a sine/cosine respectively. The intuition behind this encoding is that you can represent $PE_{(pos+k,:)}$ as a linear function of $PE_{(pos,:)}$, which might allow the model to easily attend to relative positions. The wavelengths in different dimensions range from $2\\pi$ to $10000\\cdot 2\\pi$.\n",
+ "\n",
+ "The positional encoding is implemented below. The code is taken from the [PyTorch tutorial](https://pytorch.org/tutorials/beginner/transformer_tutorial.html#define-the-model) about Transformers on NLP and adjusted for our purposes."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 8,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class PositionalEncoding(nn.Module):\n",
+ "\n",
+ " def __init__(self, d_model, max_len=5000):\n",
+ " \"\"\"\n",
+ " Inputs\n",
+ " d_model - Hidden dimensionality of the input.\n",
+ " max_len - Maximum length of a sequence to expect.\n",
+ " \"\"\"\n",
+ " super().__init__()\n",
+ "\n",
+ " # Create matrix of [SeqLen, HiddenDim] representing the positional encoding for max_len inputs\n",
+ " pe = torch.zeros(max_len, d_model)\n",
+ " position = torch.arange(0, max_len, dtype=torch.float).unsqueeze(1)\n",
+ " div_term = torch.exp(torch.arange(0, d_model, 2).float() * (-math.log(10000.0) / d_model))\n",
+ " pe[:, 0::2] = torch.sin(position * div_term)\n",
+ " pe[:, 1::2] = torch.cos(position * div_term)\n",
+ " pe = pe.unsqueeze(0)\n",
+ " \n",
+ " # register_buffer => Tensor which is not a parameter, but should be part of the modules state.\n",
+ " # Used for tensors that need to be on the same device as the module.\n",
+ " # persistent=False tells PyTorch to not add the buffer to the state dict (e.g. when we save the model) \n",
+ " self.register_buffer('pe', pe, persistent=False)\n",
+ "\n",
+ " def forward(self, x):\n",
+ " x = x + self.pe[:, :x.size(1)]\n",
+ " return x"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "To understand the positional encoding, we can visualize it below. We will generate an image of the positional encoding over hidden dimensionality and position in a sequence. Each pixel, therefore, represents the change of the input feature we perform to encode the specific position. Let's do it below."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 9,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "application/pdf": "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\n",
+ "image/svg+xml": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 2020-11-09T10:43:19.865866 \n",
+ " image/svg+xml \n",
+ " \n",
+ " \n",
+ " Matplotlib v3.3.2, https://matplotlib.org/ \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "encod_block = PositionalEncoding(d_model=48, max_len=96)\n",
+ "pe = encod_block.pe.squeeze().T.cpu().numpy()\n",
+ "\n",
+ "fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(8,3))\n",
+ "pos = ax.imshow(pe, cmap=\"RdGy\", extent=(1,pe.shape[1]+1,pe.shape[0]+1,1))\n",
+ "fig.colorbar(pos, ax=ax)\n",
+ "ax.set_xlabel(\"Position in sequence\")\n",
+ "ax.set_ylabel(\"Hidden dimension\")\n",
+ "ax.set_title(\"Positional encoding over hidden dimensions\")\n",
+ "ax.set_xticks([1]+[i*10 for i in range(1,1+pe.shape[1]//10)])\n",
+ "ax.set_yticks([1]+[i*10 for i in range(1,1+pe.shape[0]//10)])\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "You can clearly see the sine and cosine waves with different wavelengths that encode the position in the hidden dimensions. Specifically, we can look at the sine/cosine wave for each hidden dimension separately, to get a better intuition of the pattern. Below we visualize the positional encoding for the hidden dimensions $1$, $2$, $3$ and $4$."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 10,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "application/pdf": "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\n",
+ "image/svg+xml": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 2020-11-09T10:43:20.404975 \n",
+ " image/svg+xml \n",
+ " \n",
+ " \n",
+ " Matplotlib v3.3.2, https://matplotlib.org/ \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "sns.set_theme()\n",
+ "fig, ax = plt.subplots(2, 2, figsize=(12,4))\n",
+ "ax = [a for a_list in ax for a in a_list]\n",
+ "for i in range(len(ax)):\n",
+ " ax[i].plot(np.arange(1,17), pe[i,:16], color=f'C{i}', marker=\"o\", markersize=6, markeredgecolor=\"black\")\n",
+ " ax[i].set_title(f\"Encoding in hidden dimension {i+1}\")\n",
+ " ax[i].set_xlabel(\"Position in sequence\", fontsize=10)\n",
+ " ax[i].set_ylabel(\"Positional encoding\", fontsize=10)\n",
+ " ax[i].set_xticks(np.arange(1,17))\n",
+ " ax[i].tick_params(axis='both', which='major', labelsize=10)\n",
+ " ax[i].tick_params(axis='both', which='minor', labelsize=8)\n",
+ " ax[i].set_ylim(-1.2, 1.2)\n",
+ "fig.subplots_adjust(hspace=0.8)\n",
+ "sns.reset_orig()\n",
+ "plt.show()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "As we can see, the patterns between the hidden dimension $1$ and $2$ only differ in the starting angle. The wavelength is $2\\pi$, hence the repetition after position $6$. The hidden dimensions $2$ and $3$ have about twice the wavelength. "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Learning rate warm-up\n",
+ "\n",
+ "One commonly used technique for training a Transformer is learning rate warm-up. This means that we gradually increase the learning rate from 0 on to our originally specified learning rate in the first few iterations. Thus, we slowly start learning instead of taking very large steps from the beginning. In fact, training a deep Transformer without learning rate warm-up can make the model diverge and achieve a much worse performance on training and testing. Take for instance the following plot by [Liu et al. (2019)](https://arxiv.org/pdf/1908.03265.pdf) comparing Adam-vanilla (i.e. Adam without warm-up) vs Adam with a warm-up:\n",
+ "\n",
+ ":::{figure} ../image/warmup_loss_plot.svg\n",
+ ":::\n",
+ "\n",
+ "Clearly, the warm-up is a crucial hyperparameter in the Transformer architecture. Why is it so important? There are currently two common explanations. Firstly, Adam uses the bias correction factors which however can lead to a higher variance in the adaptive learning rate during the first iterations. Improved optimizers like [RAdam](https://arxiv.org/abs/1908.03265) have been shown to overcome this issue, not requiring warm-up for training Transformers. Secondly, the iteratively applied Layer Normalization across layers can lead to very high gradients during the first iterations, which can be solved by using [Pre-Layer Normalization](https://proceedings.icml.cc/static/paper_files/icml/2020/328-Paper.pdf) (similar to Pre-Activation ResNet), or replacing Layer Normalization by other techniques ([Adaptive Normalization](https://proceedings.icml.cc/static/paper_files/icml/2020/328-Paper.pdf), [Power Normalization](https://arxiv.org/abs/2003.07845)). \n",
+ "\n",
+ "Nevertheless, many applications and papers still use the original Transformer architecture with Adam, because warm-up is a simple, yet effective way of solving the gradient problem in the first iterations. There are many different schedulers we could use. For instance, the original Transformer paper used an exponential decay scheduler with a warm-up. However, the currently most popular scheduler is the cosine warm-up scheduler, which combines warm-up with a cosine-shaped learning rate decay. We can implement it below, and visualize the learning rate factor over epochs. "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 11,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class CosineWarmupScheduler(optim.lr_scheduler._LRScheduler):\n",
+ " \n",
+ " def __init__(self, optimizer, warmup, max_iters):\n",
+ " self.warmup = warmup\n",
+ " self.max_num_iters = max_iters\n",
+ " super().__init__(optimizer)\n",
+ " \n",
+ " def get_lr(self):\n",
+ " lr_factor = self.get_lr_factor(epoch=self.last_epoch)\n",
+ " return [base_lr * lr_factor for base_lr in self.base_lrs]\n",
+ " \n",
+ " def get_lr_factor(self, epoch):\n",
+ " lr_factor = 0.5 * (1 + np.cos(np.pi * epoch / self.max_num_iters))\n",
+ " if epoch <= self.warmup:\n",
+ " lr_factor *= epoch * 1.0 / self.warmup\n",
+ " return lr_factor"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 12,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "application/pdf": "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\n",
+ "image/svg+xml": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 2020-11-09T10:43:20.987233 \n",
+ " image/svg+xml \n",
+ " \n",
+ " \n",
+ " Matplotlib v3.3.2, https://matplotlib.org/ \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "# Needed for initializing the lr scheduler\n",
+ "p = nn.Parameter(torch.empty(4,4))\n",
+ "optimizer = optim.Adam([p], lr=1e-3)\n",
+ "lr_scheduler = CosineWarmupScheduler(optimizer=optimizer, warmup=100, max_iters=2000)\n",
+ "\n",
+ "# Plotting\n",
+ "epochs = list(range(2000))\n",
+ "sns.set()\n",
+ "plt.figure(figsize=(8,3))\n",
+ "plt.plot(epochs, [lr_scheduler.get_lr_factor(e) for e in epochs])\n",
+ "plt.ylabel(\"Learning rate factor\")\n",
+ "plt.xlabel(\"Iterations (in batches)\")\n",
+ "plt.title(\"Cosine Warm-up Learning Rate Scheduler\")\n",
+ "plt.show()\n",
+ "sns.reset_orig()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "In the first 100 iterations, we increase the learning rate factor from 0 to 1, whereas for all later iterations, we decay it using the cosine wave. Pre-implementations of this scheduler can be found in the popular NLP Transformer library [huggingface](https://huggingface.co/transformers/main_classes/optimizer_schedules.html?highlight=cosine#transformers.get_cosine_schedule_with_warmup)."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### PyTorch Lightning Module\n",
+ "\n",
+ "Finally, we can embed the Transformer architecture into a PyTorch lightning module. From Tutorial 5, you know that PyTorch Lightning simplifies our training and test code, as well as structures the code nicely in separate functions. We will implement a template for a classifier based on the Transformer encoder. Thereby, we have a prediction output per sequence element. If we would need a classifier over the whole sequence, the common approach is to add an additional `[CLS]` token to the sequence, representing the classifier token. However, here we focus on tasks where we have an output per element. \n",
+ "\n",
+ "Additionally to the Transformer architecture, we add a small input network (maps input dimensions to model dimensions), the positional encoding, and an output network (transforms output encodings to predictions). We also add the learning rate scheduler, which takes a step each iteration instead of once per epoch. This is needed for the warmup and the smooth cosine decay. The training, validation, and test step is left empty for now and will be filled for our task-specific models."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 13,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class TransformerPredictor(pl.LightningModule):\n",
+ "\n",
+ " def __init__(self, input_dim, model_dim, num_classes, num_heads, num_layers, lr, warmup, max_iters, dropout=0.0, input_dropout=0.0):\n",
+ " \"\"\"\n",
+ " Inputs:\n",
+ " input_dim - Hidden dimensionality of the input\n",
+ " model_dim - Hidden dimensionality to use inside the Transformer\n",
+ " num_classes - Number of classes to predict per sequence element\n",
+ " num_heads - Number of heads to use in the Multi-Head Attention blocks\n",
+ " num_layers - Number of encoder blocks to use.\n",
+ " lr - Learning rate in the optimizer\n",
+ " warmup - Number of warmup steps. Usually between 50 and 500\n",
+ " max_iters - Number of maximum iterations the model is trained for. This is needed for the CosineWarmup scheduler\n",
+ " dropout - Dropout to apply inside the model\n",
+ " input_dropout - Dropout to apply on the input features\n",
+ " \"\"\"\n",
+ " super().__init__()\n",
+ " self.save_hyperparameters()\n",
+ " self._create_model()\n",
+ "\n",
+ " def _create_model(self):\n",
+ " # Input dim -> Model dim\n",
+ " self.input_net = nn.Sequential(\n",
+ " nn.Dropout(self.hparams.input_dropout),\n",
+ " nn.Linear(self.hparams.input_dim, self.hparams.model_dim)\n",
+ " )\n",
+ " # Positional encoding for sequences\n",
+ " self.positional_encoding = PositionalEncoding(d_model=self.hparams.model_dim)\n",
+ " # Transformer\n",
+ " self.transformer = TransformerEncoder(num_layers=self.hparams.num_layers,\n",
+ " input_dim=self.hparams.model_dim,\n",
+ " dim_feedforward=2*self.hparams.model_dim,\n",
+ " num_heads=self.hparams.num_heads,\n",
+ " dropout=self.hparams.dropout)\n",
+ " # Output classifier per sequence lement\n",
+ " self.output_net = nn.Sequential(\n",
+ " nn.Linear(self.hparams.model_dim, self.hparams.model_dim),\n",
+ " nn.LayerNorm(self.hparams.model_dim),\n",
+ " nn.ReLU(inplace=True),\n",
+ " nn.Dropout(self.hparams.dropout),\n",
+ " nn.Linear(self.hparams.model_dim, self.hparams.num_classes)\n",
+ " ) \n",
+ "\n",
+ " def forward(self, x, mask=None, add_positional_encoding=True):\n",
+ " \"\"\"\n",
+ " Inputs:\n",
+ " x - Input features of shape [Batch, SeqLen, input_dim]\n",
+ " mask - Mask to apply on the attention outputs (optional)\n",
+ " add_positional_encoding - If True, we add the positional encoding to the input.\n",
+ " Might not be desired for some tasks.\n",
+ " \"\"\"\n",
+ " x = self.input_net(x)\n",
+ " if add_positional_encoding:\n",
+ " x = self.positional_encoding(x)\n",
+ " x = self.transformer(x, mask=mask)\n",
+ " x = self.output_net(x)\n",
+ " return x\n",
+ "\n",
+ " @torch.no_grad()\n",
+ " def get_attention_maps(self, x, mask=None, add_positional_encoding=True):\n",
+ " \"\"\"\n",
+ " Function for extracting the attention matrices of the whole Transformer for a single batch.\n",
+ " Input arguments same as the forward pass.\n",
+ " \"\"\"\n",
+ " x = self.input_net(x)\n",
+ " if add_positional_encoding:\n",
+ " x = self.positional_encoding(x)\n",
+ " attention_maps = self.transformer.get_attention_maps(x, mask=mask)\n",
+ " return attention_maps\n",
+ "\n",
+ " def configure_optimizers(self):\n",
+ " optimizer = optim.Adam(self.parameters(), lr=self.hparams.lr)\n",
+ " \n",
+ " # Apply lr scheduler per step\n",
+ " lr_scheduler = CosineWarmupScheduler(optimizer, \n",
+ " warmup=self.hparams.warmup, \n",
+ " max_iters=self.hparams.max_iters)\n",
+ " return [optimizer], [{'scheduler': lr_scheduler, 'interval': 'step'}]\n",
+ "\n",
+ " def training_step(self, batch, batch_idx):\n",
+ " raise NotImplementedError\n",
+ "\n",
+ " def validation_step(self, batch, batch_idx):\n",
+ " raise NotImplementedError \n",
+ "\n",
+ " def test_step(self, batch, batch_idx):\n",
+ " raise NotImplementedError "
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Experiments\n",
+ "\n",
+ "After having finished the implementation of the Transformer architecture, we can start experimenting and apply it to various tasks. In this notebook, we will focus on two tasks: parallel Sequence-to-Sequence, and set anomaly detection. The two tasks focus on different properties of the Transformer architecture, and we go through them below.\n",
+ "\n",
+ "### Sequence to Sequence\n",
+ "\n",
+ "A Sequence-to-Sequence task represents a task where the input _and_ the output is a sequence, not necessarily of the same length. Popular tasks in this domain include machine translation and summarization. For this, we usually have a Transformer encoder for interpreting the input sequence, and a decoder for generating the output in an autoregressive manner. Here, however, we will go back to a much simpler example task and use only the encoder. Given a sequence of $N$ numbers between $0$ and $M$, the task is to reverse the input sequence. In Numpy notation, if our input is $x$, the output should be $x$[::-1]. Although this task sounds very simple, RNNs can have issues with such because the task requires long-term dependencies. Transformers are built to support such, and hence, we expect it to perform very well. \n",
+ "\n",
+ "First, let's create a dataset class below."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 14,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class ReverseDataset(data.Dataset):\n",
+ "\n",
+ " def __init__(self, num_categories, seq_len, size):\n",
+ " super().__init__()\n",
+ " self.num_categories = num_categories\n",
+ " self.seq_len = seq_len\n",
+ " self.size = size\n",
+ " \n",
+ " self.data = torch.randint(self.num_categories, size=(self.size, self.seq_len))\n",
+ " \n",
+ " def __len__(self):\n",
+ " return self.size\n",
+ "\n",
+ " def __getitem__(self, idx):\n",
+ " inp_data = self.data[idx]\n",
+ " labels = torch.flip(inp_data, dims=(0,))\n",
+ " return inp_data, labels"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "We create an arbitrary number of random sequences of numbers between 0 and `num_categories-1`. The label is simply the tensor flipped over the sequence dimension. We can create the corresponding data loaders below. "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 15,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "dataset = partial(ReverseDataset, 10, 16)\n",
+ "train_loader = data.DataLoader(dataset(50000), batch_size=128, shuffle=True, drop_last=True, pin_memory=True)\n",
+ "val_loader = data.DataLoader(dataset(1000), batch_size=128)\n",
+ "test_loader = data.DataLoader(dataset(10000), batch_size=128)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Let's look at an arbitrary sample of the dataset:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 16,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Input data: tensor([9, 6, 2, 0, 6, 2, 7, 9, 7, 3, 3, 4, 3, 7, 0, 9])\n",
+ "Labels: tensor([9, 0, 7, 3, 4, 3, 3, 7, 9, 7, 2, 6, 0, 2, 6, 9])\n"
+ ]
+ }
+ ],
+ "source": [
+ "inp_data, labels = train_loader.dataset[0]\n",
+ "print(\"Input data:\", inp_data)\n",
+ "print(\"Labels: \", labels)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "During training, we pass the input sequence through the Transformer encoder and predict the output for each input token. We use the standard Cross-Entropy loss to perform this. Every number is represented as a one-hot vector. Remember that representing the categories as single scalars decreases the expressiveness of the model extremely as $0$ and $1$ are not closer related than $0$ and $9$ in our example. An alternative to a one-hot vector is using a learned embedding vector as it is provided by the PyTorch module `nn.Embedding`. However, using a one-hot vector with an additional linear layer as in our case has the same effect as an embedding layer (`self.input_net` maps one-hot vector to a dense vector, where each row of the weight matrix represents the embedding for a specific category).\n",
+ "\n",
+ "To implement the training dynamic, we create a new class inheriting from `TransformerPredictor` and overwriting the training, validation and test step functions."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 17,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class ReversePredictor(TransformerPredictor):\n",
+ " \n",
+ " def _calculate_loss(self, batch, mode=\"train\"):\n",
+ " # Fetch data and transform categories to one-hot vectors\n",
+ " inp_data, labels = batch\n",
+ " inp_data = F.one_hot(inp_data, num_classes=self.hparams.num_classes).float()\n",
+ " \n",
+ " # Perform prediction and calculate loss and accuracy\n",
+ " preds = self.forward(inp_data, add_positional_encoding=True)\n",
+ " loss = F.cross_entropy(preds.view(-1,preds.size(-1)), labels.view(-1))\n",
+ " acc = (preds.argmax(dim=-1) == labels).float().mean()\n",
+ " \n",
+ " # Logging\n",
+ " self.log(f\"{mode}_loss\", loss)\n",
+ " self.log(f\"{mode}_acc\", acc)\n",
+ " return loss, acc\n",
+ " \n",
+ " def training_step(self, batch, batch_idx):\n",
+ " loss, _ = self._calculate_loss(batch, mode=\"train\")\n",
+ " return loss\n",
+ " \n",
+ " def validation_step(self, batch, batch_idx):\n",
+ " _ = self._calculate_loss(batch, mode=\"val\")\n",
+ " \n",
+ " def test_step(self, batch, batch_idx):\n",
+ " _ = self._calculate_loss(batch, mode=\"test\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Finally, we can create a training function similar to the one we have seen in Tutorial 5 for PyTorch Lightning. We create a `pl.Trainer` object, running for $N$ epochs, logging in TensorBoard, and saving our best model based on the validation. Afterward, we test our models on the test set. An additional parameter we pass to the trainer here is `gradient_clip_val`. This clips the norm of the gradients for all parameters before taking an optimizer step and prevents the model from diverging if we obtain very high gradients at, for instance, sharp loss surfaces (see many good blog posts on gradient clipping, like [DeepAI glossary](https://deepai.org/machine-learning-glossary-and-terms/gradient-clipping)). For Transformers, gradient clipping can help to further stabilize the training during the first few iterations, and also afterward. In plain PyTorch, you can apply gradient clipping via `torch.nn.utils.clip_grad_norm_(...)` (see [documentation](https://pytorch.org/docs/stable/generated/torch.nn.utils.clip_grad_norm_.html#torch.nn.utils.clip_grad_norm_)). The clip value is usually between 0.5 and 10, depending on how harsh you want to clip large gradients. After having explained this, let's implement the training function:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 18,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def train_reverse(**kwargs):\n",
+ " # Create a PyTorch Lightning trainer with the generation callback\n",
+ " root_dir = os.path.join(CHECKPOINT_PATH, \"ReverseTask\")\n",
+ " os.makedirs(root_dir, exist_ok=True)\n",
+ " trainer = pl.Trainer(default_root_dir=root_dir, \n",
+ " callbacks=[ModelCheckpoint(save_weights_only=True, mode=\"max\", monitor=\"val_acc\")],\n",
+ " accelerator=\"gpu\" if str(device).startswith(\"cuda\") else \"cpu\",\n",
+ " devices=1,\n",
+ " max_epochs=10,\n",
+ " gradient_clip_val=5)\n",
+ " trainer.logger._default_hp_metric = None # Optional logging argument that we don't need\n",
+ " \n",
+ " # Check whether pretrained model exists. If yes, load it and skip training\n",
+ " pretrained_filename = os.path.join(CHECKPOINT_PATH, \"ReverseTask.ckpt\")\n",
+ " if os.path.isfile(pretrained_filename):\n",
+ " print(\"Found pretrained model, loading...\")\n",
+ " model = ReversePredictor.load_from_checkpoint(pretrained_filename)\n",
+ " else:\n",
+ " model = ReversePredictor(max_iters=trainer.max_epochs*len(train_loader), **kwargs)\n",
+ " trainer.fit(model, train_loader, val_loader)\n",
+ " \n",
+ " # Test best model on validation and test set\n",
+ " val_result = trainer.test(model, val_loader, verbose=False)\n",
+ " test_result = trainer.test(model, test_loader, verbose=False)\n",
+ " result = {\"test_acc\": test_result[0][\"test_acc\"], \"val_acc\": val_result[0][\"test_acc\"]}\n",
+ " \n",
+ " model = model.to(device)\n",
+ " return model, result"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Finally, we can train the model. In this setup, we will use a single encoder block and a single head in the Multi-Head Attention. This is chosen because of the simplicity of the task, and in this case, the attention can actually be interpreted as an \"explanation\" of the predictions (compared to the other papers above dealing with deep Transformers). "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def scaled_dot_product(q, k, v, mask=None):\n",
+ " d_k = q.size()[-1]\n",
+ " attn_logits = torch.matmul(q, k.transpose(-2, -1))\n",
+ " attn_logits = attn_logits / math.sqrt(d_k)\n",
+ " if mask is not None:\n",
+ " attn_logits = attn_logits.masked_fill(mask == 0, -9e15)\n",
+ " attention = F.softmax(attn_logits, dim=-1)\n",
+ " values = torch.matmul(attention, v)\n",
+ " return values, attention\n",
+ "\n",
+ "class MultiheadAttention(nn.Module):\n",
+ " \n",
+ " def __init__(self, input_dim, embed_dim, num_heads):\n",
+ " super().__init__()\n",
+ " assert embed_dim % num_heads == 0, \"Embedding dimension must be 0 modulo number of heads.\"\n",
+ " \n",
+ " self.embed_dim = embed_dim\n",
+ " self.num_heads = num_heads\n",
+ " self.head_dim = embed_dim // num_heads\n",
+ " \n",
+ " # Stack all weight matrices 1...h together for efficiency\n",
+ " # Note that in many implementations you see \"bias=False\" which is optional\n",
+ " self.qkv_proj = nn.Linear(input_dim, 3*embed_dim)\n",
+ " self.o_proj = nn.Linear(embed_dim, embed_dim)\n",
+ " \n",
+ " self._reset_parameters()\n",
+ "\n",
+ " def _reset_parameters(self):\n",
+ " # Original Transformer initialization, see PyTorch documentation\n",
+ " nn.init.xavier_uniform_(self.qkv_proj.weight)\n",
+ " self.qkv_proj.bias.data.fill_(0)\n",
+ " nn.init.xavier_uniform_(self.o_proj.weight)\n",
+ " self.o_proj.bias.data.fill_(0)\n",
+ "\n",
+ " def forward(self, x, mask=None, return_attention=False):\n",
+ " batch_size, seq_length, _ = x.size()\n",
+ " if mask is not None:\n",
+ " mask = expand_mask(mask)\n",
+ " qkv = self.qkv_proj(x)\n",
+ " \n",
+ " # Separate Q, K, V from linear output\n",
+ " qkv = qkv.reshape(batch_size, seq_length, self.num_heads, 3*self.head_dim)\n",
+ " qkv = qkv.permute(0, 2, 1, 3) # [Batch, Head, SeqLen, Dims]\n",
+ " q, k, v = qkv.chunk(3, dim=-1)\n",
+ " \n",
+ " # Determine value outputs\n",
+ " values, attention = scaled_dot_product(q, k, v, mask=mask)\n",
+ " values = values.permute(0, 2, 1, 3) # [Batch, SeqLen, Head, Dims]\n",
+ " values = values.reshape(batch_size, seq_length, self.embed_dim)\n",
+ " o = self.o_proj(values)\n",
+ " \n",
+ " if return_attention:\n",
+ " return o, attention\n",
+ " else:\n",
+ " return o"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 19,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "GPU available: True, used: True\n",
+ "TPU available: False, using: 0 TPU cores\n",
+ "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n"
+ ]
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Found pretrained model, loading...\n"
+ ]
+ }
+ ],
+ "source": [
+ "reverse_model, reverse_result = train_reverse(input_dim=train_loader.dataset.num_categories,\n",
+ " model_dim=32,\n",
+ " num_heads=1,\n",
+ " num_classes=train_loader.dataset.num_categories,\n",
+ " num_layers=1,\n",
+ " dropout=0.0,\n",
+ " lr=5e-4,\n",
+ " warmup=50)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "The warning of PyTorch Lightning regarding the number of workers can be ignored for now. As the data set is so simple and the `__getitem__` finishes a neglectable time, we don't need subprocesses to provide us the data (in fact, more workers can slow down the training as we have communication overhead among processes/threads). First, let's print the results:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 20,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Val accuracy: 100.00%\n",
+ "Test accuracy: 100.00%\n"
+ ]
+ }
+ ],
+ "source": [
+ "print(f\"Val accuracy: {(100.0 * reverse_result['val_acc']):4.2f}%\")\n",
+ "print(f\"Test accuracy: {(100.0 * reverse_result['test_acc']):4.2f}%\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "As we would have expected, the Transformer can correctly solve the task. However, how does the attention in the Multi-Head Attention block looks like for an arbitrary input? Let's try to visualize it below."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 21,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "data_input, labels = next(iter(val_loader))\n",
+ "inp_data = F.one_hot(data_input, num_classes=reverse_model.hparams.num_classes).float()\n",
+ "inp_data = inp_data.to(device)\n",
+ "attention_maps = reverse_model.get_attention_maps(inp_data)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "The object `attention_maps` is a list of length $N$ where $N$ is the number of layers. Each element is a tensor of shape [Batch, Heads, SeqLen, SeqLen], which we can verify below."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 22,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/plain": [
+ "torch.Size([128, 1, 16, 16])"
+ ]
+ },
+ "execution_count": 22,
+ "metadata": {},
+ "output_type": "execute_result"
+ }
+ ],
+ "source": [
+ "attention_maps[0].shape"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Next, we will write a plotting function that takes as input the sequences, attention maps, and an index indicating for which batch element we want to visualize the attention map. We will create a plot where over rows, we have different layers, while over columns, we show the different heads. Remember that the softmax has been applied for each row separately."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 23,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def plot_attention_maps(input_data, attn_maps, idx=0):\n",
+ " if input_data is not None:\n",
+ " input_data = input_data[idx].detach().cpu().numpy()\n",
+ " else:\n",
+ " input_data = np.arange(attn_maps[0][idx].shape[-1])\n",
+ " attn_maps = [m[idx].detach().cpu().numpy() for m in attn_maps]\n",
+ " \n",
+ " num_heads = attn_maps[0].shape[0]\n",
+ " num_layers = len(attn_maps)\n",
+ " seq_len = input_data.shape[0]\n",
+ " fig_size = 4 if num_heads == 1 else 3\n",
+ " fig, ax = plt.subplots(num_layers, num_heads, figsize=(num_heads*fig_size, num_layers*fig_size))\n",
+ " if num_layers == 1:\n",
+ " ax = [ax]\n",
+ " if num_heads == 1:\n",
+ " ax = [[a] for a in ax]\n",
+ " for row in range(num_layers):\n",
+ " for column in range(num_heads):\n",
+ " ax[row][column].imshow(attn_maps[row][column], origin='lower', vmin=0)\n",
+ " ax[row][column].set_xticks(list(range(seq_len)))\n",
+ " ax[row][column].set_xticklabels(input_data.tolist())\n",
+ " ax[row][column].set_yticks(list(range(seq_len)))\n",
+ " ax[row][column].set_yticklabels(input_data.tolist())\n",
+ " ax[row][column].set_title(f\"Layer {row+1}, Head {column+1}\")\n",
+ " fig.subplots_adjust(hspace=0.5)\n",
+ " plt.show()"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Finally, we can plot the attention map of our trained Transformer on the reverse task:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 24,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "application/pdf": "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\n",
+ "image/svg+xml": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 2020-11-09T10:43:26.716937 \n",
+ " image/svg+xml \n",
+ " \n",
+ " \n",
+ " Matplotlib v3.3.2, https://matplotlib.org/ \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "plot_attention_maps(data_input, attention_maps, idx=0)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "The model has learned to attend to the token that is on the flipped index of itself. Hence, it actually does what we intended it to do. We see that it however also pays some attention to values close to the flipped index. This is because the model doesn't need the perfect, hard attention to solve this problem, but is fine with this approximate, noisy attention map. The close-by indices are caused by the similarity of the positional encoding, which we also intended with the positional encoding."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Set Anomaly Detection\n",
+ "\n",
+ "Besides sequences, sets are another data structure that is relevant for many applications. In contrast to sequences, elements are unordered in a set. RNNs can only be applied on sets by assuming an order in the data, which however biases the model towards a non-existing order in the data. [Vinyals et al. (2015)](https://arxiv.org/abs/1511.06391) and other papers have shown that the assumed order can have a significant impact on the model's performance, and hence, we should try to not use RNNs on sets. Ideally, our model should be permutation-equivariant/invariant such that the output is the same no matter how we sort the elements in a set. \n",
+ "\n",
+ "Transformers offer the perfect architecture for this as the Multi-Head Attention is permutation-equivariant, and thus, outputs the same values no matter in what order we enter the inputs (inputs and outputs are permuted equally). The task we are looking at for sets is _Set Anomaly Detection_ which means that we try to find the element(s) in a set that does not fit the others. In the research community, the common application of anomaly detection is performed on a set of images, where $N-1$ images belong to the same category/have the same high-level features while one belongs to another category. Note that category does not necessarily have to relate to a class in a standard classification problem, but could be the combination of multiple features. For instance, on a face dataset, this could be people with glasses, male, beard, etc. An example of distinguishing different animals can be seen below. The first four images show foxes, while the last represents a different animal. We want to recognize that the last image shows a different animal, but it is not relevant which class of animal it is.\n",
+ "\n",
+ " \n",
+ "\n",
+ ":::{figure} ../image/warmup_loss_plot.svg\n",
+ ":::\n",
+ "\n",
+ "In this tutorial, we will use the CIFAR100 dataset. CIFAR100 has 600 images for 100 classes each with a resolution of 32x32, similar to CIFAR10. The larger amount of classes requires the model to attend to specific features in the images instead of coarse features as in CIFAR10, therefore making the task harder. We will show the model a set of 9 images of one class, and 1 image from another class. The task is to find the image that is from a different class than the other images.\n",
+ "Using the raw images directly as input to the Transformer is not a good idea, because it is not translation invariant as a CNN, and would need to learn to detect image features from high-dimensional input first of all. Instead, we will use a pre-trained ResNet34 model from the torchvision package to obtain high-level, low-dimensional features of the images. The ResNet model has been pre-trained on the [ImageNet](http://image-net.org/) dataset which contains 1 million images of 1k classes and varying resolutions. However, during training and testing, the images are usually scaled to a resolution of 224x224, and hence we rescale our CIFAR images to this resolution as well. Below, we will load the dataset, and prepare the data for being processed by the ResNet model."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 25,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Files already downloaded and verified\n",
+ "Files already downloaded and verified\n"
+ ]
+ }
+ ],
+ "source": [
+ "# ImageNet statistics\n",
+ "DATA_MEANS = np.array([0.485, 0.456, 0.406])\n",
+ "DATA_STD = np.array([0.229, 0.224, 0.225])\n",
+ "# As torch tensors for later preprocessing\n",
+ "TORCH_DATA_MEANS = torch.from_numpy(DATA_MEANS).view(1,3,1,1)\n",
+ "TORCH_DATA_STD = torch.from_numpy(DATA_STD).view(1,3,1,1)\n",
+ "\n",
+ "# Resize to 224x224, and normalize to ImageNet statistic\n",
+ "transform = transforms.Compose([transforms.Resize((224,224)),\n",
+ " transforms.ToTensor(),\n",
+ " transforms.Normalize(DATA_MEANS, DATA_STD)\n",
+ " ])\n",
+ "# Loading the training dataset. \n",
+ "train_set = CIFAR100(root=DATASET_PATH, train=True, transform=transform, download=True)\n",
+ "\n",
+ "# Loading the test set\n",
+ "test_set = CIFAR100(root=DATASET_PATH, train=False, transform=transform, download=True)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Next, we want to run the pre-trained ResNet model on the images, and extract the features before the classification layer. These are the most high-level features, and should sufficiently describe the images. CIFAR100 has some similarity to ImageNet, and thus we are not retraining the ResNet model in any form. However, if you would want to get the best performance and have a very large dataset, it would be better to add the ResNet to the computation graph during training and finetune its parameters as well. As we don't have a large enough dataset and want to train our model efficiently, we will extract the features beforehand. Let's load and prepare the model below."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 26,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "import os\n",
+ "os.environ[\"TORCH_HOME\"] = CHECKPOINT_PATH\n",
+ "pretrained_model = torchvision.models.resnet34(weights='IMAGENET1K_V1')\n",
+ "# Remove classification layer\n",
+ "# In some models, it is called \"fc\", others have \"classifier\"\n",
+ "# Setting both to an empty sequential represents an identity map of the final features.\n",
+ "pretrained_model.fc = nn.Sequential()\n",
+ "pretrained_model.classifier = nn.Sequential()\n",
+ "# To GPU\n",
+ "pretrained_model = pretrained_model.to(device)\n",
+ "\n",
+ "# Only eval, no gradient required\n",
+ "pretrained_model.eval()\n",
+ "for p in pretrained_model.parameters():\n",
+ " p.requires_grad = False"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "We will now write a extraction function for the features below. This cell requires access to a GPU, as the model is rather deep and the images relatively large. The GPUs on GoogleColab are sufficient, but running this cell can take 2-3 minutes. Once it is run, the features are exported on disk so they don't have to be recalculated every time you run the notebook. However, this requires >150MB free disk space. So it is recommended to run this only on a local computer if you have enough free disk and a GPU (GoogleColab is fine for this). If you do not have a GPU, you can download the features from the [GoogleDrive folder](https://drive.google.com/drive/folders/1DF7POc6j03pRiWQPWSl5QJX5iY-xK0sV?usp=sharing)."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 27,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "@torch.no_grad()\n",
+ "def extract_features(dataset, save_file):\n",
+ " if not os.path.isfile(save_file):\n",
+ " data_loader = data.DataLoader(dataset, batch_size=128, shuffle=False, drop_last=False, num_workers=4)\n",
+ " extracted_features = []\n",
+ " for imgs, _ in tqdm(data_loader):\n",
+ " imgs = imgs.to(device)\n",
+ " feats = pretrained_model(imgs)\n",
+ " extracted_features.append(feats)\n",
+ " extracted_features = torch.cat(extracted_features, dim=0)\n",
+ " extracted_features = extracted_features.detach().cpu()\n",
+ " torch.save(extracted_features, save_file)\n",
+ " else:\n",
+ " extracted_features = torch.load(save_file)\n",
+ " return extracted_features\n",
+ "\n",
+ "train_feat_file = os.path.join(CHECKPOINT_PATH, \"train_set_features.tar\")\n",
+ "train_set_feats = extract_features(train_set, train_feat_file)\n",
+ "\n",
+ "test_feat_file = os.path.join(CHECKPOINT_PATH, \"test_set_features.tar\")\n",
+ "test_feats = extract_features(test_set, test_feat_file)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Let's verify the feature shapes below. The training should have 50k elements, and the test 10k images. The feature dimension is 512 for the ResNet34. If you experiment with other models, you likely see a different feature dimension."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 28,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Train: torch.Size([50000, 512])\n",
+ "Test: torch.Size([10000, 512])\n"
+ ]
+ }
+ ],
+ "source": [
+ "print(\"Train:\", train_set_feats.shape)\n",
+ "print(\"Test: \", test_feats.shape)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "As usual, we want to create a validation set to detect when we should stop training. In this case, we will split the training set into 90% training, 10% validation. However, the difficulty is here that we need to ensure that the validation set has the same number of images for all 100 labels. Otherwise, we have a class imbalance which is not good for creating the image sets. Hence, we take 10% of the images for each class, and move them into the validation set. The code below does exactly this."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 29,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "## Split train into train+val\n",
+ "# Get labels from train set\n",
+ "labels = train_set.targets\n",
+ "\n",
+ "# Get indices of images per class\n",
+ "labels = torch.LongTensor(labels)\n",
+ "num_labels = labels.max()+1\n",
+ "sorted_indices = torch.argsort(labels).reshape(num_labels, -1) # [classes, num_imgs per class]\n",
+ "\n",
+ "# Determine number of validation images per class\n",
+ "num_val_exmps = sorted_indices.shape[1] // 10\n",
+ "\n",
+ "# Get image indices for validation and training\n",
+ "val_indices = sorted_indices[:,:num_val_exmps].reshape(-1)\n",
+ "train_indices = sorted_indices[:,num_val_exmps:].reshape(-1)\n",
+ "\n",
+ "# Group corresponding image features and labels\n",
+ "train_feats, train_labels = train_set_feats[train_indices], labels[train_indices]\n",
+ "val_feats, val_labels = train_set_feats[val_indices], labels[val_indices]"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Now we can prepare a dataset class for the set anomaly task. We define an epoch to be the sequence in which each image has been exactly once as an \"anomaly\". Hence, the length of the dataset is the number of images in it. For the training set, each time we access an item with `__getitem__`, we sample a random, different class than the image at the corresponding index `idx` has. In a second step, we sample $N-1$ images of this sampled class. The set of 10 images is finally returned. The randomness in the `__getitem__` allows us to see a slightly different set during each iteration. However, we can't use the same strategy for the test set as we want the test dataset to be the same every time we iterate over it. Hence, we sample the sets in the `__init__` method, and return those in `__getitem__`. The code below implements exactly this dynamic."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 30,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class SetAnomalyDataset(data.Dataset):\n",
+ " \n",
+ " def __init__(self, img_feats, labels, set_size=10, train=True):\n",
+ " \"\"\"\n",
+ " Inputs:\n",
+ " img_feats - Tensor of shape [num_imgs, img_dim]. Represents the high-level features.\n",
+ " labels - Tensor of shape [num_imgs], containing the class labels for the images\n",
+ " set_size - Number of elements in a set. N-1 are sampled from one class, and one from another one.\n",
+ " train - If True, a new set will be sampled every time __getitem__ is called.\n",
+ " \"\"\"\n",
+ " super().__init__()\n",
+ " self.img_feats = img_feats\n",
+ " self.labels = labels\n",
+ " self.set_size = set_size-1 # The set size is here the size of correct images\n",
+ " self.train = train\n",
+ " \n",
+ " # Tensors with indices of the images per class\n",
+ " self.num_labels = labels.max()+1\n",
+ " self.img_idx_by_label = torch.argsort(self.labels).reshape(self.num_labels, -1)\n",
+ " \n",
+ " if not train:\n",
+ " self.test_sets = self._create_test_sets()\n",
+ " \n",
+ " \n",
+ " def _create_test_sets(self):\n",
+ " # Pre-generates the sets for each image for the test set\n",
+ " test_sets = []\n",
+ " num_imgs = self.img_feats.shape[0]\n",
+ " np.random.seed(42)\n",
+ " test_sets = [self.sample_img_set(self.labels[idx]) for idx in range(num_imgs)]\n",
+ " test_sets = torch.stack(test_sets, dim=0)\n",
+ " return test_sets\n",
+ " \n",
+ " \n",
+ " def sample_img_set(self, anomaly_label):\n",
+ " \"\"\"\n",
+ " Samples a new set of images, given the label of the anomaly. \n",
+ " The sampled images come from a different class than anomaly_label\n",
+ " \"\"\"\n",
+ " # Sample class from 0,...,num_classes-1 while skipping anomaly_label as class\n",
+ " set_label = np.random.randint(self.num_labels-1)\n",
+ " if set_label >= anomaly_label:\n",
+ " set_label += 1\n",
+ " \n",
+ " # Sample images from the class determined above\n",
+ " img_indices = np.random.choice(self.img_idx_by_label.shape[1], size=self.set_size, replace=False)\n",
+ " img_indices = self.img_idx_by_label[set_label, img_indices]\n",
+ " return img_indices\n",
+ " \n",
+ " \n",
+ " def __len__(self):\n",
+ " return self.img_feats.shape[0]\n",
+ " \n",
+ " \n",
+ " def __getitem__(self, idx):\n",
+ " anomaly = self.img_feats[idx]\n",
+ " if self.train: # If train => sample\n",
+ " img_indices = self.sample_img_set(self.labels[idx])\n",
+ " else: # If test => use pre-generated ones\n",
+ " img_indices = self.test_sets[idx]\n",
+ " \n",
+ " # Concatenate images. The anomaly is always the last image for simplicity\n",
+ " img_set = torch.cat([self.img_feats[img_indices], anomaly[None]], dim=0)\n",
+ " indices = torch.cat([img_indices, torch.LongTensor([idx])], dim=0)\n",
+ " label = img_set.shape[0]-1\n",
+ " \n",
+ " # We return the indices of the images for visualization purpose. \"Label\" is the index of the anomaly\n",
+ " return img_set, indices, label"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Next, we can setup our datasets and data loaders below. Here, we will use a set size of 10, i.e. 9 images from one category + 1 anomaly. Feel free to change it if you want to experiment with the sizes. "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 31,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "SET_SIZE = 10\n",
+ "test_labels = torch.LongTensor(test_set.targets)\n",
+ "\n",
+ "train_anom_dataset = SetAnomalyDataset(train_feats, train_labels, set_size=SET_SIZE, train=True)\n",
+ "val_anom_dataset = SetAnomalyDataset(val_feats, val_labels, set_size=SET_SIZE, train=False)\n",
+ "test_anom_dataset = SetAnomalyDataset(test_feats, test_labels, set_size=SET_SIZE, train=False)\n",
+ "\n",
+ "train_anom_loader = data.DataLoader(train_anom_dataset, batch_size=64, shuffle=True, drop_last=True, num_workers=4, pin_memory=True)\n",
+ "val_anom_loader = data.DataLoader(val_anom_dataset, batch_size=64, shuffle=False, drop_last=False, num_workers=4)\n",
+ "test_anom_loader = data.DataLoader(test_anom_dataset, batch_size=64, shuffle=False, drop_last=False, num_workers=4)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "To understand the dataset a little better, we can plot below a few sets from the test dataset. Each row shows a different input set, where the first 9 are from the same class."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 32,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "application/pdf": "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\n",
+ "image/svg+xml": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 2020-11-09T10:43:30.487860 \n",
+ " image/svg+xml \n",
+ " \n",
+ " \n",
+ " Matplotlib v3.3.2, https://matplotlib.org/ \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "def visualize_exmp(indices, orig_dataset):\n",
+ " images = [orig_dataset[idx][0] for idx in indices.reshape(-1)]\n",
+ " images = torch.stack(images, dim=0)\n",
+ " images = images * TORCH_DATA_STD + TORCH_DATA_MEANS\n",
+ " \n",
+ " img_grid = torchvision.utils.make_grid(images, nrow=SET_SIZE, normalize=True, pad_value=0.5, padding=16)\n",
+ " img_grid = img_grid.permute(1, 2, 0)\n",
+ "\n",
+ " plt.figure(figsize=(12,8))\n",
+ " plt.title(\"Anomaly examples on CIFAR100\")\n",
+ " plt.imshow(img_grid)\n",
+ " plt.axis('off')\n",
+ " plt.show()\n",
+ " plt.close()\n",
+ "\n",
+ "_, indices, _ = next(iter(test_anom_loader))\n",
+ "visualize_exmp(indices[:4], test_set)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "We can already see that for some sets the task might be easier than for others. Difficulties can especially arise if the anomaly is in a different, but yet visually similar class (e.g. train vs bus, flour vs worm, etc.).\n",
+ "\n",
+ "After having prepared the data, we can look closer at the model. Here, we have a classification of the whole set. For the prediction to be permutation-equivariant, we will output one logit for each image. Over these logits, we apply a softmax and train the anomaly image to have the highest score/probability. This is a bit different than a standard classification layer as the softmax is applied over images, not over output classes in the classical sense. However, if we swap two images in their position, we effectively swap their position in the output softmax. Hence, the prediction is equivariant with respect to the input. We implement this idea below in the subclass of the Transformer Lightning module."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 33,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "class AnomalyPredictor(TransformerPredictor):\n",
+ " \n",
+ " def _calculate_loss(self, batch, mode=\"train\"):\n",
+ " img_sets, _, labels = batch\n",
+ " preds = self.forward(img_sets, add_positional_encoding=False) # No positional encodings as it is a set, not a sequence!\n",
+ " preds = preds.squeeze(dim=-1) # Shape: [Batch_size, set_size]\n",
+ " loss = F.cross_entropy(preds, labels) # Softmax/CE over set dimension\n",
+ " acc = (preds.argmax(dim=-1) == labels).float().mean()\n",
+ " self.log(f\"{mode}_loss\", loss)\n",
+ " self.log(f\"{mode}_acc\", acc, on_step=False, on_epoch=True)\n",
+ " return loss, acc\n",
+ " \n",
+ " def training_step(self, batch, batch_idx):\n",
+ " loss, _ = self._calculate_loss(batch, mode=\"train\")\n",
+ " return loss\n",
+ " \n",
+ " def validation_step(self, batch, batch_idx):\n",
+ " _ = self._calculate_loss(batch, mode=\"val\")\n",
+ " \n",
+ " def test_step(self, batch, batch_idx):\n",
+ " _ = self._calculate_loss(batch, mode=\"test\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Finally, we write our train function below. It has the exact same structure as the reverse task one, hence not much of an explanation is needed here."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 34,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "def train_anomaly(**kwargs):\n",
+ " # Create a PyTorch Lightning trainer with the generation callback\n",
+ " root_dir = os.path.join(CHECKPOINT_PATH, \"SetAnomalyTask\")\n",
+ " os.makedirs(root_dir, exist_ok=True)\n",
+ " trainer = pl.Trainer(default_root_dir=root_dir, \n",
+ " callbacks=[ModelCheckpoint(save_weights_only=True, mode=\"max\", monitor=\"val_acc\")],\n",
+ " accelerator=\"gpu\" if str(device).startswith(\"cuda\") else \"cpu\",\n",
+ " devices=1,\n",
+ " max_epochs=100,\n",
+ " gradient_clip_val=2)\n",
+ " trainer.logger._default_hp_metric = None # Optional logging argument that we don't need\n",
+ " \n",
+ " # Check whether pretrained model exists. If yes, load it and skip training\n",
+ " pretrained_filename = os.path.join(CHECKPOINT_PATH, \"SetAnomalyTask.ckpt\")\n",
+ " if os.path.isfile(pretrained_filename):\n",
+ " print(\"Found pretrained model, loading...\")\n",
+ " model = AnomalyPredictor.load_from_checkpoint(pretrained_filename)\n",
+ " else:\n",
+ " model = AnomalyPredictor(max_iters=trainer.max_epochs*len(train_anom_loader), **kwargs)\n",
+ " trainer.fit(model, train_anom_loader, val_anom_loader)\n",
+ " model = AnomalyPredictor.load_from_checkpoint(trainer.checkpoint_callback.best_model_path)\n",
+ " \n",
+ " # Test best model on validation and test set\n",
+ " train_result = trainer.test(model, train_anom_loader, verbose=False)\n",
+ " val_result = trainer.test(model, val_anom_loader, verbose=False)\n",
+ " test_result = trainer.test(model, test_anom_loader, verbose=False)\n",
+ " result = {\"test_acc\": test_result[0][\"test_acc\"], \"val_acc\": val_result[0][\"test_acc\"], \"train_acc\": train_result[0][\"test_acc\"]}\n",
+ " \n",
+ " model = model.to(device)\n",
+ " return model, result"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Let's finally train our model. We will use 4 layers with 4 attention heads each. The hidden dimensionality of the model is 256, and we use a dropout of 0.1 throughout the model for good regularization. Note that we also apply the dropout on the input features, as this makes the model more robust against image noise and generalizes better. Again, we use warmup to slowly start our model training. "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 35,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stderr",
+ "output_type": "stream",
+ "text": [
+ "GPU available: True, used: True\n",
+ "WARNING: Logging before flag parsing goes to stderr.\n",
+ "I1109 10:43:31.036801 139648634296128 distributed.py:49] GPU available: True, used: True\n",
+ "TPU available: False, using: 0 TPU cores\n",
+ "I1109 10:43:31.038146 139648634296128 distributed.py:49] TPU available: False, using: 0 TPU cores\n",
+ "LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n",
+ "I1109 10:43:31.039162 139648634296128 accelerator_connector.py:385] LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0]\n"
+ ]
+ }
+ ],
+ "source": [
+ "anomaly_model, anomaly_result = train_anomaly(input_dim=train_anom_dataset.img_feats.shape[-1],\n",
+ " model_dim=256,\n",
+ " num_heads=4,\n",
+ " num_classes=1,\n",
+ " num_layers=4,\n",
+ " dropout=0.1,\n",
+ " input_dropout=0.1,\n",
+ " lr=5e-4,\n",
+ " warmup=100)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "We can print the achieved accuracy below."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 36,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Train accuracy: 97.77%\n",
+ "Val accuracy: 94.38%\n",
+ "Test accuracy: 94.30%\n"
+ ]
+ }
+ ],
+ "source": [
+ "print(f\"Train accuracy: {(100.0*anomaly_result['train_acc']):4.2f}%\")\n",
+ "print(f\"Val accuracy: {(100.0*anomaly_result['val_acc']):4.2f}%\")\n",
+ "print(f\"Test accuracy: {(100.0*anomaly_result['test_acc']):4.2f}%\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "With ~94% validation and test accuracy, the model generalizes quite well. It should be noted that you might see slightly different scores depending on what computer/device you are running this notebook. This is because despite setting the seed before generating the test dataset, it is not the same across platforms and numpy versions. Nevertheless, we can conclude that the model performs quite well and can solve the task for most sets. Before trying to interpret the model, let's verify that our model is permutation-equivariant, and assigns the same predictions for different permutations of the input set. For this, we sample a batch from the test set and run it through the model to obtain the probabilities. "
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 37,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Preds\n",
+ " [5.4543594e-05 1.4208173e-04 6.6922468e-05 7.6413504e-05 7.7112330e-05\n",
+ " 8.7848457e-05 6.6820685e-05 9.9929154e-01 7.3219831e-05 6.3545609e-05]\n",
+ "Permuted preds\n",
+ " [5.4543532e-05 1.4208158e-04 6.6922395e-05 7.6413417e-05 7.7112243e-05\n",
+ " 8.7848362e-05 6.6820678e-05 9.9929142e-01 7.3219751e-05 6.3545544e-05]\n"
+ ]
+ }
+ ],
+ "source": [
+ "inp_data, indices, labels = next(iter(test_anom_loader))\n",
+ "inp_data = inp_data.to(device)\n",
+ "\n",
+ "anomaly_model.eval()\n",
+ "\n",
+ "with torch.no_grad():\n",
+ " preds = anomaly_model.forward(inp_data, add_positional_encoding=False)\n",
+ " preds = F.softmax(preds.squeeze(dim=-1), dim=-1)\n",
+ "\n",
+ " # Permut input data\n",
+ " permut = np.random.permutation(inp_data.shape[1])\n",
+ " perm_inp_data = inp_data[:,permut]\n",
+ " perm_preds = anomaly_model.forward(perm_inp_data, add_positional_encoding=False)\n",
+ " perm_preds = F.softmax(perm_preds.squeeze(dim=-1), dim=-1)\n",
+ "\n",
+ "assert (preds[:,permut] - perm_preds).abs().max() < 1e-5, \"Predictions are not permutation equivariant\"\n",
+ "\n",
+ "print(\"Preds\\n\", preds[0,permut].cpu().numpy())\n",
+ "print(\"Permuted preds\\n\", perm_preds[0].cpu().numpy())"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "You can see that the predictions are almost exactly the same, and only differ because of slight numerical differences inside the network operation.\n",
+ "\n",
+ "To interpret the model a little more, we can plot the attention maps inside the model. This will give us an idea of what information the model is sharing/communicating between images, and what each head might represent. First, we need to extract the attention maps for the test batch above, and determine the discrete predictions for simplicity."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 38,
+ "metadata": {},
+ "outputs": [],
+ "source": [
+ "attention_maps = anomaly_model.get_attention_maps(inp_data, add_positional_encoding=False)\n",
+ "predictions = preds.argmax(dim=-1)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Below we write a plot function which plots the images in the input set, the prediction of the model, and the attention maps of the different heads on layers of the transformer. Feel free to explore the attention maps for different input examples as well."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 39,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "application/pdf": "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\n",
+ "image/svg+xml": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 2020-11-09T10:43:35.755092 \n",
+ " image/svg+xml \n",
+ " \n",
+ " \n",
+ " Matplotlib v3.3.2, https://matplotlib.org/ \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Prediction: 9\n"
+ ]
+ },
+ {
+ "data": {
+ "application/pdf": "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\n",
+ "image/svg+xml": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 2020-11-09T10:43:36.649271 \n",
+ " image/svg+xml \n",
+ " \n",
+ " \n",
+ " Matplotlib v3.3.2, https://matplotlib.org/ \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ }
+ ],
+ "source": [
+ "def visualize_prediction(idx):\n",
+ " visualize_exmp(indices[idx:idx+1], test_set)\n",
+ " print(\"Prediction:\", predictions[idx].item())\n",
+ " plot_attention_maps(input_data=None, attn_maps=attention_maps, idx=idx)\n",
+ "\n",
+ "visualize_prediction(0)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "Depending on the random seed, you might see a slightly different input set. For the version on the website, we compare 9 tree images with a volcano. We see that multiple heads, for instance, Layer 2 Head 1, Layer 2 Head 3, and Layer 3 Head 1 focus on the last image. Additionally, the heads in Layer 4 all seem to ignore the last image and assign a very low attention probability to it. This shows that the model has indeed recognized that the image doesn't fit the setting, and hence predicted it to be the anomaly. Layer 3 Head 2-4 seems to take a slightly weighted average of all images. That might indicate that the model extracts the \"average\" information of all images, to compare it to the image features itself. \n",
+ "\n",
+ "Let's try to find where the model actually makes a mistake. We can do this by identifying the sets where the model predicts something else than 9, as in the dataset, we ensured that the anomaly is always at the last position in the set."
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 40,
+ "metadata": {},
+ "outputs": [
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Indices with mistake: [36 49]\n"
+ ]
+ }
+ ],
+ "source": [
+ "mistakes = torch.where(predictions != 9)[0].cpu().numpy()\n",
+ "print(\"Indices with mistake:\", mistakes)"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "As our model achieves ~94% accuracy, we only have very little number of mistakes in a batch of 64 sets. Still, let's visualize one of them, for example the last one:"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": 41,
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "application/pdf": "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\n",
+ "image/svg+xml": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 2020-11-09T10:43:37.728184 \n",
+ " image/svg+xml \n",
+ " \n",
+ " \n",
+ " Matplotlib v3.3.2, https://matplotlib.org/ \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Prediction: 2\n"
+ ]
+ },
+ {
+ "data": {
+ "application/pdf": "JVBERi0xLjQKJazcIKu6CjEgMCBvYmoKPDwgL1BhZ2VzIDIgMCBSIC9UeXBlIC9DYXRhbG9nID4+CmVuZG9iago4IDAgb2JqCjw8IC9FeHRHU3RhdGUgNCAwIFIgL0ZvbnQgMyAwIFIgL1BhdHRlcm4gNSAwIFIKL1Byb2NTZXQgWyAvUERGIC9UZXh0IC9JbWFnZUIgL0ltYWdlQyAvSW1hZ2VJIF0gL1NoYWRpbmcgNiAwIFIKL1hPYmplY3QgNyAwIFIgPj4KZW5kb2JqCjEwIDAgb2JqCjw8IC9Bbm5vdHMgWyBdIC9Db250ZW50cyA5IDAgUgovR3JvdXAgPDwgL0NTIC9EZXZpY2VSR0IgL1MgL1RyYW5zcGFyZW5jeSAvVHlwZSAvR3JvdXAgPj4KL01lZGlhQm94IFsgMCAwIDY3MC4zOTc3OTM5NzIzIDY5OC41MTY4NzUgXSAvUGFyZW50IDIgMCBSIC9SZXNvdXJjZXMgOCAwIFIKL1R5cGUgL1BhZ2UgPj4KZW5kb2JqCjkgMCBvYmoKPDwgL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0xlbmd0aCAxMSAwIFIgPj4Kc3RyZWFtCnicvZ1djyvHdUXf+Sv4mABBq+urPx4t2FES5EXJBfxsyNeKBV0HsoEE+fc5Tc6w9j5TVZtNQxcXRqCTIRfJ6a5aTa6ZCdefLt/8Jlx//Nt1vv5k//vfa7h+d/3mt5//588/fP6P7769/vC3y2zzL5dlnae0r+ue7T9/xv9c9m0qYdnWYvOZ//O/Lpe/XOz+7Tbf2V3/eLnEeSplT/b/LWWdyrbFtNvdh7RPYUlzCDj/GefLukxhD/c7rndDY+P96fLLtQHZlrTkawjbtMxpuf/z//nXz9ffX/9y/eY38XhFgr04wV6R+cMr8ovdcL8er8vxf7uwH75cv/nXcP3tf1+/v3x//eX9fmd7OcLxak/b273b5BKXKW+73QO/LjjO0/z+sly+PR7a5dtP12/+OVzDfP30p0tMU5r3bdlT2OK15Djte87b/bF8+uPlH+Z/vH766fq7T5cb+ZK2KeV9jwyEqeClMoV1jinuJZcWLzCvzFMMc9wyA3EsiHmd5jSXfSn7vLeIkYlLnOZ1XvLGRBwLYtmmzVjLHOOaWsTExDVN9vK7o/gLjQXRzqu1hHQ/hlrEzMStTOsePpxPOBbELU5LuJ+wDVxh3L5OSwmr3RnhcCxwe57yGucSwoFoEBcmhnm39eQ4RNyKgXPBDPMypRRzyOsa1xZ0ddAYprTGvCwOinMFtcUl7HFdtzzvoQXdHDSlyU6nLfmlsY4VMs3TXFLIcc6l+e3cGQlr15Ls5dxyPIh2nvlp52idjld/n9YYbBE4/rVWGqCsxV6/zV4SxtTxgLMGe1GCnY7Hv9YKA5zNbpH2YOc4cep4wNnSlLOd8Me/1FpXgLPv07zvdjAzp44HnN2e9BzfN5/GagKbXJinvezHOYocGPc5y7weW2hr8QBAjNMW57cjrgLqeAAI9mxjuj+P5qoBnJSnZZttQWVOHQ84dnRvW1rv35nWWgGcvEzFzjs7UYlTxwOOnTZrzvfjLLSWB+DYDpFnO0UZ85gOKHYyL3O+nzWltR6g49gmuQTb2hlTxwPOYgphm124/eNFIF7/7S5+NxVhyemIWke8Lv/ZUbgvXYWzm5yUQboF3NeQMd+e41257jr34+NFijePiVMocyxlOV6fxbbDt5sfL9W//+H/Pv/1Gv7p+i+f//DH62O9eZPmu9nexflhuMHO65jW2/ZNZhXStOVgxy4/K5jTY4f7+ei4h08+vqCY9C2vqu7lrrrXZ1R3TLz5bnzWd+dg37H1eMnd8QVztd2ZduzbZruevTTNbd1vRPZSm2raieCgOFeabd6x5S3vaTOheMJ7Y8y2bG03SSYozhXUvGO1ryhrXubmxu63qLTYGrbdPJmgOJcXFLb5L7uxbcHcnrDfmI9jbHdH+BeeK2guU7ZLii2vZmxPCHBcZlvf9g+nG80V1G4YN+Ns6XZcSw22p2JC6YDvMwVb9mPBSSHOto4/YcB2bWRn/3HuOiDOFXQL02wavBbb3ZpXMn6L2xfbZuabLhMU5wq6mz2sYU72rd+a56fb8ZKpyh6CP2hxrC5K5zytJsBz3MMzAoyLNRrwbuebH3cWIXtlzVztcHtbE5UNI5J0GJjKhz1UqjFCyY0BquTYQ6UnI5REGaDKlD1USjNum2TNFSq12UEHBo00UmigKYf2NKnTCCWfBqgSag+Vbo1QkmuAKrv2UCnaCEXTBqZQbY+U1k3ahdoNTOXdHvqigvdMsmeGZrttyfzSlczD20/6Kt+ENHxAkRoe58VOk62EpZSQxx4en/Fw28pt1yjb7nYw2w1sQd8L798/45wePNxP28MfX7CsaQ9fw8PHxJuHpyc93C6l7FtmK5zfc3GuNl0zHzvG7aHMpf026Ye3ntfpeAc2f3jzGeYKavZTogmNSV5qyoXbApNtNWU5FkUHxbmCmv3Yg7RvYmo+T7cB5hBMZldbKdwhiHP1pve8TTGbswdbV5tvmLoN0PzOZHYt3k9prqB2nR1m849iO0hT/d0+mI9PA/L68WTDuYKmMNmFhl1gxdR+w9Ttg/n4QGC2ZdYxYayQOd0WnfV4YZq+6HbBfHwisGxLdJ9i0FxBS7G9edvTUo6TWbt4Pj4UsOuTzV0r01xBF7uhXbTOswle87rK7YP2VUbaUwl++YS5/MBmn24+YDKbm9dVTsdxzQYdT8cG6se9T6XMQcpSYnhbGpWOIxJ1HJlKxz1U6jhCUccRqnTcQ6WOIxR1HKFKxz1U6jjunqjjAJU67qADHUca6jjSlI57mtRxhKKOI1TpuIdKHUco6jhClY57qNRxhIKOI1PouEdKHSf7Ah1HptJxD31Rx3tC2RNEs962a37puuZh8Ce1lW9COj6g6HfF7RKmzMu6ruYC+1jH0zM6XrL52n58y9yn8bagh3tJQs8L5vTg4X7aOv74gtUuv9avoeNj4k3H85M6XspsVpuT/4Cd5ipaMP9JOZtd7HFp+oXbAssSzWrNJtxRSnMFtY02zmUOdqK137f1OciazWuPQ8xBca6gttHOS0nrHvb2+7ZuCyybHTFzyf6JwlgWIcaI9iDj1u4z3P5Xjs8oluIO7y88V8w9T+tm2+SS1/abt24bXI7PKGxJ/HCu4VxAF7setx3Evvthbr9567bBJQbb445V1uc2MFdQOx3tW29XV2suzUsAtw0ux2cU2WzafbhCcwW1zTYtqy02tug1r7DcNrgcn1HMa8zuJKW5ghrFZK3YdzWszSsstxEux2cUtuIFd81McwW9fcG6R1sccvMkdT6Oizb4eMl23vlx59i1i7qY8v3NsPUJH0ck+jgylY97qPRxhKKPI1T5uIdKH0co+jhClY97qPRx3D7RxwEqfdxBBz6ONPRxpCkf9zTp4whFH0eo8nEPlT6OUPRxhCof91Dp4wgFH0em8HGPlD5O+gU+jkzl4x76oo/3jLJniKa9bdn80pVNu8lZb+WbkI8PKNLHTaumsJU5bXGfy9jH8zM+Xksa0yc7GeLbp5C1pIE5hdh536a4Lbclv94LTh8u3oLc1uBTIfblOQ3vwm4GXk6H2PS64Pj4TsT+7sMhdlqK7bn7nPJj8+mF2ASEqeBxiN3gdUNsAuJYEDnEbhC7ITYRcSyIHGI3iN0Qm4g4FkQOsRvEbohNRBwLIoTYDVw3xCYcjgWOQ+wGsR9i84qBc8F0IXYD2g+xGYpzBeUQuwHththuaaxjheQQu4HshtjJjhiTeV+uwrjfeh5v2mxhQWPoptjJDpWjseKktE4HlD1Ntn+hJHRDbHvC9g1PvlyFcZ+T5zKVtKAXdEPsbMdD2O0WK3PqeMCxr7fnfcfEcYid7RiYS3aZfJ0OKHGfol25wJVPt8bOJhD78bEZZ/IwHnBMxEJY8WKnW2Nnk4rVVjmXycN4wCl28WgKB9c33RrbLsSnJRWfycN4wFnytN0W4cclTbfGzts85b34TB7GA46d08veXA9QcuKUyuIzeRgPAMe7sfb1cOFyMsRmIWmaF3fYPbNztzipgnSLelcjwskIu8z5fuv5ca7cDTe+GGKzWdVKhZ4XzPHhw900HddF0S+K7t+RYbdsdzmfYbujC+Zqs+MMWwsv5NYMxbmSbM6wtfVCbs1QnCsoZ9hafCG3ZijO5eUEZdjafSG3ZijOFZQzbK2/kFu76yaYKyhn2FqC35JrBr7PFIwzbO2/kFszEOcKyhm29l/IrRmKcwXlDFv7b+2ted2EsbokpQxb6y8u1eS/NbqUAuyiS+3CCEUZBqawYY+UYox7CZlxZUo1dlBtyQglTQao8mQPlcqMUHRmYApp9kjpz4gkgQamMmgPlTKNULJpgCqd9lBp1ggltQaocmsPlZqNUPJsgCrR9tC+c5N1oXQDTVm3p70o4D2P7IghV9g98/TSftJV+SYo4X3G2QRbWvjZDJueGcQp9Mxgjo8f7qZp4S6J/goWPibeLHw9H2Hzdx/nasvlCPuZt50fsbV74xnmCsoRtrZwiK0ZinMFxQhbOzjE1nwA4ly94c0RtnZwiK0ZinMF5QhbOzjE1gzFuYJyhK0dvNbWzISxQnKErU0cYmtm4lxBOcLWJg6xNUNxrqAcYWsTh9jaLZ4wlx/WUIStZRxXbJRxSC6ljLvkUss4QkHGkSlk3COljOOWgjIOTCnjDqplHKEo4whVMu6hUsYRCjKOTCHjHillHJEo48hUMu6hUsYRijKOUCXjHiplHKEo4whVMu6hUsYRijKOUCXjHtqXcZIvkHGkKRn3tBdlvKeTHT/kBrsnoF7fTyor3wRlvM84G2BLGT8bYfMn8bVMoWcGc3z8cDdNGXdB9FeQ8THxJuPb+QSb+wicq1yBE2wt45BaMxTnCsoJ9hMhSE2tGYpzBeUEW+t4ba2ZCWPZgmCCrV0cUmtG4lwxOcHWLg6pNZ9pOBdQl2BrF4fU2oU2MFdQTrC1jUNqzVCcKygn2NrGIbVmKM4VlBNsbeOQWjMU5wrKCba2cVyy0cYhuJQ27oJLbeMIBRtHprBxj5Q2jnsK2jgwpY07qLZxhKKNI1TZuIdKG0co2DgyhY17pLRxRKKNI1PZuIdKG0co2jhClY17qLRxhKKNI1TZuIdKG0co2jhClY17aN/Gyb7AxpGmbNzTXrTxnk92BJEL7J6BulucdVa+Cdp4n3E2v5Y2fjLBPnILM/3754+1oqljCrCT3TBtbz9IAPEOjhsJ9vu93Trh8isn2A52M/D9dIKNr0udhuOb9/ay6AA7bKt965J53mPr6QXYiHsMFY3z6watm18jrk4Vj+PrBq8bXyOvThWP0+sGr5teI69OFY/D6wavG14jr04VD7LrBqybXSOsThWMo+sGrx9d0xpRx4rokusGsp9cE7KOJZKD6wayG1zzQvg+lUDOrRvAbm4d5+PTrtWnqTDuF51xtoPaXl8Qg25uHcPxAdfq01QYDzjBDuY1jEvrmI6PszafpsJ4AIh2AKeA23+3tI75+ARr87/BGcYDju1heQ/4Y3Td0jqW40OrzXfwMB5wih2rJeIFTre1tgu3aZ5318HX6YBiOhFDxGuabmltbjIdEYLr4GE84Kx2jbhGvIzpltZmJ9OaZt/Bw3jA2RZboxNeuXRL6zRHM6zZdfB1OqDYubzuCS9Wurm1ndFTLrPv4GE8+CmFMNuinPD65GRuTfLRtizOrTsW529xSvs4z673NCSczK1TXO43L4/T5W6z6cXcmjyq9ij4vDBswUePIUzTZ138/KLU/h25dcts7bA73Vvz4fUYy72Oa2ttt1BVE7KOpU9za60VF5pqQtaxRHJprS0XimpC1rG+aqDOWosu9NSErGOJ5Mpauy7U1Hxp9BhLJDfW2njfemrC3UcSxYW1ll0oqQlXxxLJfbWWXeioCVnHEsl1tZbdmlHTMvmYyutNaqu169K6jLJbS0ppu66k1OKLUDJfgCr19dC+BSONNBhoyoM9TSoxQsmJAaqk2EOlHyOUBBmgypA9VMoyQtGWgSl02SOlOSOS1BmYyp09VGo0QsmjAapE2kOlU6PwoFTDTyEIq/ZIKdiIJMMGplJs/6MPL9p2Rxp7FshtdUczPxj6KTF1MTYa94Bxtq2Wyn22rcZnBtUJPjPMV/DhY+7SVG5XOn8F5R4T78r97F9WhIiavv11LHdcTqufeUP5kVDzW8qPsURyWK2VGwJqQtaxRGJWrYUb8mk69upYvpHNUbUWboinCVnHEslJtRZuSKcJWccSyUG1Fu5aThPxMZVAzqm1dkM2TcQ6lkiOqbV2QzRNyDqWSE6ptXZDMs1r5WOsP3ihkFqbNy3PYN6QTUrzdtmkNm+EonkjVJm3h/bNG2lo3khT5u1p0rwRiuaNUGXeHirNG6Fo3ghV5u2h0rwRCuaNTGHeHinNG5Fo3shU5u2h0rwRiuaNUGXeHirNG70HzBt/5ECYt0dK80YkmjcylXn7n3N40bw77tiTQQ6pO7b5wdVP+akrr9G8B4yzIbU077MhNX2qXgsTfGaYquDDx7Slad4ua/4K5j0m3s372b/xCMU0pQ51LNsD7qi1eUMvTcg6lkiuqJ/oOWotTcg6lkhuqLV711iaiI+pTjqwoNbiDaU0AetYErmf1uINnTSdYHWskK6e1uINlTSXMo+xRHI7rdUbGmlC1rFEcjmt1RsKaULWsURyN63VG/poQtaxRHI1rdWb1mdQb2gkpXq7RlKrN0JRvRGq1NtD++qNNFRvpCn19jSp3ghF9UaoUm8PleqNUFRvhCr19lCp3ggF9UamUG+PlOqNSFRvZCr19lCp3ghF9UaoUm8PleqN4gPqjT9fINTbI6V6IxLVG5lKvf0PNbyo3h157NkgV9Md3fS3OCeoLrNG9R4wzlbTUr3PVtPp+OT1w1+Ar2OqpoNtW3l9y/7rndC4VU0/7u1UMH28JJdTwfSDc5fsZ/+AI8TS8HJgQj2l91dDx9LmvusWosikAQTt9JjDmTRx+oE0gDCbHpM4kCZSP40GEgbTYxKn0UTqR9FAwlR6TOIomkj9HBpIGEmPSZBDE6YfQgMG8+gxhkNoIg0SaDzhsYwes1wCTbBB/IwwbKIFjONngvWzZ1rJagwtUJw9E6ofPNsF1Bw/BqLv01HvarpTcFfv1s62py5L9L8l+X04iDZXs5uCu3i3dy52iRSjT0PrtA/Jx5+gKbhrd2PnpdiGlfzvR67TPmQ5/uTMgj+o1i2dV/v6nHyBXqd9yHr8iZkFr0i6mfNm+6Dtwy4/r9M+ZDv+pMw6zpvDPE/7kVxxfA7j/v3vx1+PWfGao1s3h3D8rZjs23MY9zHHejHHFS8zunWzWbpd8mffnsN4wLG72bYNryy6fbOtfXadX1x7XqcDSjr+NsyGFxNn62a0hbYNubq5bVv+Fqf0jL4e7mlIOFk323n6dvPHWXK3zvxq3Yz6AxEJPC8Y06OHe2l7p2+NX5PPJ39arwu7G+izf7MQm2Y6qCB1FhsZN80jCcWaGWEYOQvh5Zp5ZKLYMSMM82YB4455JKNYMCMMw2al8lQwj3wU22WEYdIsYNwuj5QUq2W6RoGYWcC4Wh6J6XuvjKD3hFlAuFceOSmWygjCgFnAuFQeOSk2ygjDdFnAuFEeOSnUybjMQbMsLvOoTh4pKdWW4KQYW46l9ENVKvwUf2MqCCr8RtmxofrmUckq8MhWK1Dpqv8NtspcgUjqWonKXR1RaiwQyWMrUYmsI0qnBSJJbSUqq3XEvuDidk2GW1lScR1M2i4ySXeBqXzXMbX6IpTcF6BKfj1UejBCUYSBKUzYI1+U4o7b9WTNRchtG/wg0qf8kW9AYjxgnI2QpRmfjpDhmWHkAc8MxvTw4V7aZuyT4F/VjLuwuxk/+7cEMT3GbzoWyWI75fR4/PZsjY7pDVpokQWMo+ORGWNujDCskAUMc+ORF2NojMcX9sfiDWEOjUdejIkxwrA8FjBOjEdejHExwrA5FjCOi0deDFkxsiA2FijOikd2jEExsrAzFjAOikd2jCkxwrAwFjBOiUd2jBExrXXQFqsPJygiHgkyRZFVkKmJHAvyh/hTCDL+FtMqyPhbXseC7NNEJcjAQ0EGoBJk/1tllSADEQUZiEqQHVEKMhBRkIGoBNkRpSADEQUZiEqQHbEvyLhroyADSwqyg0lBRiYKMjKVIDumFmSEoiAjVAmyh0pBRigIMjKFIHvki4LcUbyes7lWuC2FH5T6lEbyDUiQB4yzrbAU5NOtMH7MDCkGPDMY08OHe2kLsi93f1VB7sLugvzsnx/EQhg/68dwWHwEz4XwSJCxDUYYJsMCxm3wMGKAKhhhGAsLGFfBI0WGHhhZUAmrjgF74JEfYwmMKAyEBYtL4JEfYwOMJw6mwWOYa4BHfoz1L8UgEAULGNe/I0PG7hdhmAMLGHe/I0PG4hdhGAILGBe/I0PG1hdhmAALGLe+I0OmdrEaMqWLY0P+0GgKQ8bfLFoNGX/z6tiQfUGoDBl4aMgAVIbsf9OrMmQgoiEDURmyI0pDBiIaMhCVITuiNGQgoiEDURmyI/YNGbdtNGRgSUN2MGnIyERDRqYyZMfUhoxQNGSEKkP2UGnICAVDRqYwZI980ZA7jteTNpf0tq3Q3+KcR/INyJAHjLNJrzTkx5n3/eX/ATcNo6IKZW5kc3RyZWFtCmVuZG9iagoxMSAwIG9iago2MDczCmVuZG9iagozMiAwIG9iago8PCAvRmlsdGVyIC9GbGF0ZURlY29kZSAvTGVuZ3RoIDc3ID4+CnN0cmVhbQp4nDM3NVIwULC0ABJmpiYK5kaWCimGXEA+iJXLZWhpDmblgFkmxgZAlqmpKRILIgvTC2HB5GC0sYk51AQECyQHtjYHZlsOVxoAnuAbmgplbmRzdHJlYW0KZW5kb2JqCjMzIDAgb2JqCjw8IC9GaWx0ZXIgL0ZsYXRlRGVjb2RlIC9MZW5ndGggNTkgPj4Kc3RyZWFtCnicMzU1VzBQsLQAEqamRgrmRpYKKYZcQD6IlctlaGkOZuWAWRbGQAZIGZxhAKTBmnNgenK40gCp4RBaCmVuZHN0cmVhbQplbmRvYmoKMzQgMCBvYmoKPDwgL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0xlbmd0aCAzMDQgPj4Kc3RyZWFtCnicPZI7ksMwDEN7nYIXyIz4k+TzZCeV9/7tPjLJVoBJiQAoL3WZsqY8IGkmCf/R4eFiO+V32J7NzMC1RC8TyynPoSvE3EX5spmNurI6xarDMJ1b9Kici4ZNk5rnKksZtwuew7WJ55Z9xA83NKgHdY1Lwg3d1WhZCs1wdf87vUfZdzU8F5tU6tQXjxdRFeb5IU+ih+lK4nw8KCFcezBGFhLkU9FAjrNcrfJeQvYOtxqywkFqSeezJzzYdXpPLm4XzRAPZLlU+E5R7O3QM77sSgk9ErbhWO59O5qx6RqbOOx+70bWyoyuaCF+yFcn6yVg3FMmRRJkTrZYbovVnu6hKKZzhnMZIOrZioZS5mJXq38MO28sL9ksyJTMCzJGp02eOHjIfo2a9HmV53j9AWzzczsKZW5kc3RyZWFtCmVuZG9iagozNSAwIG9iago8PCAvRmlsdGVyIC9GbGF0ZURlY29kZSAvTGVuZ3RoIDY2ID4+CnN0cmVhbQp4nDM2tFAwUDA3V9A1NDRVMDIyUDA0MlFIMeQyNDQHM3O5YII5YJaJAZBhCCTBGnK4YFpzwDogslCtOVxpAE04EfUKZW5kc3RyZWFtCmVuZG9iagozNiAwIG9iago8PCAvRmlsdGVyIC9GbGF0ZURlY29kZSAvTGVuZ3RoIDIyNyA+PgpzdHJlYW0KeJw1TzuyAyEM6zmFLpAZjG1gz7OZVC/3b59ksg0S/kjy9ERHJl7myAis2fG2FhmIGfgWU/GvPe3DhOo9uIcI5eJCmGEknDXruJun48W/XeUz1sG7Db5ilhcEtjCT9ZXFmct2wVgaJ3FOshtj10RsY13r6RTWEUwoAyGd7TAlyBwVKX2yo4w5Ok7kiediqsUuv+9hfcGmMaLCHFcFT9BkUJY97yagHRf039WN30k0i14CMpFgYZ0k5s5ZTvjVa0fHUYsiMSekGeQyEdKcrmIKoQnFOjsKKhUFl+pzyt0+/2hdW00KZW5kc3RyZWFtCmVuZG9iagozNyAwIG9iago8PCAvRmlsdGVyIC9GbGF0ZURlY29kZSAvTGVuZ3RoIDI0NSA+PgpzdHJlYW0KeJxFULuNQzEM6z0FFwhg/Sx7nndIldu/PUpGcIUhWj+SWhKYiMBLDLGUb+JHRkE9C78XheIzxM8XhUHOhKRAnPUZEJl4htpGbuh2cM68wzOMOQIXxVpwptOZ9lzY5JwHJxDObZTxjEK6SVQVcVSfcUzxqrLPjdeBpbVss9OR7CGNhEtJJSaXflMq/7QpWyro2kUTsEjkgZNNNOEsP0OSYsyglFH3MLWO9HGykUd10MnZnDktmdnup+1MfA9YJplR5Smd5zI+J6nzXE597rMd0eSipVX7nP3ekZbyIrXbodXpVyVRmY3Vp5C4PP+Mn/H+A46gWT4KZW5kc3RyZWFtCmVuZG9iagozOCAwIG9iago8PCAvRmlsdGVyIC9GbGF0ZURlY29kZSAvTGVuZ3RoIDM5MiA+PgpzdHJlYW0KeJw9UktuBTEI288puECl8E1ynqne7t1/W5vMVKoKLwO2MZSXDKklP+qSiDNMfvVyXeJR8r1samfmIe4uNqb4WHJfuobYctGaYrFPHMkvyLRUWKFW3aND8YUoEw8ALeCBBeG+HP/xF6jB17CFcsN7ZAJgStRuQMZD0RlIWUERYfuRFeikUK9s4e8oIFfUrIWhdGKIDZYAKb6rDYmYqNmgh4SVkqod0vGMpPBbwV2JYVBbW9sEeGbQENnekY0RM+3RGXFZEWs/PemjUTK1URkPTWd88d0yUvPRFeik0sjdykNnz0InYCTmSZjncCPhnttBCzH0ca+WT2z3mClWkfAFO8oBA7393pKNz3vgLIxc2+xMJ/DRaaccE62+HmL9gz9sS5tcxyuHRRSovCgIftdBE3F8WMX3ZKNEd7QB1iMT1WglEAwSws7tMPJ4xnnZ3hW05vREaKNEHtSOET0ossXlnBWwp/yszbEcng8me2+0j5TMzKiEFdR2eqi2z2Md1Hee+/r8AS4AoRkKZW5kc3RyZWFtCmVuZG9iagozOSAwIG9iago8PCAvRmlsdGVyIC9GbGF0ZURlY29kZSAvTGVuZ3RoIDI0NyA+PgpzdHJlYW0KeJxNUbttRDEM698UXOAA62t5ngtSXfZvQ8kIkMIgoS8ppyUW9sZLDOEHWw++5JFVQ38ePzHsMyw9yeTUP+a5yVQUvhWqm5hQF2Lh/WgEvBZ0LyIrygffj2UMc8734KMQl2AmNGCsb0kmF9W8M2TCiaGOw0GbVBh3TRQsrhXNM8jtVjeyOrMgbHglE+LGAEQE2ReQzWCjjLGVkMVyHqgKkgVaYNfpG1GLgiuU1gl0otbEuszgq+f2djdDL/LgqLp4fQzrS7DC6KV7LHyuQh/M9Ew7d0kjvfCmExFmDwVSmZ2RlTo9Yn23QP+fZSv4+8nP8/0LFShcKgplbmRzdHJlYW0KZW5kb2JqCjQwIDAgb2JqCjw8IC9GaWx0ZXIgL0ZsYXRlRGVjb2RlIC9MZW5ndGggOTAgPj4Kc3RyZWFtCnicTY1BEsAgCAPvvCJPUETQ/3R60v9fq9QOvcBOAokWRYL0NWpLMO64MhVrUCmYlJfAVTBcC9ruosr+MklMnYbTe7cDg7LxcYPSSfv2cXoAq/16Bt0P0hwiWAplbmRzdHJlYW0KZW5kb2JqCjQxIDAgb2JqCjw8IC9GaWx0ZXIgL0ZsYXRlRGVjb2RlIC9MZW5ndGggMzIwID4+CnN0cmVhbQp4nDVRu3HFMAzrNQUX8J34lTSPc6/K278NQDsVYRoEQKq8ZEq5XOqSVbLC5EeH6hRN+T5gpvwO9ZDj6B7ZIbpT1pZ7GAjLxDyljlhNlnu4BYEvDE2JuYXz9wjoKwajMBOBusXfP0CzJDBpcPBTkGutWmKJDjwsFlizK8ytGilUyFV8Oza5BwVycbPQpxyaFLfcgvBliGRHarGvy2Up8rv1CRiEFeaITxSJheeBDmYi8ScDYnv22WJXVy+qERnWSYcHUgTSbG4SMDRFsuqDG9hXxzU/T0fZwclBv4rB+DY4mS9JeV8FoRCPF/4Oz9nIsZJDJBTyfbXAiCNsgBGhT+0jEGUgNEX37plSPiZViu8ARiEcfapXMrwXkdlqhs3/GV3ZKgoGVVkfn0ZwJoNJOPNkowrTUrXTv/vc4/MHY2N6gAplbmRzdHJlYW0KZW5kb2JqCjQyIDAgb2JqCjw8IC9GaWx0ZXIgL0ZsYXRlRGVjb2RlIC9MZW5ndGggODAgPj4Kc3RyZWFtCnicRYy7DcAwCER7pmAEfiZmnyiVs38bIErccE+6e7g6EjJT3mGGhwSeDCyGU/EGmaNgNbhGUo2d7KOwbl91geZ6U6v19wcqT3Z2cT3Nyxn0CmVuZHN0cmVhbQplbmRvYmoKNDMgMCBvYmoKPDwgL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0xlbmd0aCAxNTcgPj4Kc3RyZWFtCnicRZC5EUMxCERzVUEJErAI6rHH0Xf/qRf5SrRvAC2HryVTqh8nIqbc12j0MHkOn00lVizYJraTGnIbFkFKMZh4TjGro7ehmYfU67ioqrh1ZpXTacvKxX/zaFczkz3CNeon8E3o+J88tKnoW6CvC5R9QLU4nUlQMX2vYoGjnHZ/IpwY4D4ZR5kpI3Fibgrs9xkAZr5XuMbjBd0BN3kKZW5kc3RyZWFtCmVuZG9iago0NCAwIG9iago8PCAvRmlsdGVyIC9GbGF0ZURlY29kZSAvTGVuZ3RoIDY4ID4+CnN0cmVhbQp4nDMzNlMwULAwAhKmpoYK5kaWCimGXEA+iJXLBRPLAbPMLMyBLCMLkJYcLkMLYzBtYmykYGZiBmRZIDEgutIAcvgSkQplbmRzdHJlYW0KZW5kb2JqCjQ1IDAgb2JqCjw8IC9GaWx0ZXIgL0ZsYXRlRGVjb2RlIC9MZW5ndGggMzE3ID4+CnN0cmVhbQp4nDVSS3JDMQjbv1Nwgc6Yv32edLJq7r+thCcrsC1AQi4vWdJLftQl26XD5Fcf9yWxQj6P7ZrMUsX3FrMUzy2vR88Rty0KBFETPfgyJxUi1M/U6Dp4YZc+A68QTikWeAeTAAav4V94lE6DwDsbMt4Rk5EaECTBmkuLTUiUPUn8K+X1pJU0dH4mK3P5e3KpFGqjyQgVIFi52AekKykeJBM9iUiycr03VojekFeSx2clJhkQ3SaxTbTA49yVtISZmEIF5liA1XSzuvocTFjjsITxKmEW1YNNnjWphGa0jmNkw3j3wkyJhYbDElCbfZUJqpeP09wJI6ZHTXbtwrJbNu8hRKP5MyyUwccoJAGHTmMkCtKwgBGBOb2wir3mCzkWwIhlnZosDG1oJbt6joXA0JyzpWHG157X8/4HRVt7owplbmRzdHJlYW0KZW5kb2JqCjQ2IDAgb2JqCjw8IC9GaWx0ZXIgL0ZsYXRlRGVjb2RlIC9MZW5ndGggMTcgPj4Kc3RyZWFtCnicMza0UDCAwxRDLgAalALsCmVuZHN0cmVhbQplbmRvYmoKNDcgMCBvYmoKPDwgL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0xlbmd0aCAzMzggPj4Kc3RyZWFtCnicNVI5rt1ADOt9Cl0ggHbNnOcFqX7u34aUXwpDtFaKmo4WlWn5ZSFVLZMuv+1JbYkb8vfJCokTklcl2qUMkVD5PIVUv2fLvL7WnBEgS5UKk5OSxyUL/gyX3i4c52NrP48jdz16YFWMhBIByxQTo2tZOrvDmo38PKYBP+IRcq5YtxxjFUgNunHaFe9D83nIGiBmmJaKCl1WiRZ+QfGgR61991hUWCDR7RxJcIyNUJGAdoHaSAw5sxa7qC/6WZSYCXTtiyLuosASScycYl06+g8+dCyovzbjy6+OSvpIK2tM2nejSWnMIpOul0VvN299PbhA8y7Kf17NIEFT1ihpfNCqnWMomhllhXccmgw0xxyHzBM8hzMSlPR9KH5fSya6KJE/Dg2hf18eo4ycBm8Bc9GftooDF/HZYa8cYIXSxZrkfUAqE3pg+v/X+Hn+/AMctoBUCmVuZHN0cmVhbQplbmRvYmoKNDggMCBvYmoKPDwgL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0xlbmd0aCAyNDggPj4Kc3RyZWFtCnicLVE5kgNBCMvnFXpCc9PvscuR9//pCsoBg4ZDIDotcVDGTxCWK97yyFW04e+ZGMF3waHfynUbFjkQFUjSGFRNqF28Hr0HdhxmAvOkNSyDGesDP2MKN3pxeEzG2e11GTUEe9drT2ZQMisXccnEBVN12MiZw0+mjAvtXM8NyLkR1mUYpJuVxoyEI00hUkih6iapM0GQBKOrUaONHMV+6csjnWFVI2oM+1xL29dzE84aNDsWqzw5pUdXnMvJxQsrB/28zcBFVBqrPBAScL/bQ/2c7OQ33tK5s8X0+F5zsrwwFVjx5rUbkE21+Dcv4vg94+v5/AOopVsWCmVuZHN0cmVhbQplbmRvYmoKNDkgMCBvYmoKPDwgL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0xlbmd0aCAxMzggPj4Kc3RyZWFtCnicPY9BDgMxCAPveYU/ECl2Qljes1VP2/9fS5rdXtAIjDEWQkNvqGoOm4INx4ulS6jW8CmKiUoOyJlgDqWk0h1nkXpiOBjcHrQbzuKx6foRu5JWfdDmRrolaIJH7FNp3JZxE8QDNQXqKepco7wQuZ+pV9g0kt20spJrOKbfveep6//TVd5fX98ujAplbmRzdHJlYW0KZW5kb2JqCjUwIDAgb2JqCjw8IC9GaWx0ZXIgL0ZsYXRlRGVjb2RlIC9MZW5ndGggMjEwID4+CnN0cmVhbQp4nDVQyw1DMQi7ZwoWqBQCgWSeVr11/2tt0DthEf9CWMiUCHmpyc4p6Us+OkwPti6/sSILrXUl7MqaIJ4r76GZsrHR2OJgcBomXoAWN2DoaY0aNXThgqYulUKBxSXwmXx1e+i+Txl4ahlydgQRQ8lgCWq6Fk1YtDyfkE4B4v9+w+4t5KGS88qeG/kbnO3wO7Nu4SdqdiLRchUy1LM0xxgIE0UePHlFpnDis9Z31TQS1GYLTpYBrk4/jA4AYCJeWYDsrkQ5S9KOpZ9vvMf3D0AAU7QKZW5kc3RyZWFtCmVuZG9iagozMCAwIG9iago8PCAvQmFzZUZvbnQgL0RlamFWdVNhbnMgL0NoYXJQcm9jcyAzMSAwIFIKL0VuY29kaW5nIDw8Ci9EaWZmZXJlbmNlcyBbIDMyIC9zcGFjZSA0NCAvY29tbWEgNDggL3plcm8gL29uZSAvdHdvIC90aHJlZSAvZm91ciAvZml2ZSAvc2l4IC9zZXZlbgovZWlnaHQgL25pbmUgNzIgL0ggNzYgL0wgOTcgL2EgMTAwIC9kIC9lIDExNCAvciAxMjEgL3kgXQovVHlwZSAvRW5jb2RpbmcgPj4KL0ZpcnN0Q2hhciAwIC9Gb250QkJveCBbIC0xMDIxIC00NjMgMTc5NCAxMjMzIF0gL0ZvbnREZXNjcmlwdG9yIDI5IDAgUgovRm9udE1hdHJpeCBbIDAuMDAxIDAgMCAwLjAwMSAwIDAgXSAvTGFzdENoYXIgMjU1IC9OYW1lIC9EZWphVnVTYW5zCi9TdWJ0eXBlIC9UeXBlMyAvVHlwZSAvRm9udCAvV2lkdGhzIDI4IDAgUiA+PgplbmRvYmoKMjkgMCBvYmoKPDwgL0FzY2VudCA5MjkgL0NhcEhlaWdodCAwIC9EZXNjZW50IC0yMzYgL0ZsYWdzIDMyCi9Gb250QkJveCBbIC0xMDIxIC00NjMgMTc5NCAxMjMzIF0gL0ZvbnROYW1lIC9EZWphVnVTYW5zIC9JdGFsaWNBbmdsZSAwCi9NYXhXaWR0aCAxMzQyIC9TdGVtViAwIC9UeXBlIC9Gb250RGVzY3JpcHRvciAvWEhlaWdodCAwID4+CmVuZG9iagoyOCAwIG9iagpbIDYwMCA2MDAgNjAwIDYwMCA2MDAgNjAwIDYwMCA2MDAgNjAwIDYwMCA2MDAgNjAwIDYwMCA2MDAgNjAwIDYwMCA2MDAgNjAwCjYwMCA2MDAgNjAwIDYwMCA2MDAgNjAwIDYwMCA2MDAgNjAwIDYwMCA2MDAgNjAwIDYwMCA2MDAgMzE4IDQwMSA0NjAgODM4IDYzNgo5NTAgNzgwIDI3NSAzOTAgMzkwIDUwMCA4MzggMzE4IDM2MSAzMTggMzM3IDYzNiA2MzYgNjM2IDYzNiA2MzYgNjM2IDYzNiA2MzYKNjM2IDYzNiAzMzcgMzM3IDgzOCA4MzggODM4IDUzMSAxMDAwIDY4NCA2ODYgNjk4IDc3MCA2MzIgNTc1IDc3NSA3NTIgMjk1CjI5NSA2NTYgNTU3IDg2MyA3NDggNzg3IDYwMyA3ODcgNjk1IDYzNSA2MTEgNzMyIDY4NCA5ODkgNjg1IDYxMSA2ODUgMzkwIDMzNwozOTAgODM4IDUwMCA1MDAgNjEzIDYzNSA1NTAgNjM1IDYxNSAzNTIgNjM1IDYzNCAyNzggMjc4IDU3OSAyNzggOTc0IDYzNCA2MTIKNjM1IDYzNSA0MTEgNTIxIDM5MiA2MzQgNTkyIDgxOCA1OTIgNTkyIDUyNSA2MzYgMzM3IDYzNiA4MzggNjAwIDYzNiA2MDAgMzE4CjM1MiA1MTggMTAwMCA1MDAgNTAwIDUwMCAxMzQyIDYzNSA0MDAgMTA3MCA2MDAgNjg1IDYwMCA2MDAgMzE4IDMxOCA1MTggNTE4CjU5MCA1MDAgMTAwMCA1MDAgMTAwMCA1MjEgNDAwIDEwMjMgNjAwIDUyNSA2MTEgMzE4IDQwMSA2MzYgNjM2IDYzNiA2MzYgMzM3CjUwMCA1MDAgMTAwMCA0NzEgNjEyIDgzOCAzNjEgMTAwMCA1MDAgNTAwIDgzOCA0MDEgNDAxIDUwMCA2MzYgNjM2IDMxOCA1MDAKNDAxIDQ3MSA2MTIgOTY5IDk2OSA5NjkgNTMxIDY4NCA2ODQgNjg0IDY4NCA2ODQgNjg0IDk3NCA2OTggNjMyIDYzMiA2MzIgNjMyCjI5NSAyOTUgMjk1IDI5NSA3NzUgNzQ4IDc4NyA3ODcgNzg3IDc4NyA3ODcgODM4IDc4NyA3MzIgNzMyIDczMiA3MzIgNjExIDYwNQo2MzAgNjEzIDYxMyA2MTMgNjEzIDYxMyA2MTMgOTgyIDU1MCA2MTUgNjE1IDYxNSA2MTUgMjc4IDI3OCAyNzggMjc4IDYxMiA2MzQKNjEyIDYxMiA2MTIgNjEyIDYxMiA4MzggNjEyIDYzNCA2MzQgNjM0IDYzNCA1OTIgNjM1IDU5MiBdCmVuZG9iagozMSAwIG9iago8PCAvSCAzMiAwIFIgL0wgMzMgMCBSIC9hIDM0IDAgUiAvY29tbWEgMzUgMCBSIC9kIDM2IDAgUiAvZSAzNyAwIFIKL2VpZ2h0IDM4IDAgUiAvZml2ZSAzOSAwIFIgL2ZvdXIgNDAgMCBSIC9uaW5lIDQxIDAgUiAvb25lIDQyIDAgUiAvciA0MyAwIFIKL3NldmVuIDQ0IDAgUiAvc2l4IDQ1IDAgUiAvc3BhY2UgNDYgMCBSIC90aHJlZSA0NyAwIFIgL3R3byA0OCAwIFIgL3kgNDkgMCBSCi96ZXJvIDUwIDAgUiA+PgplbmRvYmoKMyAwIG9iago8PCAvRjEgMzAgMCBSID4+CmVuZG9iago0IDAgb2JqCjw8IC9BMSA8PCAvQ0EgMCAvVHlwZSAvRXh0R1N0YXRlIC9jYSAxID4+Ci9BMiA8PCAvQ0EgMSAvVHlwZSAvRXh0R1N0YXRlIC9jYSAxID4+ID4+CmVuZG9iago1IDAgb2JqCjw8ID4+CmVuZG9iago2IDAgb2JqCjw8ID4+CmVuZG9iago3IDAgb2JqCjw8IC9JMSAxMiAwIFIgL0kxMCAyMSAwIFIgL0kxMSAyMiAwIFIgL0kxMiAyMyAwIFIgL0kxMyAyNCAwIFIgL0kxNCAyNSAwIFIKL0kxNSAyNiAwIFIgL0kxNiAyNyAwIFIgL0kyIDEzIDAgUiAvSTMgMTQgMCBSIC9JNCAxNSAwIFIgL0k1IDE2IDAgUgovSTYgMTcgMCBSIC9JNyAxOCAwIFIgL0k4IDE5IDAgUiAvSTkgMjAgMCBSID4+CmVuZG9iagoxMiAwIG9iago8PCAvQml0c1BlckNvbXBvbmVudCA4IC9Db2xvclNwYWNlIC9EZXZpY2VSR0IKL0RlY29kZVBhcm1zIDw8IC9Db2xvcnMgMyAvQ29sdW1ucyAxMTkgL1ByZWRpY3RvciAxMCA+PgovRmlsdGVyIC9GbGF0ZURlY29kZSAvSGVpZ2h0IDExOSAvTGVuZ3RoIDUxIDAgUiAvU3VidHlwZSAvSW1hZ2UKL1R5cGUgL1hPYmplY3QgL1dpZHRoIDExOSA+PgpzdHJlYW0KeJzt3D1KxFAUQOHERAaslEHFTmwtBBEEfyoLG7dga+OOXIWFGxCs7awEHQSRQcRCEQaNugPvLR5HnTlffUnC4TbzSKbuHjarSLO9EM7oB1O//QATwcoEKxOsTLAywcoEKxOsTLAywcoEKxPaUmcU3cUwnGl2l+ILfX0WeJo/xl0mWJlgZYKVCVYmWJlgZYKVCVYmWJlgZUI9ut0Kh3p7/fhK7Uw88/GWeKQx5C4TrEywMsHKBCsTrEywMsHKBCsTrEywMqFNnVEkdOeDcGZiv09xlwlWJliZYGWClQlWJliZYGWClQlWJliZUPv/GAB3mWBlgpUJViZYmWBlgpUJViZYmWBlgpUJbbO/kRi7Cyeezt7Dmf7BdOJeY8hdJliZYGWClQlWJliZYGWClQlWJliZ0Fav8a/njNm558TUhL5x4C4TrEywMsHKBCsTrEywMsHKBCsTrEywMqEtdaF/91XE8DSeWTxciYdebsIRd5lgZYKVCVYmWJlgZYKVCVYmWJlgZYKVCcXOMS5P5sOZ9aPH+EJNLxjoRrknCuwcL8dDiTOKDHeZYGWClQlWJliZYGWClQlWJliZYGWClQnFzjHWVq8SU4l3NgodU4Su7wfMjSp3mWFlgpUJViZYmWBlgpUJViZYmWBlgpUJ38pMLA0KZW5kc3RyZWFtCmVuZG9iago1MSAwIG9iago0OTcKZW5kb2JqCjEzIDAgb2JqCjw8IC9CaXRzUGVyQ29tcG9uZW50IDggL0NvbG9yU3BhY2UgL0RldmljZVJHQgovRGVjb2RlUGFybXMgPDwgL0NvbG9ycyAzIC9Db2x1bW5zIDExOSAvUHJlZGljdG9yIDEwID4+Ci9GaWx0ZXIgL0ZsYXRlRGVjb2RlIC9IZWlnaHQgMTE5IC9MZW5ndGggNTIgMCBSIC9TdWJ0eXBlIC9JbWFnZQovVHlwZSAvWE9iamVjdCAvV2lkdGggMTE5ID4+CnN0cmVhbQp4nO3dsQkCQRBAUc+zBsHYIgzUGmxZsBirOLUD10A+Iu/FA7N8Jt9ptb+shp6P8cxPmebxzHP5yqrleh/OrL+yifdULqhcULmgckHlgsoFlQsqF1QuqFyYlvthODQft8FT/phbLqhcULmgckHlgsoFlQsqF1QuqFxQuaByQeWCygWVCyoXVC6oXFC5oHJB5YLKBZULKhdULqhcULmgckHlgsoFlQsqF1QuqFxQuaByQeWCygWVCyoXVC6oXFC5oHJB5YLKBZULKhdULqhcULmgckHlgsoFlQsqF1QuqFzYzKfdB2PjTyeW2/hDifn8ya4/5JYLKhdULqhcULmgckHlgsoFlQsqF1Qu+Emj4JYLKhdULqhcULmgckHlgsoFlQsqF1QuqFxQuaByQeWCygWVCyoXVC6oXFC5oHJB5YLKBZULKhdULqhcULmgckHlgsoFlQsqF1QuvAAewBE0CmVuZHN0cmVhbQplbmRvYmoKNTIgMCBvYmoKMzQ5CmVuZG9iagoxNCAwIG9iago8PCAvQml0c1BlckNvbXBvbmVudCA4IC9Db2xvclNwYWNlIC9EZXZpY2VSR0IKL0RlY29kZVBhcm1zIDw8IC9Db2xvcnMgMyAvQ29sdW1ucyAxMTkgL1ByZWRpY3RvciAxMCA+PgovRmlsdGVyIC9GbGF0ZURlY29kZSAvSGVpZ2h0IDExOSAvTGVuZ3RoIDUzIDAgUiAvU3VidHlwZSAvSW1hZ2UKL1R5cGUgL1hPYmplY3QgL1dpZHRoIDExOSA+PgpzdHJlYW0KeJzt3btqVFEcRvF9xtwGMQSiiBcsMqVoKYJgYZMiCHmENKYRG0lhJT6AnVgp+BqKqKS108pGEXUUNYoyXiYZk/gG823xsLRYv3b+7HNYs6vNnDNN6S2XVnQm2lln79Ew8PN9XmT0Lc/MzOeZyX1x5P7V3TjTyVfSX7MywcoEKxOsTLAywcoEKxOsTLAywcqEpiyc/9f38Geurx6PM2u3nsWZS0sLcebG3df5hn4N44h7mWBlgpUJViZYmWBlgpUJViZYmWBlgpUJLf2IolJT86U24z9ePP0uLrF2O1/o4ZNRvpftrTxTwb1MsDLBygQrE6xMsDLBygQrE6xMsDLBygT2HGN3J8+ks44Tqz/iGqMHb+LM5GI330zNDVdwLxOsTLAywcoEKxOsTLAywcoEKxOsTLAygT3HqJGODrbXX8U19pw5mC80FX740SL3MsHKBCsTrEywMsHKBCsTrEywMsHKBCsT/r/nStKrNfv9qbxIJ797s3T355mtr3mmgnuZYGWClQlWJliZYGWClQlWJliZYGUC+wbK2V6eGX4c//mV5SNxjcH3/FaGm/fyUxE1b5es4V4mWJlgZYKVCVYmWJlgZYKVCVYmWJlgZULz9NFKHDp5sWKl6bk8M8gPK5SddATRrXiaoeZfQ0HuZYKVCVYmWJlgZYKVCVYmWJlgZYKVCVYmNGeX8iFF71B+MOLOej9fbeZAnhkNxn9+6thcXOPx84p/8pyezzObnyvWyffjXiZYmWBlgpUJViZYmWBlgpUJViZYmWBlwsS1C/klB+cuv2jnapsbeSb9CefLLxXnDxUOz4ZXRJRS3n4ID7mUUsrwUxxxLxOsTLAywcoEKxOsTLAywcoEKxOsTLAy4TedcFUECmVuZHN0cmVhbQplbmRvYmoKNTMgMCBvYmoKNjYwCmVuZG9iagoxNSAwIG9iago8PCAvQml0c1BlckNvbXBvbmVudCA4IC9Db2xvclNwYWNlIC9EZXZpY2VSR0IKL0RlY29kZVBhcm1zIDw8IC9Db2xvcnMgMyAvQ29sdW1ucyAxMTkgL1ByZWRpY3RvciAxMCA+PgovRmlsdGVyIC9GbGF0ZURlY29kZSAvSGVpZ2h0IDExOSAvTGVuZ3RoIDU0IDAgUiAvU3VidHlwZSAvSW1hZ2UKL1R5cGUgL1hPYmplY3QgL1dpZHRoIDExOSA+PgpzdHJlYW0KeJzt3b9Kw1AYQPGmTUHQpVRFxEGLgyAOxUn6COLk5uTk7hv4HE7irpuDszgqLm4FdbCKfxCkg1Da+gZ+FxoOlp7f/JEbDncKyU12dnpYiuwetcMZTlYJR7abS+HMxe1TwlrleGY4CEcSrqKRWZlgZYKVCVYmWJlgZYKVCVYmWJlgZUJWauwUcqH+9Vs4U2nNF7LW2HEvE6xMsDLBygQrE6xMsDLBygQrE6xMsDIhT5pKeAViYp9R9K9ewhn3MsHKBCsTrEywMsHKBCsTrEywMsHKBCsT0p5jDPvhyM3xXDizefCetNzfytV4ZtArYKE0G3tb4Yx7mWBlgpUJViZYmWBlgpUJViZYmWBlgpUJhX1XMrFSPqhxLxOsTLAywcoEKxOsTLAywcoEKxOsTLAyIU85ynJhdjmemY7f2bh7TDhac9ycX7bCGfcywcoEKxOsTLAywcoEKxOsTLAywcqEPOVHEB+9LJx5ff4s4n7GT3OtE864lwlWJliZYGWClQlWJliZYGWClQlWJliZgH7xcH9SC2fW97+AO0m0urgSzrQ7D+GMe5lgZYKVCVYmWJlgZYKVCVYmWJlgZYKVCXmpOhNP9bqFLPbdnSrgKlP1eOYn4eWQhJMsU55R1OqNeKn4bjQyKxOsTLAywcoEKxOsTLAywcoEKxOsTPAEylF5AuV/YWWClQlWJliZYGWClQlWJliZYGWClQm/lZ86jgplbmRzdHJlYW0KZW5kb2JqCjU0IDAgb2JqCjU0MQplbmRvYmoKMTYgMCBvYmoKPDwgL0JpdHNQZXJDb21wb25lbnQgOCAvQ29sb3JTcGFjZSAvRGV2aWNlUkdCCi9EZWNvZGVQYXJtcyA8PCAvQ29sb3JzIDMgL0NvbHVtbnMgMTE5IC9QcmVkaWN0b3IgMTAgPj4KL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0hlaWdodCAxMTkgL0xlbmd0aCA1NSAwIFIgL1N1YnR5cGUgL0ltYWdlCi9UeXBlIC9YT2JqZWN0IC9XaWR0aCAxMTkgPj4Kc3RyZWFtCnic7d29ahRRHIbxs9lZNYlJUIwoKgYREgIiimIhWugVeAdWVtpb2NvZehXegZWdIFpYBBtFswnRrB/ZZDebT72E94DwVM+v/jOzPDlNDjNnWvN3npTk0/rvOFPK3zhx88KJOHPhVBgYjPKNrs1PxpnX74Zx5jBOlPL+23qcGau4jv6XlQlWJliZYGWClQlWJliZYGWClQlWJjRfh+04NH08bS6U0j/IN+sN88ZA+1eY6Q3yD759Na+erd1WnFmuuNeD67NxxrVMsDLBygQrE6xMsDLBygQrE6xMsDLByoRmup0fb7i/kPcfNof5Ou+7eVvgz3bYXhjtx2uUNx8248zyoBNnBhUPZOxV/B7XMsHKBCsTrEywMsHKBCsTrEywMsHKBCsTmh97+bGEI5088/xxP85ceTSKM6vjZ+JM1F3q5aGjJ/NM+2gc2dnLGziuZYKVCVYmWJlgZYKVCVYmWJlgZYKVCVYmNE3eoigrvfzSyLOXU3Hm0uxMnJk5Fu61McorY+FyfhHm42p+OGR5O46UTpNnXMsEKxOsTLAywcoEKxOsTLAywcoEKxOsTKjax5iZzEMVlyn9imc/2mNhZrfiII7hTp7ZPqj5ydnURF6prmWClQlWJliZYGWClQlWJliZYGWClQnNqOIf1p3dPDN/8Vic6S3lUw7mpsMLBJ2KhdH9k58CmOrkNxV6FadU9jby8Q6uZYKVCVYmWJlgZYKVCVYmWJlgZYKVCVYmtMriwzy1P8gzB3mz497i+TiztRO2Dr5v5ZXx+dXbOHPxwa040+1vxZma0x1cywQrE6xMsDLBygQrE6xMsDLBygQrE6xMaEozmad28+mSZaziwxRpj6KUMn8u/OFPVXx79OmLu3Hmxlw+lmH/Sz6OYm2Qr+NaJliZYGWClQlWJliZYGWClQlWJliZYGVCU/YrTlmcOJtnDvM7I2crvl3RH4bXNGq+wNlq5Q2TlZ/5OmvD/JDJ6YnxOONaJliZYGWClQlWJliZYGWClQlWJliZYGXCP4QAi28KZW5kc3RyZWFtCmVuZG9iago1NSAwIG9iago3NzIKZW5kb2JqCjE3IDAgb2JqCjw8IC9CaXRzUGVyQ29tcG9uZW50IDggL0NvbG9yU3BhY2UgL0RldmljZVJHQgovRGVjb2RlUGFybXMgPDwgL0NvbG9ycyAzIC9Db2x1bW5zIDExOSAvUHJlZGljdG9yIDEwID4+Ci9GaWx0ZXIgL0ZsYXRlRGVjb2RlIC9IZWlnaHQgMTE5IC9MZW5ndGggNTYgMCBSIC9TdWJ0eXBlIC9JbWFnZQovVHlwZSAvWE9iamVjdCAvV2lkdGggMTE5ID4+CnN0cmVhbQp4nO3dsUuUcQDG8d95v/O6M70OWyryxMpKw83EIoegf6GgMWhoLodaGloahKIhaGlxi4ZCcDBwCTSnAkmXsobII7TrNLw77zz7D3xeCJ7p+5kfXl6+vcu9ve9r6t7kZFCmZtblJuaPys21/qbcDJzoPHjw7MGCPMjwjYtyMzGclps/23ty82rhu9x0yAX+H5UdqOxAZQcqO1DZgcoOVHagsgOVHajsEJe/7CRY5eXkXI/+yV86Ju5RhBDqjfbBg+k3E/Ig44O7cjM6pE/m9vOy3IydLskN17IDlR2o7EBlByo7UNmByg5UdqCyA5UdqOwQ89mUHBVzOX2gDn0fY6emN83W/sGD+m6CEy5k5Obbuj5OSEU56e0WJxy4lj2o7EBlByo7UNmByg5UdqCyA5UdqOxAZYdYOKxDV7Y35OZSqSg3qQR3DjozYnR19Ic8yJPXvXJz/+YhfTbprJzMfq7KDdeyA5UdqOxAZQcqO1DZgcoOVHagsgOVHajsEH9V9DMSIau/fbGxpY9T6NKPLrTFayVh5n2fPMjl8zW5qWwnuMKaW3Jya0LH4Vp2oLIDlR2o7EBlByo7UNmByg5UdqCyA5UdYkwneEhiry4nmzv60YWRnP5H/avePTlzUt+jWF2Tk5DL6psqIaW/Bfrxq76Bw7XsQGUHKjtQ2YHKDlR2oLIDlR2o7EBlh5jR3yYIIa1fDugrtuSmsqUeAwihOy9+1O7v61sC5U19MpmY4Bd2Wn+yotHS58O17EBlByo7UNmByg5UdqCyA5UdqOxAZQcqO8TFNf2f4aH2U07mV/Qfr3h5d1Bu5hbFUwmbVX1T5dEd/TWF6dkEX25IYKWqA3ItO1DZgcoOVHagsgOVHajsQGUHKjtQ2YHKDvH6uH4g4+lcQW7G+vVtgXdL+s2JYo84n3ZbP//w8EWX3AwN6Cts7Lj+ixyrv+WEa9mCyg5UdqCyA5UdqOxAZQcqO1DZgcoOVHaI858aetXWz1oslfWXG66M6Nc0mi3x7smFUxV5kA/Lebk526ffK3n8Vn+BslQ8Ijdcyw5UdqCyA5UdqOxAZQcqO1DZgcoOVHagssM/sG2GJQplbmRzdHJlYW0KZW5kb2JqCjU2IDAgb2JqCjgxOAplbmRvYmoKMTggMCBvYmoKPDwgL0JpdHNQZXJDb21wb25lbnQgOCAvQ29sb3JTcGFjZSAvRGV2aWNlUkdCCi9EZWNvZGVQYXJtcyA8PCAvQ29sb3JzIDMgL0NvbHVtbnMgMTE5IC9QcmVkaWN0b3IgMTAgPj4KL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0hlaWdodCAxMTkgL0xlbmd0aCA1NyAwIFIgL1N1YnR5cGUgL0ltYWdlCi9UeXBlIC9YT2JqZWN0IC9XaWR0aCAxMTkgPj4Kc3RyZWFtCnic7d0/a5NRHEfxmzRKba00qDQUoVUcFBTEycXBRRAUR9FFnHRzcXB1cxLcnB1cXB3cpasE30GgBGzwTxVp2pqnvoTvhdZDA+cz/3ju5XCXXJ4krXLpcUle3u/Gmefvx3FmZa6JM4Ofv8JE8zc+5MryYpzpr4/iTDman1N204ZLaeenaN+sTLAywcoEKxOsTLAywcoEKxOsTLAyoVPG+eN80+R7jLK3F0cGo2F+Tr6myAv1Rwt5oUm+eCnNdhx5ems5zniWCVYmWJlgZYKVCVYmWJlgZYKVCVYmWJnQKufuxKHJ2kacmbl5Na/2Z71mT4dI+0gcObt0Jj/mIPaiwMoEKxOsTLAywcoEKxOsTLAywcoEKxM6NUOv396IM0+u59cb3nysWW3K9Bbyl2U8ywQrE6xMsDLBygQrE6xMsDLBygQrE6xMqLrHeHT3c5zp3ru4781Mpa2dPONZJliZYGWClQlWJliZYGWClQlWJliZYGVC1T3Giflv/3sf02tzO59UzzLBygQrE6xMsDLBygQrE6xMsDLByoSqT9iD4fk4szJf8QuUv2tWmzKLs37j4XCwMsHKBCsTrEywMsHKBCsTrEywMsHKhE7Njyw+fHEqzgy+DvJqs6crtpSuBSb53y1uXz4ZZz58yX8h0uv24sy1C74pcDhYmWBlgpUJViZYmWBlgpUJViZYmWBlQqd05uLQp2G+6yjHVyuWy68ulNZMGKi4x1j/XrGXY/mOosnvmJR2xUH1LBOsTLAywcoEKxOsTLAywcoEKxOsTLAyoVN2NuPQu2dLcebBqx95tXF+BeJA9LcqLl6a3TiyUVbjzGTSijOeZYKVCVYmWJlgZYKVCVYmWJlgZYKVCVYm/AP421qlCmVuZHN0cmVhbQplbmRvYmoKNTcgMCBvYmoKNjU5CmVuZG9iagoxOSAwIG9iago8PCAvQml0c1BlckNvbXBvbmVudCA4IC9Db2xvclNwYWNlIC9EZXZpY2VSR0IKL0RlY29kZVBhcm1zIDw8IC9Db2xvcnMgMyAvQ29sdW1ucyAxMTkgL1ByZWRpY3RvciAxMCA+PgovRmlsdGVyIC9GbGF0ZURlY29kZSAvSGVpZ2h0IDExOSAvTGVuZ3RoIDU4IDAgUiAvU3VidHlwZSAvSW1hZ2UKL1R5cGUgL1hPYmplY3QgL1dpZHRoIDExOSA+PgpzdHJlYW0KeJzt3cFKVFEcx/GjXjWrkZGykILSEDdBELhw2SqQHqJdBNEr9ATtWrRsFfQW1cK1Fm0KRCIqSNQcNccZZ3qDfgeCLy6+n21/7r1+u5t7OPfOSLn9qCTLc+NxZv7qSJxpxuJIeb0RhhZbp/EgrclhnPm6n69m++B3nCn9P3FkNB9F/83KBCsTrEywMsHKBCsTrEywMsHKBCsTmtI/jEMfdy7HmdnpXpyZu9TkKzrZ+/e/f9mbisdYaucTbXd2Ky5mP448WV2IM97LBCsTrEywMsHKBCsTrEywMsHKBCsTrExoyti5OPRgqR9nFm9cjDPv1w/yFQ3yekj07aidh067eabJayYfNvNxvJcJViZYmWBlgpUJViZYmWBlgpUJViZYmdCUibz+sLaZX9N48/xtnBlbXclXNDhJR8lrC4d7W/lEFWsUNWsdn7fz3g/vZYKVCVYmWJlgZYKVCVYmWJlgZYKVCVYmNOUk75G4eyvv2Xj24l6cuTOTz7VxlJYFTo/jQS60b8aZw86POFNG84dBVubz9zq8lwlWJliZYGWClQlWJliZYGWClQlWJliZ0JSxyTjU7eX9GAvX8uP86SB/C7QMB2Gg4oIPu/nbm2Wk4g4b5hdqdjo5jvcywcoEKxOsTLAywcoEKxOsTLAywcqEJj/RlnLcy/8Z3ZP89DzMz6IVLyKMtyqOUmGQn9TLIP/hx738h3svE6xMsDLBygQrE6xMsDLBygQrE6xMsDKh6pc0OsfTcebxy+9x5un9K3Hm07t2mBjmLQkPl/Pawqu1ijtskHcKtKbcKXA2WJlgZYKVCVYmWJlgZYKVCVYmWJlgZULVfoz1nxVfd7jejjODfKqKjz6OTsRj7HbyWkfVFyj7nTgyfd79GGeDlQlWJliZYGWClQlWJliZYGWClQlWJuRfgSilLMzk/Rizrbx0MCwVL5bEXwSt+HLD1q98ntLLaxQ1b8L0K5ZMvJcJViZYmWBlgpUJViZYmWBlgpUJViZYmfAXHkWJBwplbmRzdHJlYW0KZW5kb2JqCjU4IDAgb2JqCjc2MAplbmRvYmoKMjAgMCBvYmoKPDwgL0JpdHNQZXJDb21wb25lbnQgOCAvQ29sb3JTcGFjZSAvRGV2aWNlUkdCCi9EZWNvZGVQYXJtcyA8PCAvQ29sb3JzIDMgL0NvbHVtbnMgMTE5IC9QcmVkaWN0b3IgMTAgPj4KL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0hlaWdodCAxMTkgL0xlbmd0aCA1OSAwIFIgL1N1YnR5cGUgL0ltYWdlCi9UeXBlIC9YT2JqZWN0IC9XaWR0aCAxMTkgPj4Kc3RyZWFtCnic7d3PahNRHEfxO83YNLYmIqLFKlYQXAhCrQiufQqfwU3Xvo5r8RlculDoG4h/0Vqr1aRJO218A78XDGd1Pusfd6Yns5nbmaQp956WaPwpzwyuxJGtq8tx5td06d8D25vzuMibd00+mZtncebF7jTOlOleHAl/khbCygQrE6xMsDLBygQrE6xMsDLBygQrE9rStHmqdy6OPFjPexQ/jvKH+n0WZiazLi6y3Isj5XCc90NKN84zy6M44rVMsDLBygQrE6xMsDLBygQrE6xMsDLByoSmvb8ThzbP50cXnj/Ljy482jmoOadgKW+YlHk+4Sq9QRxZHQzjjNcywcoEKxOsTLAywcoEKxOsTLAywcoEKxPabvozDq0ML8SZh1u7+Whdfvcka47zTM0+RlNxhVWsM57lh1W8lglWJliZYGWClQlWJliZYGWClQlWJliZ0Jb+xTg0Ocm3869eb1ccbb/ilNKxlvoVi9SoeK+k4ljX1lbyMjWno/9kZYKVCVYmWJlgZYKVCVYmWJlgZYKVCW2bvxGz5oa/XBpOKhY6rVhpEYss6r2Sin2Mz7Nc0GuZYGWClQlWJliZYGWClQlWJliZYGVC2x3nO+Oun/8ZfmvjYz7a2VrNOS3Aou6wz2YLWcZrmWBlgpUJViZYmWBlgpUJViZYmWBlgpUJTbnzJE91FU8BlPwP88d3N+LM3u+wzu31fCrv8490lhuX88zLt1/z0OlJHPFaJliZYGWClQlWJliZYGWClQlWJliZYGVCW+b5BzarHm/o5ZcDDo/yXsfkJHzwo9W8yPRLPuHRWsW7HhV7FDUvhHgtE6xMsDLBygQrE6xMsDLBygQrE6xMsDKhrfpCx16+VR+OrseZecWeySx9L8PBn/zNDYP84xblw7eKzZk2/yJozQaO1zLBygQrE6xMsDLBygQrE6xMsDLBygQrE/4CtZtvwgplbmRzdHJlYW0KZW5kb2JqCjU5IDAgb2JqCjcxNAplbmRvYmoKMjEgMCBvYmoKPDwgL0JpdHNQZXJDb21wb25lbnQgOCAvQ29sb3JTcGFjZSAvRGV2aWNlUkdCCi9EZWNvZGVQYXJtcyA8PCAvQ29sb3JzIDMgL0NvbHVtbnMgMTE5IC9QcmVkaWN0b3IgMTAgPj4KL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0hlaWdodCAxMTkgL0xlbmd0aCA2MCAwIFIgL1N1YnR5cGUgL0ltYWdlCi9UeXBlIC9YT2JqZWN0IC9XaWR0aCAxMTkgPj4Kc3RyZWFtCnic7d1PbhJhGMfxAd4ZQAQRLGKtJrpo3ZjoyjN4B0/gBdx5AlcuPYTHcOFOE2vwT9qkDU2lhfKvhQGKN+A3ickvLr6f9ZORfPtu5nEYcu/fvYmU82FOzkxnKznz6mVPznQOG5sHkvhaXmTn7kTOdHs35czRaVHODKc6Tl5O4N9R2YHKDlR2oLIDlR2o7EBlByo7UNmByg5hOtO34bNUrw6uZks5Ewd9nWpFXGex1CejPyzLmVKiFy9JvJYzdb0O4SxbUNmByg5UdqCyA5UdqOxAZQcqO1DZgcoO4UFL7x8ebevb+W5PP7rw5ceWnOmPwuaBWkXvHw7H+sPUbuilyv2tmZw5u0jkDGfZgcoOVHagsgOVHajsQGUHKjtQ2YHKDlR2CJMrHfrZ7kDOjKa35cz4siBncurxkMFYX2Se6odM5rGeadT0HqPbK8kZzrIDlR2o7EBlByo7UNmByg5UdqCyA5UdqOwQ5N4giqJPX5tyZjTV13m+p9cCw4l4HqNeXciLnJ7r5zFajVTOnJxV5My1flaFs2xBZQcqO1DZgcoOVHagsgOVHajsQGUHKjuEQobOOy19y//zWH+9YrHM8L5Q9XxIutA7iiwGo1jOtO/oxUuiFi8RZ9mDyg5UdqCyA5UdqOxAZQcqO1DZgcoOYanfgxCtM/xneKpvwqN5qv+o8qWPWd4KObnU/1C1ot/cUE70ay0WK35J4/9AZQcqO1DZgcoOVHagsgOVHajsQGUHKjuE38d6LfD9QC8pahX9RYS3rz/Lmf3Ok80D5dJcXiQJ+sMMJ/onMD581F/1KBZ1QM6yA5UdqOxAZQcqO1DZgcoOVHagsgOVHajsEPJ5/TjBcqVv1ZNY/8EGfb0WqFfHmwfWa/2B9w+25cyLpx0506zfkzN/+nqvwll2oLIDlR2o7EBlByo7UNmByg5UdqCyA5UdQvOWDv2wrV9yUMywxzjp6V8EvRiXNw8U8nqpMk/1r218+/VYzrSb+rsncazfJMFZdqCyA5UdqOxAZQcqO1DZgcoOVHagsgOVHf4CQtKUwwplbmRzdHJlYW0KZW5kb2JqCjYwIDAgb2JqCjc4NgplbmRvYmoKMjIgMCBvYmoKPDwgL0JpdHNQZXJDb21wb25lbnQgOCAvQ29sb3JTcGFjZSAvRGV2aWNlUkdCCi9EZWNvZGVQYXJtcyA8PCAvQ29sb3JzIDMgL0NvbHVtbnMgMTE5IC9QcmVkaWN0b3IgMTAgPj4KL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0hlaWdodCAxMTkgL0xlbmd0aCA2MSAwIFIgL1N1YnR5cGUgL0ltYWdlCi9UeXBlIC9YT2JqZWN0IC9XaWR0aCAxMTkgPj4Kc3RyZWFtCnic7d1NbhJxAIbxKQzDQBVpa03UNqbxI26MiYuuXJh4Am/jEVx4AQ/hDVy6MG7VxE1rLEKhIFgKdBg+6g14/0mTNyye3/rNUB/YMM4MG+/fvY2UxVJOosVCb54+mMrNzq1s9aDR2ZQHebg3lJvTnj7O+SiWm8m0IDd6geujsgOVHajsQGUHKjtQ2YHKDlR2oLIDlR3i2VyPmp1cbgoB79ebV125aXTq+kDKVm0sN+2/VbkpJ/oMzlXA38Nn2YHKDlR2oLIDlR2o7EBlByo7UNmByg5Udojv7+oTGWlSkpvheENujpvbcnPaS1cPsly/0El7S256/xK5ef6kLzc/jvRr8Vl2oLIDlR2o7EBlByo7UNmByg5UdqCyA5Ud4uVSnxbYrumbRpJSUW7OR/p8SKEgLnCYzvQnozsoy00p1ldSyD8m4nqM9UFlByo7UNmByg5UdqCyA5UdqOxAZQcqO+jnP0RR1O7r2XCsv9A/3tfPx2h1K6sH1bK+1+Pg3khufv6uyU2e6394PtMngvgsO1DZgcoOVHagsgOVHajsQGUHKjtQ2YHKDvEy4HqCJODShVJRf52fL/SbKp8pOl/oF5rO9MUhISqpfjBICD7LDlR2oLIDlR2o7EBlByo7UNmByg5Udoirqf6P90mm34wsD/i5jQBJSXybXwacE7iY6Fsr4qI+zs3NS7np9PkljfVAZQcqO1DZgcoOVHagsgOVHajsQGUHKjvE3470Rf3HjYncpIl+w549+iM3WS4euvCruSMP8vrwi9x8/f5Cbj583Jebu7flhM+yBZUdqOxAZQcqO1DZgcoOVHagsgOVHajsEBcL+gaCLA+44yHkrohY/2pHJc1WD1pndXmQRudAbm5U9WMk6jV950R3EPAkS7nA9VHZgcoOVHagsgOVHajsQGUHKjtQ2YHKDkFPoEwTfa7jcqq/zi+v9HFaZ3dWDwYX+lcyXu6dyM2nz4dys1vXN8vERe4rWQ9UdqCyA5UdqOxAZQcqO1DZgcoOVHagssN/N4qPuwplbmRzdHJlYW0KZW5kb2JqCjYxIDAgb2JqCjc3MQplbmRvYmoKMjMgMCBvYmoKPDwgL0JpdHNQZXJDb21wb25lbnQgOCAvQ29sb3JTcGFjZSAvRGV2aWNlUkdCCi9EZWNvZGVQYXJtcyA8PCAvQ29sb3JzIDMgL0NvbHVtbnMgMTE5IC9QcmVkaWN0b3IgMTAgPj4KL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0hlaWdodCAxMTkgL0xlbmd0aCA2MiAwIFIgL1N1YnR5cGUgL0ltYWdlCi9UeXBlIC9YT2JqZWN0IC9XaWR0aCAxMTkgPj4Kc3RyZWFtCnic7d05blNhGEbhm/jaGTzEmQQOSAlSREVDRUfLGlgJRTbFEqioQDSIAiGFQQpE4CR2TDzEjtkBryWiI4rz1J+cm+O/ub/vsHR09KJIRqNZnLm7U8aZzvY0ztTXwszycvyM4sHeRZz5/H0jznw4Xosz40k+ngUOWf/MygQrE6xMsDLBygQrE6xMsDLBygQrE8rt1lIcGq3mPYpp3upYSLdX+/tA73deGdfTzThz3q/Gma3WTZw56FzFGdcywcoEKxOsTLAywcoEKxOsTLAywcoEKxPKxno+Vb+zlTcprkaVOPPzIs/M52FfJW+7FMXHrytxpt2Yx5n9zjDOXF7l/RDXMsHKBCsTrEywMsHKBCsTrEywMsHKBCsTytkC11E06/l+kLKStwXOL8O1FkVRxE8p815IMRjmg9ls5pnJdV6FtTJvBLmWCVYmWJlgZYKVCVYmWJlgZYKVCVYmWJlQLi1wfcPxyeqt/LFWfYFT/tv43ivL+b9qN/MOznSWj+asnzdWXMsEKxOsTLAywcoEKxOsTLAywcoEKxOsTCjn+aqE4t5ufpblaJxP509+cc/ZiE7P8sEc3h/Fmc3WOM64lglWJliZYGWClQlWJliZYGWClQlWJpTxSQlFUXR7+eEEN/kigGKllk/n19P3Xi3zh/QGefXstvNtHOMF7ng46+e3bbiWCVYmWJlgZYKVCVYmWJlgZYKVCVYmWJlQfvmRtwU+fcsvi9ht5x/enz/LP6o36mGmP8i3Xzx98jbOvH7zOM68fNWIM+OJT274P1iZYGWClQlWJliZYGWClQlWJliZYGVCWavm0Ku1PNOo52s2hpN8O0M1PdBxdpMvIHn3/lGcKSt5/2GjmTdnhqO8EeRaJliZYGWClQlWJliZYGWClQlWJliZYGVCuZGvOCge7q/HmcoC39dpN7+o8yK9baM3yI+I2NvJ+w/dXn6tx0En33tyPc37Kq5lgpUJViZYmWBlgpUJViZYmWBlgpUJVib8AcrghKgKZW5kc3RyZWFtCmVuZG9iago2MiAwIG9iago3NTcKZW5kb2JqCjI0IDAgb2JqCjw8IC9CaXRzUGVyQ29tcG9uZW50IDggL0NvbG9yU3BhY2UgL0RldmljZVJHQgovRGVjb2RlUGFybXMgPDwgL0NvbG9ycyAzIC9Db2x1bW5zIDExOSAvUHJlZGljdG9yIDEwID4+Ci9GaWx0ZXIgL0ZsYXRlRGVjb2RlIC9IZWlnaHQgMTE5IC9MZW5ndGggNjMgMCBSIC9TdWJ0eXBlIC9JbWFnZQovVHlwZSAvWE9iamVjdCAvV2lkdGggMTE5ID4+CnN0cmVhbQp4nO3dT2pTURxH8ZvklhpTSq0QlCIIBetMKOiko66hC+liXIALcOAC1IEDscWh4ESoVB0aipRoS/65A78XfBxEzmf8497Xkzt6eenr3Xp8XJKLyVmcabG5vRtn7g4Xfx64WvTiItcNMxtrqzjz5umXOLNzFC64lNKPE/p7ViZYmWBlgpUJViZYmWBlgpUJViZYmVAvfl1hmz3ZmcWZ6XW4BfF9mjfaG+d7C+eTfMKev9rLm5WPccKzTLAywcoEKxOsTLAywcoEKxOsTLAywcqEWuoQ2+zmen5MYrEMj0mM1vJGG8O80bDheYwPn7q5yeNZJliZYGWClQlWJliZYGWClQlWJliZYGVCJTdbLjtYpJdvUZRFw0Yt6wwGDUMNPMsEKxOsTLAywcoEKxOsTLAywcoEKxOsTKhllR9L6Eq/i8+05Xr7DbcfwL/bs4ywMsHKBCsTrEywMsHKBCsTrEywMsHKhNr0XMK/pKvrbVlnNu/mZodnmWBlgpUJViZYmWBlgpUJViZYmWBlQq0Nnecdbdby5Tz2Bb5PCvxvrEywMsHKBCsTrEywMsHKBCsTrEywMqEe3stvt3j5Y5BXWuWXV5ye53XujMI6teFa3n3OQ/e38gXvP7wRZ569ztfjWSZYmWBlgpUJViZYmWBlgpUJViZYmWBlQq88Os5Tl2edbLa5vRtnbq+H/x45nXXzk4eWB1G+vjiJM4ODcZzxLBOsTLAywcoEKxOsTLAywcoEKxOsTLAyoZbZJbbZg638C5V5egnGcpVPxniUX6Ux+ZnXeft+P86U8i1OeJYJViZYmWBlgpUJViZYmWBlgpUJViZYmfAb7NJYXwplbmRzdHJlYW0KZW5kb2JqCjYzIDAgb2JqCjYzMgplbmRvYmoKMjUgMCBvYmoKPDwgL0JpdHNQZXJDb21wb25lbnQgOCAvQ29sb3JTcGFjZSAvRGV2aWNlUkdCCi9EZWNvZGVQYXJtcyA8PCAvQ29sb3JzIDMgL0NvbHVtbnMgMTE5IC9QcmVkaWN0b3IgMTAgPj4KL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0hlaWdodCAxMTkgL0xlbmd0aCA2NCAwIFIgL1N1YnR5cGUgL0ltYWdlCi9UeXBlIC9YT2JqZWN0IC9XaWR0aCAxMTkgPj4Kc3RyZWFtCnic7d3BSlRRHMfxa3MdzZRCSgZzpIRaREF7qXWbHiHa9wg9QYsWLXoBN23btGhVi8JFmzCIItGgsAwVMzDLmdHeYH5ncflC8P2s/5wZv3NWxzN3RqqF2xVldmYhzjy+d2L4wMMnh3GR5dWNODN4/T3OtG5240w1+BtHwp+kRliZYGWClQlWJliZYGWClQlWJliZYGVCXbVP56nDvUZerDM5iDPftseHD8xO9+IiV+by+cP9R5fjzK2r+3Hm+cqXOONeJliZYGWClQlWJliZYGWClQlWJliZYGVCfXdxOg4tvWzmHGP3IH+o71bD9Yaf+8dxkamx/GY+fP4dZ+Zm6rxQAfcywcoEKxOsTLAywcoEKxOsTLAywcoEKxPqpTf5ekNTjvIJRHX9cjiD2Nzpx0XWtvPuWbzWjjPbu/k7LCXcywQrE6xMsDLBygQrE6xMsDLBygQrE6xMqG90c+hXH5t5sXYrH2SsbxwNH+gP8iJnJ8IiVVX92MlnFCfHRuJMCfcywcoEKxOsTLAywcoEKxOsTLAywcoEKxPqyfA4iiYdH+djgV4/HEEM8hFFNVJw/NDr5/OQifFmdqF7mWBlgpUJViZYmWBlgpUJViZYmWBlQr2Xn1/QmJJvPLRHwwdf8r/7Xn7SZVW3St6MNwX+H1YmWJlgZYKVCVYmWJlgZYKVCVYmWJlQL3/KPwTRlM5UPl+4NB+ODt6v5xd6+3Urzjx9kK8cXLhT8CjLAu5lgpUJViZYmWBlgpUJViZYmWBlgpUJViY080sRhVa2RuPMfOdg+MDWr4I7EoM/caQ7t5nX6Z/PMwXcywQrE6xMsDLBygQrE6xMsDLBygQrE6xMQM8xuqfyfYy6Fa5JdM4UnGOMn4sjz15czOtM5ZFqby2OuJcJViZYmWBlgpUJViZYmWBlgpUJViZYmfAPsrJohQplbmRzdHJlYW0KZW5kb2JqCjY0IDAgb2JqCjcwMgplbmRvYmoKMjYgMCBvYmoKPDwgL0JpdHNQZXJDb21wb25lbnQgOCAvQ29sb3JTcGFjZSAvRGV2aWNlUkdCCi9EZWNvZGVQYXJtcyA8PCAvQ29sb3JzIDMgL0NvbHVtbnMgMTE5IC9QcmVkaWN0b3IgMTAgPj4KL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0hlaWdodCAxMTkgL0xlbmd0aCA2NSAwIFIgL1N1YnR5cGUgL0ltYWdlCi9UeXBlIC9YT2JqZWN0IC9XaWR0aCAxMTkgPj4Kc3RyZWFtCnic7d3LalNRHEbxnZzTC4kFrXirEigoioID6cyBA30Vhz6Az+EzOPYZOuzADosYEasTK0VQ1JC7b5Bvg2GBsH7jP9m7K3vSndO0c//Ji5KcfPkUZ2rc3tuPM3evL1YP/Bot44scvj+LM7cu78WZ0zdHcaZ5fDXOdOOE/p2VCVYmWJlgZYKVCVYmWJlgZYKVCVYmtCdn37HFDvbzFcROv109MJ2Fi45SynR+Lc7cG3TizNPnB3GmlM9xwrNMsDLBygQrE6xMsDLBygQrE6xMsDLByoRwabBeG22+OhhPwjVFzYv0tvJmtjbyCfv6cz2n0LNMsDLBygQrE6xMsDLBygQrE6xMsDLBygT0HqNG04RrijhQSvkzzgstl/nhkM0mz9TwLBOsTLAywcoEKxOsTLAywcoEKxOsTLAyoS3L/Gca69LJNxClpJuDiuuH0jY1m8m7mS1qdpx5lglWJliZYGWClQlWJliZYGWClQlWJliZ0JYOF7rmCmI2T0MVVwu9zYrN5JGyWM/jGJ5lhJUJViZYmWBlgpUJViZYmWBlgpUJbVlMsMW6Fe9pN32A3634DXte8fRDU7GZydwnBf4fViZYmWBlgpUJViZYmWBlgpUJViZYmdA+GuR/OnH88XQtix0N8yfvg9356oHxNC90ODyPM+++5R98+PptnNl+thtnPMsEKxOsTLAywcoEKxOsTLAywcoEKxOsTGiPz7kvoaz5QseLF8Ib/+N3zVdN5JlemzczmvQr1so8ywQrE6xMsDLBygQrE6xMsDLBygQrE6xMaMvoDFvszo08c+VSuFfpbYcHNkop/Z2bcebhIL/Oy1cP4kwpH+KEZ5lgZYKVCVYmWJlgZYKVCVYmWJlgZYKVCX8BZvtaEgplbmRzdHJlYW0KZW5kb2JqCjY1IDAgb2JqCjY2MwplbmRvYmoKMjcgMCBvYmoKPDwgL0JpdHNQZXJDb21wb25lbnQgOCAvQ29sb3JTcGFjZSAvRGV2aWNlUkdCCi9EZWNvZGVQYXJtcyA8PCAvQ29sb3JzIDMgL0NvbHVtbnMgMTE5IC9QcmVkaWN0b3IgMTAgPj4KL0ZpbHRlciAvRmxhdGVEZWNvZGUgL0hlaWdodCAxMTkgL0xlbmd0aCA2NiAwIFIgL1N1YnR5cGUgL0ltYWdlCi9UeXBlIC9YT2JqZWN0IC9XaWR0aCAxMTkgPj4Kc3RyZWFtCnic7d1PTlNRAEbx2/ZBqdCiiQgSjRpnRBMTo8YY1+EmjBNnLsKRS3HuBhw51IASB6jEPwQohNdSd8B3E8mJg/Mbf3mFw5v08qCd6c6jkvSeXImbGtdXb8XN18Pp2YPFubl4kYfrbdy82zyMmxsXR3Gz/e1z3HTjQv/OygQrE6xMsDLBygQrE6xMsDLBygQrE5pnLx9UrHbzZjKOk421cEZRSlnZ75w96DeTeJHBfJyUe2tLcbM8yF/w9vd8p3ovE6xMsDLBygQrE6xMsDLBygQrE6xMsDKh+fIjjxaHV+PmcPwrbtp8AlHm0s99NssX2ctnKmUhP9ZRjttwqFJKKUs348R7mWBlgpUJViZYmWBlgpUJViZYmWBlgpUJzcoonwu8/1lxpelRnMzKMF/mNAy6FTdGv2IzPsmb4ULFoUm7FyfeywQrE6xMsDLBygQrE6xMsDLBygQrE6xMaP6M8yMHTcVTCZNu/luObsV1eunn3qm4SMXpQ5nv5U07rfmKB3HivUywMsHKBCsTrEywMsHKBCsTrEywMsHKhCY+/1AqzzFKHtX8Sci5qHmhmvOQKh3/P8b/wcoEKxOsTLAywcoEKxOsTLAywcqEJv5mvpQy6OZ3rMcVbzRPK974xk3NG+OaF6o6Wqh4mqB0mjjxXiZYmWBlgpUJViZYmWBlgpUJViZYmWBlQvP6Rf4PBo+f75zLi+3uL8fN5kE4O1jt5wOIu+t58/ZTPhG5cylOStnfihPvZYKVCVYmWJlgZYKVCVYmWJlgZYKVCVYmNPc3PlTMVs7lxY4m+ejguG3PHvzu5I/AODrJn+RZ2oM4GbejfJ0K3ssEKxOsTLAywcoEKxOsTLAywcoEKxOsTGhevXmaV52PeTPLj0DcvpyPF3rpkzKG/fy5oqML+cDk2iifUawO8ze1VfGsivcywcoEKxOsTLAywcoEKxOsTLAywcoEKxP+AjWxbYYKZW5kc3RyZWFtCmVuZG9iago2NiAwIG9iago3MjcKZW5kb2JqCjIgMCBvYmoKPDwgL0NvdW50IDEgL0tpZHMgWyAxMCAwIFIgXSAvVHlwZSAvUGFnZXMgPj4KZW5kb2JqCjY3IDAgb2JqCjw8IC9DcmVhdGlvbkRhdGUgKEQ6MjAyMDExMDkxMDQzMzkrMDInMDAnKQovQ3JlYXRvciAoTWF0cGxvdGxpYiB2My4zLjIsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcpCi9Qcm9kdWNlciAoTWF0cGxvdGxpYiBwZGYgYmFja2VuZCB2My4zLjIpID4+CmVuZG9iagp4cmVmCjAgNjgKMDAwMDAwMDAwMCA2NTUzNSBmIAowMDAwMDAwMDE2IDAwMDAwIG4gCjAwMDAwMjg2NDIgMDAwMDAgbiAKMDAwMDAxMzQ3MyAwMDAwMCBuIAowMDAwMDEzNTA1IDAwMDAwIG4gCjAwMDAwMTM2MDQgMDAwMDAgbiAKMDAwMDAxMzYyNSAwMDAwMCBuIAowMDAwMDEzNjQ2IDAwMDAwIG4gCjAwMDAwMDAwNjUgMDAwMDAgbiAKMDAwMDAwMDQwMyAwMDAwMCBuIAowMDAwMDAwMjA4IDAwMDAwIG4gCjAwMDAwMDY1NTEgMDAwMDAgbiAKMDAwMDAxMzg1MCAwMDAwMCBuIAowMDAwMDE0NTk2IDAwMDAwIG4gCjAwMDAwMTUxOTQgMDAwMDAgbiAKMDAwMDAxNjEwMyAwMDAwMCBuIAowMDAwMDE2ODkzIDAwMDAwIG4gCjAwMDAwMTc5MTQgMDAwMDAgbiAKMDAwMDAxODk4MSAwMDAwMCBuIAowMDAwMDE5ODg5IDAwMDAwIG4gCjAwMDAwMjA4OTggMDAwMDAgbiAKMDAwMDAyMTg2MSAwMDAwMCBuIAowMDAwMDIyODk2IDAwMDAwIG4gCjAwMDAwMjM5MTYgMDAwMDAgbiAKMDAwMDAyNDkyMiAwMDAwMCBuIAowMDAwMDI1ODAzIDAwMDAwIG4gCjAwMDAwMjY3NTQgMDAwMDAgbiAKMDAwMDAyNzY2NiAwMDAwMCBuIAowMDAwMDEyMTcwIDAwMDAwIG4gCjAwMDAwMTE5NzAgMDAwMDAgbiAKMDAwMDAxMTU1MiAwMDAwMCBuIAowMDAwMDEzMjIzIDAwMDAwIG4gCjAwMDAwMDY1NzIgMDAwMDAgbiAKMDAwMDAwNjcyMSAwMDAwMCBuIAowMDAwMDA2ODUyIDAwMDAwIG4gCjAwMDAwMDcyMjkgMDAwMDAgbiAKMDAwMDAwNzM2NyAwMDAwMCBuIAowMDAwMDA3NjY3IDAwMDAwIG4gCjAwMDAwMDc5ODUgMDAwMDAgbiAKMDAwMDAwODQ1MCAwMDAwMCBuIAowMDAwMDA4NzcwIDAwMDAwIG4gCjAwMDAwMDg5MzIgMDAwMDAgbiAKMDAwMDAwOTMyNSAwMDAwMCBuIAowMDAwMDA5NDc3IDAwMDAwIG4gCjAwMDAwMDk3MDcgMDAwMDAgbiAKMDAwMDAwOTg0NyAwMDAwMCBuIAowMDAwMDEwMjM3IDAwMDAwIG4gCjAwMDAwMTAzMjYgMDAwMDAgbiAKMDAwMDAxMDczNyAwMDAwMCBuIAowMDAwMDExMDU4IDAwMDAwIG4gCjAwMDAwMTEyNjkgMDAwMDAgbiAKMDAwMDAxNDU3NiAwMDAwMCBuIAowMDAwMDE1MTc0IDAwMDAwIG4gCjAwMDAwMTYwODMgMDAwMDAgbiAKMDAwMDAxNjg3MyAwMDAwMCBuIAowMDAwMDE3ODk0IDAwMDAwIG4gCjAwMDAwMTg5NjEgMDAwMDAgbiAKMDAwMDAxOTg2OSAwMDAwMCBuIAowMDAwMDIwODc4IDAwMDAwIG4gCjAwMDAwMjE4NDEgMDAwMDAgbiAKMDAwMDAyMjg3NiAwMDAwMCBuIAowMDAwMDIzODk2IDAwMDAwIG4gCjAwMDAwMjQ5MDIgMDAwMDAgbiAKMDAwMDAyNTc4MyAwMDAwMCBuIAowMDAwMDI2NzM0IDAwMDAwIG4gCjAwMDAwMjc2NDYgMDAwMDAgbiAKMDAwMDAyODYyMiAwMDAwMCBuIAowMDAwMDI4NzAyIDAwMDAwIG4gCnRyYWlsZXIKPDwgL0luZm8gNjcgMCBSIC9Sb290IDEgMCBSIC9TaXplIDY4ID4+CnN0YXJ0eHJlZgoyODg1OQolJUVPRgo=\n",
+ "image/svg+xml": [
+ "\n",
+ "\n",
+ "\n",
+ "\n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " 2020-11-09T10:43:38.742709 \n",
+ " image/svg+xml \n",
+ " \n",
+ " \n",
+ " Matplotlib v3.3.2, https://matplotlib.org/ \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n",
+ " \n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {
+ "needs_background": "light"
+ },
+ "output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "Probabilities:\n",
+ "Image 0: 0.06%\n",
+ "Image 1: 1.63%\n",
+ "Image 2: 89.63%\n",
+ "Image 3: 0.01%\n",
+ "Image 4: 0.01%\n",
+ "Image 5: 0.01%\n",
+ "Image 6: 0.01%\n",
+ "Image 7: 0.01%\n",
+ "Image 8: 0.01%\n",
+ "Image 9: 8.63%\n"
+ ]
+ }
+ ],
+ "source": [
+ "visualize_prediction(mistakes[-1])\n",
+ "print(\"Probabilities:\")\n",
+ "for i, p in enumerate(preds[mistakes[-1]].cpu().numpy()):\n",
+ " print(f\"Image {i}: {100.0*p:4.2f}%\")"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "In this example, the model confuses a palm tree with a building, giving a probability of ~90% to image 2, and 8% to the actual anomaly. However, the difficulty here is that the picture of the building has been taken at a similar angle as the palms. Meanwhile, image 2 shows a rather unusual palm with a different color palette, which is why the model fails here. Nevertheless, in general, the model performs quite well."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Conclusion\n",
+ "\n",
+ "In this tutorial, we took a closer look at the Multi-Head Attention layer which uses a scaled dot product between queries and keys to find correlations and similarities between input elements. The Transformer architecture is based on the Multi-Head Attention layer and applies multiple of them in a ResNet-like block. The Transformer is a very important, recent architecture that can be applied to many tasks and datasets. Although it is best known for its success in NLP, there is so much more to it. We have seen its application on sequence-to-sequence tasks and set anomaly detection. Its property of being permutation-equivariant if we do not provide any positional encodings, allows it to generalize to many settings. Hence, it is important to know the architecture, but also its possible issues such as the gradient problem during the first iterations solved by learning rate warm-up."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Your turn! 🚀\n",
+ "You can practice your cnn skills by following the assignment [complete the transformer architecture](../../assignments/llm/basic/transformer-architecture.ipynb)."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Self study\n",
+ "\n",
+ "You can refer to those YouTube videos for further study:\n",
+ "\n",
+ "* [Transformer: A Novel Neural Network Architecture for Language Understanding (Jakob Uszkoreit, 2017)](https://ai.googleblog.com/2017/08/transformer-novel-neural-network.html) - The original Google blog post about the Transformer paper, focusing on the application in machine translation.\n",
+ "* [The Illustrated Transformer (Jay Alammar, 2018)](http://jalammar.github.io/illustrated-transformer/) - A very popular and great blog post intuitively explaining the Transformer architecture with many nice visualizations. The focus is on NLP.\n",
+ "* [Attention? Attention! (Lilian Weng, 2018)](https://lilianweng.github.io/lil-log/2018/06/24/attention-attention.html) - A nice blog post summarizing attention mechanisms in many domains including vision.\n",
+ "* [Illustrated: Self-Attention (Raimi Karim, 2019)](https://towardsdatascience.com/illustrated-self-attention-2d627e33b20a) - A nice visualization of the steps of self-attention. Recommended going through if the explanation below is too abstract for you.\n",
+ "* [The Transformer family (Lilian Weng, 2020)](https://lilianweng.github.io/lil-log/2020/04/07/the-transformer-family.html) - A very detailed blog post reviewing more variants of Transformers besides the original one."
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "### Research trend\n",
+ "\n",
+ "Attention is all you need; Attentional Neural Network Models | Łukasz Kaiser | Masterclass:\n",
+ "\n",
+ "\n",
+ " VIDEO \n",
+ "
\n",
+ "\n",
+ "The Narrated Transformer Language Model:\n",
+ "\n",
+ "VIDEO \n",
+ "
"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "metadata": {},
+ "source": [
+ "## Acknowledgments\n",
+ "\n",
+ "Thanks to [Phillip Lippe](https://github.com/phlippe) for creating the open-source course [UvA DL Notebooks](https://github.com/phlippe/uvadlc_notebooks). It inspires the majority of the content in this chapter.\n"
+ ]
+ }
+ ],
+ "metadata": {
+ "kernelspec": {
+ "display_name": "Python 3 (ipykernel)",
+ "language": "python",
+ "name": "python3"
+ },
+ "language_info": {
+ "codemirror_mode": {
+ "name": "ipython",
+ "version": 3
+ },
+ "file_extension": ".py",
+ "mimetype": "text/x-python",
+ "name": "python",
+ "nbconvert_exporter": "python",
+ "pygments_lexer": "ipython3",
+ "version": "3.10.4"
+ }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}
diff --git a/open-machine-learning-jupyter-book/llm/image/attention_example.svg b/open-machine-learning-jupyter-book/llm/image/attention_example.svg
new file mode 100644
index 000000000..45fd2897a
--- /dev/null
+++ b/open-machine-learning-jupyter-book/llm/image/attention_example.svg
@@ -0,0 +1,9628 @@
+
+
+
+
+
+ image/svg+xml
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
diff --git a/open-machine-learning-jupyter-book/llm/image/cifar100_example_anomaly.png b/open-machine-learning-jupyter-book/llm/image/cifar100_example_anomaly.png
new file mode 100644
index 0000000000000000000000000000000000000000..7e06e5a5d24157731d1976feca48d9bcb492e50c
GIT binary patch
literal 155392
zcmcF~MNnPg((QqRySqd1;O_3ho!}na-QC?C4(<-Y-66O;!QGudckoZ{o4l%5uXcA2
zy1v?b@~!Ib)oVv6%1a``;lTj_07PjiF=YS%_|E~&Fi`)ZTHJI#0D$J|sjBI$Z0JT}
z?`UUcX>Cg4>|t+8V(M;b1^~FPRcBZ_l5;19e2vh&!mJP=5kOb2Zsf;pZ^N@RvQ+*u
z`#q)2B8`C?4cN%*`SzUc_c`P9*{3m9%s
z_i^!JzV`GZqSSfxsBUNfjuR=jRdDNiAJ5_Q^L@U%7WBwJ`xHV3oO=1t_eCOGiDrHj~4^toB62lHj5=xAnf=HV(TZ}=(vy;R#b&d>e2
zl~L^NZLjCs=R?1sMzXJ;;iCOlL(bJqSI4H8KnH%Y^6aa)s=KeMOw{3dUH^gC5D~g~w*tMx2&x_5)ix%!Ed2?^k(~_P->3Fe{)eW9SJit$
zyW5Ss=@`8l=iF=$o{yih-nHx7XVoihx3L77WK>a|fBimuKWnkSlCF-niqVF!>mRBy
z&tyaxCKQkEvx+RFI=8!5b#{oiw;#jw{foyK2ufHuIk@`YvyHUhfjTc2PKR9kSLU>P
z`)^+0GK0-ALrmb`Lh%}h@PbMYn3YE)f}rRW+CkFLg<5O$EX*YA{JA)|HB(-E-6NE;
z>?f9O5fV7_2lpjM!i_S_*Zai*1%|7U8EQ^eaW44m%gWVDr)m_Od|nQBrCxVb>w4&rto-hft4p!2DD#;1ULrehG5A`pJsbvc^XmE*Gs9`d^ao~Q
zxl&G+Q^{)ktx9KFu2l)LhNx?DJTs_TYt8rm(Q&z(DzkKd;z-&iRbctJal_+}RI0OQ
zN7Qx;i>%aHwz9b?m`zShn}D^6Ex=)^I0Bn`>NjLw^7y3cU;9s}Z
z@4h(A7w4`5AvDJwl^K7&R?NgR#ZGx3@
zlqUFUnZ{(7^?54X-}XS&4ZR>kEr+U9#wB*K*zF1=^=>AJkRKBnp29WbhS;AC%R
z%A_Y6=rUeR{00^Ib@C6D#2jz0?a24L1^FW`i3!)SR-V!lK9~vaf}nSnlIfv-_%|QP
z%*>1Sl12ASLre?tHq&X9-g#iv2>gkV;k?2(waGM)b!A}$w=2DAIh}h
zsHhED9RI4evHN`q&lRAen7An0d+1~oZ*l)M`neiKWzTm^$F7m;G672he$hPFaT4
ze!o}i&0s=dwO+hn7X5{m*qUejUfr^LvXX{}svTzB6i4j`L0;tw27B?)M~`nj{}>T-
zq)f-k9BLy2Pzc2h-|>nXEjX5+bVawaF!Sh*b1;-h`E_N
z4mXFwX1B-HTsmA74iR=GyiydA0&J+uU4QXj8AGaJ^_r>Mb2JP^;(Vy%MCBS_yJ<50
z@oOEAKpUJxb~iE`_n**m?4z3I4RfkkM>gT$a5S+>>mD$%7||jn5)1S>$w~7_zI^
zid_a=*rty5F5{K%>i@}1K&=ST);1u=)bfwyjdoIAF*US)^vDO`vK$(v-N%|w@-Of%
zIOiW-eFH;xh2}9B)hiJm><8zuHI7IaZ;URKALIMG{L77IAcmqJoV{^Rc+0&6aV>ox
zfPQ1rh6)*By-H%FNFTmGD@W&7&Nwk?z-kS0M#@tHc!)Y!)u10{q&+5K^ouC_28!XT1-mR^-CVAP;2
z35;+{m`)5iJEP@6io%6K@zsDibDKMR!8xCbpfilF7l-EbM3fN8JVV&9zGc4Uvylw-
zgoe@on#BC=b!Z2rtr>k}c0;cCkcz!DLR!ac@KaufRjFWNEq~PqxD177^CuPx+(dkj
z<>Qj`P}
z$+{h%^q^tSpp2jPNuBcaC_nsizzT67S$}eg?`Gnn4@vP+DL9o7ex<@<;&0L1j#e3c
z=WdXfsgY~~q+!!@{xm0BRR6o71Deyi;AAB6+`X?d22YH(=wZy(j@13Pv&
zg%sCMN1Jw3G(C)iO$-~XMN;3Enx`5qHAJ2^J&u3pQX;thX>AgV^hQ({)3Fjwg&iiQ
zgs#l91QVFi
zJO3kNdO^)5)l+FTw0C2+Xv~IG&4Y{?OlvI`HrRjs9c9rBXhFpG>
zK$amf6P8hZCF*W2%qt^m3;eell5SHfJhgQ+%;SJ#|3m>~dR+i8%*Xi%5wK6_E^4j@
zGrL>mzu!A4M$L$iX=^V7>ivx0p*4V^J6e_{y@=kvjK?VFjRWhwasLCDC&FQNPraY4
z8i-}GRKK$4jH{~d8$HMwvbZ`EA>aoWL0QwWy+>k-zvB@mWHF^4Q;%Mg@+E{1@$oq@
zwIK7&y9Q43wigm5!jk4r2|uBbOBJ!1Ft1HuC|fVXN}F>|#t_!uAgq>X?s6*|4ny+X
zcKIkuGKf>k*>l6(ho5DkR>GnuaYS}ds5=MnuEPfeTg_G?b?pI74&p8)a7PQ=i=YRE
zn=)5D`A)+@aZ~Ip@DJv7I9SnG!V2F*ryI<44=1MZBSUG2dclbc!QCYBk-(~RFPL|jZ&BK_SwN@ut5wA!N_q-$SNg#mzZn$>{DV$
zjqcl|kkT4d-^QF9W=hB3cxl`*Pn>M>?<*K>JO0rt={!2|40Zk$T}anLNal;^9go3_
z(Je(&dytSZ>-%1h09Q&)C{VIr_tG5MX++{+a!NPNhlP5_&t
zKMRF&?OGf6z%@0PtBh4uY+c}?%IBwB!V40
z9*AokNLK{OUP*-Y%y5Cv-Z9H?55bQLvu6NTeZ%xE#-`dMnxY2O<&IQ`Eum~Np&xkf
zwP>*-bs>6Y^eH!N&MhR3DLYd!ZyhcbxaqDVcc{v~3QU<{JPAJi7;;$o(6Ck;FlFB8?
zq#MtDB`M1ghIGA)Cs}1ihi3-8lBtCEv{E4=jCRLR3Ktz|fA5v+P0pEPC5fISeH@V
zQB;V{c>aXZ(ugY&9tJkgMu=q_20}*-u%8G_uhh{>G~Yy%%j1xv(CiQ25;YvoR>yra
z>?J^}93Nf4CAS7+=EG$Qv_H@5g5O@c`fg~cA=9QDJgcijKH_}iQr
z1Qf4jOsD9&!>Iid=~07zgsp$(Nx5-|fv#9^)iU+|Gkzd*
zz;|a3K+J%-A++dPov|D$I8K48vRInVPkUUI7sk;oS7;C
zC8O7(L2^pTNk;7l_*0C#eR-Q1${T{v`*XFno>`YVSd$%>O9gzqs6{aPl^}{cDyrvI
z!g@_<%)%>$F|7>VN>gms;nRHg_ZB3zvmH~DVsJ1js7hvfjqz?&6WLXJ2F{^Ur6&h)
zGk+SRw4A>xhd{@K#m70aHtE2yIf5cPd&^+p-`u~mTirp{vBQLHeBW&-&rv29=D|Dk
zHn7kk7|6c|eBC(Idi!mC+x9h&1ZsO@Z4AkbZC7^Z&Fh^{4~g6fal(wby;r)ObxCol
z88q>DAg>0>!b+EPjNEbPw#Hj{A4hf9h*d2gk+0ZB>yhfHc(DL}mf*)oX{5-aT~C-2
znhZ$pgC0b(>~T7J=>r|;XAISL6lbUZR>9dWUY8^ert*0%R6wvO
ziRVkTxEV>MiT8*>VIDr?b6^WOM)5S)W@dfHvNoFSLYN!q=M?Qoc&`6E)___-No$5f
zGDs?h33wuptvlhqS#lqAfvt8i@o^6xJrNM7InV|6?8y6MZ|C`eT_fRKVHa~$0svqk
zmZGAH(xRgO%Rm7D)CoQb{89r#1Vj2N#o}~?=Olaa1(D&9N~n_6BB->%+68LRNwR{$
zl!($%QAIo2_SeSsr#pL9l|hvxFuwj=+&nbDr6Rs$;l3O3ab1pdc*S+N`71IqPo3fb
z*}9BM&>WbPbWwq|#}X)p(FU4ksmyfJ((bXfUeC{-*!x;rN1p`zzow1+6)18IL$yft
zsB+h;aI&8w$g8L;2ag7eD0SnX6V4(h?6TpJ;Q3gMN>rQ%QH|>V49nm*7M7zbph>Fz
zDD@`YZy5K8YE-XiIOKnU7v=Wxh&W?|j_(D>A{7n}q~%Z)fn?TpQKko32wv`9xxJ&c
zV>sIQCm0i%TEr6yJ`=xRE{!WhcSsj-{dqB3Ck_%az_wVhetTPK!Kd>7!fl(=`E(>q
z|D@s9Lv$Y&2vTq08^W!$q@FwP!#Lob)%7hkZj;_{R)F8OchD0lxn8L0>4KA2T9^hMoG!?N978mkbt4Fx!qT@sijCuv}ItKUD^_vsB5S_
zb-6+1bFP@EnI`tZAiMFp#u5T^_qqRX>@zXwG|KON#Bbr~#69-tzYJ$Y@MB(Z_U7u@
zL;L@Ho3Axrz=9Ls$bT8S6#svxoc=!?9Ql9x&U**Wi=>~h+D*^9!WZjSw=X0B#veR{
znnY3RO*-(G-=^Nl?nn2XQv~%VtQa6%=yOd!U?49p4-p{q;Dkx?=B@SsR4St#{XDsr
zGY~`)`RIH`Z1N!hBdHm&iF;lXgo6`m>IqdAB*yD`d;Wf^kg&$cx<12%p6;Iwb2JgX
zJQ0{S5uG*>IByA+3`Q|9lUa8aIHSLP{r3#NyR&QmE{zig;*7^w&z!y#c7jJVLNT&1
z+<-BXEYb%3yyLmiXdBXLJl&~3lWKBRFX>07>xJ5f4@e*?CurC6!R}4}tMv=hbTr@)#s97YwAc1{mHL7(;3%1lG`Y*LliRD4lzW;(4Y
zOlU#RioZ2XTIEYmEi40gag-h|oZTz;nO&EwPG5>ydKdZ7wC60af9p6rF#*X~i{06o
z?I4iz`f>QQ^Ke1v_j#9h!}xW}_?+hVg8lxI_Z3~M-?Qe1wB2=R7mh22ke_>YJ`yjS
z911D2o`G8S`f{^;;~T!6xAtREUg`_B$L?G@%&$kE#rZjh=kxRGi}fA(Yv1on_wLs9
zE%NK*>scSUQtwMWPF?Kwi+QoRrkfIFo%O9>sJm`~9o0qPrN3y<{|gDAsLO=pru(yJ
zxM$&-E!WDY3F_4fPkp!Bw!!>*ZM*!@MZNih_pN#ez`}aSXE?sPl$y-^H`=nH+3Z$Q
zH*Y_;J}N(av!{dhQQ027PaUpB*`5wEt5$r7xC17VR_uZY_Vvsc5?Lb}$^4G1nJs(d
z_fgs7$B6#AXMyW>*~&K1^~)lU(?K>`mmJ0U!N<*|67Cd&Wa3lLd|@7ivdhM2WgH-Wg`
zZM+Wm*k!N||5z<#cotHP3o90#43q6Y2{HZ1-C__fN%^Z$jtYgv9~Oj5vJD&~u5-!d
z4xk}bYG5B-zIgltzDO8J*o$rdGz>cV)rDP01gIbeI!N$mK{GAooz1X
zA^f;Z%PV&NCR4x6K@~Nn_1$&rtQsX3j3lK;GSTn3#^WWg<}3Q^?(FNZBV@`YG*G$^
zLnHTfyKN+4L`4~*0e;+erh%?0oZl5JvdO3ElWHvwe(=gCPDQn#P+$s!M8pV+*<+(O
zJvG`ghq;>da6Z^FqnA4XEdVX(Lu+t*AWs|+{4wV%lmH`zKmnfaJ-%cxx9qJJH?(Mc
z!_hN(iS75}+C%u<2?pgm%6IC|D(dd;vf6z(%Hd@512#h$bjQr^ww7Myn{F=pMr0_B
z7fKJWiW_5Z3@gjZq1B%p*bzABKhl*Y4x=Nem`05xXgI=TZ>-2sdbA7YVxO~#H7Y-o04%T;3=)sJ^E7ND7S0^f89o8*4YJKd>}5@>R*38AnFVH1(XR*1d_3;
zVgq=Eg1d0SD`wUyca`d{BEOp^(}>Vs^khl0l7v@<&I6D$d{BW+u8Y
zI`wNWcv6)WAB|>^azV*ZpnsmLEc2LMid0DOlgY^|?op5IVy#JB7vjy;1;xdMqs~zkojqg_B~~aBNvO%(s~cWQ3DJnx+HPh|bfbH^A5F6g
zi>W0gzsnRn!e1?3H9mh7!h}o?Vp1+0jkk_CK$WcT5*ZEtT?{@L4$hSc4hCEV0nTUZ
z^&v{=lxzLems2bz?%I4b|HKfe_;5TDkX(%mm%uXl@S2a3g6tn@0Ym5KWNQ?5d>Hm!
z5rg{T5g{NRI+42Nundd_9vqeytE`?naZO8ACU^DDL#yu#0HT+oum;!wwgK%TCs2J*
z!ostJ9p@_`cY~BQR;555Me?j`UpoBrYJ+j*O2|KBHT97fewcFQ)xucF&kxmOurNZd
zMCaZZ)UMDft*)1mw$z^Dv-qxE!fNr{Umrmth`Jk(%Xb9jA2x4|qIGJCkFn>7B)j4A2ibkJq4BeT%T^j4f
zX~m5TqEF`L*TdEK3wCxCQ;R}MPM8STOq}sOK@D&VTBwRqea>UAXNqQnEwkp}hxyCg
z9v^P|z=##aOq&da5z&$Y}0m)$)+)sG%#+2cBF!B3ml40xCKZ?Ca;lAag^@55*!Za1*AGxR8u
zOYd>%S@*}AJs!G#Z<}A~Ax*VzYc#JoDj%`%U;W}^(B0P+hScmP#Pj16Mo{?vwo=jP36*VOyvluV7Fa1&
zz(GC@D|u;f_IpPA7}KvMMb+PIz}zkvL%H{9jE2NuX51|-b_G7RawBuc?@?!d(-{e@
z#+hg(lj%pn3bk~;FV6jpdFt=Cp8bCt0OPsU4G0St>b+wPkmB%LcAV$M{^%OfP&7uP
zldR-peQE9-4axBzZqhO-IheDQQzxAA@qN!f#6ZFgx0;2I{Kg$N?r*n4tlB&>Z(`Xo
zj+S{VXF2G~Yd_{%&68f*H)N_Fn);C#4rNVBoCM)g#OB~AL`%Kj`L4Ma*9
z)Q4z6^Fav@-4Se9J03%+@E>V>D>CCU(YAn)jR|oqha6fyo>R7^H%87Sw#i=zG&zwo
zj8`c0CDB@_a}BuEr0qwc3kgvyP)$MZt$|aAjPjSU+=jK*QluTyYW01mT&G*fvDy4N
z?grn`owdELv_+L$Or%*({)K+NqR=N!X;^QYVA(t*cnSBm5
zw7FUrGb@Wct{r%%sMRoG=Y&&Xw0*2TI^6g5kf;k6M`$er+cFB4|Zk2{>`Atw`T
z=3$?Zkr2VG0P8fZY34P_?G>-1IDRSI0(n{7tG`aUpdpAyNkC}oT_9A!&?<*>)+_ZR
zE~+AzWYVd>!K7*;*khWj8vQ>gs_n_WF4doYewfE)WQ133w{fnBemX;nw=vUMy?*#0
zTp0xg)cAcT`)(=3M3N%($GVAIrPPmoR2jiOeA&VIu+p5GE}4C
z>r;PphnqI>pVZk&u&I_rL?pP*^yt_O>BCw%?N`Ule#{BN#evhM&P{6Np`q
zk;2%D
zT6^&CDL{1J6gM7ZUXmlsa841l7R18D*!-YIP^Ecf$h+}Jr$woqKVd7#adOQ!Y-l1Z
zv_8Xd%TEFUjf3%MW)V9ssnsg>drPL2pt$NMwpNybJk$tfMwHNDU~&OMC`_>7yd?E&
zOAIp{?GDgtaJQb!mOB#CuWRUcIBJ^(>=Lj5nI8g}UzG@TBGr_X8-j$eCm$PlkMgqq
zXb=8*WcjK9BViikL{2twxF%vc5lHE{l);)(O%cgvXOwSxvaA~$Y=9x&xpQdnr3%mq
zX-ktQBOH4;9>r)yz|%bRy+{1)uPROkN2;?V?N?M;A#Hh4=)aMBb&Mg4iNe)H5xP)P
zNF@trFaH!1o}~ySk1uS2Gz7gF+dA})(DaMN^A{@xi0253pi~@=V)RUBJwCHbJs34T
z58|NkIUqL#VkFqxJD~8rdlpa@{_0&D9I|LSwwr{!GyUb0j{UHu)Fny!MHH|w5~2M{
zNv>uE6i9A+SIl+-m6=L?L!m*_LA5d)PR1$q=(o&q5SZesW|f~XF|b^>U*|l33$BPV4QU7IF?QQv)gAZCD2um589}
zkOZ^@>KDVf(n<{~%8SNmZBmb%t<;8`N($|)@4I&Fou$oOJR$hsdj^#WhJNV+WvhwC
z!Bi{obSwW_!rQuEa+6%(E3AHxIx?VTsOdrq6WTi>NzY|7CHcLe*aQW2JX8Y%mf;E|
z>Zjr?J>cE=`**NqZhaN&()>Sc!T-b7_3H$ko?zGB!#q{CJl&GZ&`Hx?Q%(9&yXByj
z>kLk{wpFU1oiX)AH>zrh@CV<2`u$8zp&xub_;oGa|7eJfexTfV*cE2`5L8VyJPH~LC*q`?Co
z>B*ZayCZ6x(ys;`BnkAXX!_r=G%*JB)D&zK(xX;Mr=F9uI|YF>^6{_ysyr}HnNQj(
zY)t$Q(>KNiaC7PFuYD$Cu?5P?yloNR5L~)g42t<})Y0GLp;X#31YB6@Vaxa%4rrLv
z3ZI1@i63d*zuV=Ais2w?E}-$|4kieBt15=KnQ=5v&%RSN^oJe%?fF4cJjJ)0Ki_6v
zxE6SG6Em!oxEJR&UNa5PBh`T{f_^KlP!y!$^rCEO%Na2Sa-=cTMVl@i%f6nbn?AxV
zK3}ngU5=^WglWaM@1ohN9b#zHEE(Ofub}9kt$ze0gz`Me->5_P!-ym<3o+9uH;8na
z%KxiwdJOOkX_$EeIoED}zjp-RPawZ`6&Rm81aodq=5Z^6$skZgl=_!j@om>bTc7Cu
zae&5pl-9*fdoC`bKq%60S=SLUL-C*$^6-izO^E)~P16%Jfk>@e(O$^~!Udz9+kxnX
z7Dx%WOHcMKgZNJbMyzi}iZXg!c_VTgp*1>TEJHz7QUSse;8mfOKu3-`?b(-oYgL!9X_s&1`dB6TeUo!1$
zpifW3J!>!+$JGX>BMnSub##3i821%p!W7UK6_USH6*oX`z!77~tmTV;d;p(|gS>LH
zf*@jbkHc^yo*?rR&;mUpTIw681kOUtD^~eIDD&52^r}7o(K|C+1_hQStXm{tQUoBtJM)>WtLRpkfTif{8?d>mxcfJ(6L_h#>mahW-FKx!e;b`#Rb=_-6aaUJ!>k;m
zwlG^)l4iV%@=u8`nOm0D*s4lz*0%bk*7U>e8Br@SyLj#b0$3rr9pCC*_-!
z>Jzk31td6_+-`y*wSQS0rT~8ozghR$!BsvjGbz<@Hm!GB&VV84&_csVe+^pz@@hvWI9F<-nM)pC0K^z7;Pbnoi-
zbnZH5)?W;nI_Q=_*R8Igky43FM=N7+J5*-i^(EWH{lt-iORN}j@qmJj#Uda1@n{~n
z^x0Bf#tM@r`n#2NM(w|$U)iD-V$~5TIHIG
z3WTa+tAhbu1V0#uwP~WyJ>R~WNle4uu^94KbLq?5I(-Hzt0Vi>
zS7#SFRi8{C#U5eps4*6(43otxd9lE^gCJ!&BQL=iR6xQ;iELc58uS+}i5+L0e`9IY
zuaaf@Q@5p8%4X>2PNa61TELom!nl*au%`
zl#W)>ksRY<_2pJ*`5ZwOJ1cu@<`OM!c-O|!rH=xNCHmx!dl;MAQde*B;(*djQ^Cs)
ztOd-43^L5JwsmagARV*cD96iIrD}KjX?Zzo<9^$zs~ZPEn5sQLhYsw#N@$t^K4i)z&!rGV$E
z;u-=8xDl+hH%pT&(u$$B@Tbhw+EtZ?*hHJtqeqHTk8J`5@q)2K=0!DSs~6C&>`T-E
znO1hq?MnfQ1*m`}t0Ifo;1P7JZ>e}kS;LjH@U2oEfIB;6P+?XiRC(eYN+-KFH^%so
zPhK}HXfKa01=e)C?t*t~
zyVl*SXYb-8^yv)q>MCnd4)YdT5iWD9^MOgxotYkwNqUkp&G98>tORDYPjwm!q^G|O
zm9M9-;r;47d}i9mLd67Ty6)iBAri6KkO(G2sXrO*lHtI^=*L(r_X3HYJDmzi9
zSsQuY1;JrmHCp~$FRjDGCiNbUquhwDHXn@?yU31(3ByD^ddP&8<{P)wa}Le(J{!0U
z#I#k0OS^MmQQYYzF->|?gI{duSPgzkVexWGOB_US>f$cEXIT16_(i}G3R5VtL?%=|
z&@X;ZZCOx2HkO8P8THDiz?Gw(J)@3pim*VZr>F1G-nrx5ZJ;}o?r3{A69^#U6GaHS
zwXpENawx&8)KS}a6)x~jAlm4_AOQlDu%F9H!Hnyw%vRsQI_D0A9*S;4wnirVl8XcD^
z#KNdcZ4m+|H#1(lQ>mOZ`|qM$Ju
zCyzgW$ezx(R->eh9!(1<{itZH&GWv>^Cv-7BSi3SYIyI~Qw1_&-Qr-AxO~SB;SsM+
zoDtA@LG~AZc>*gNaHjNL={Y%87Y&c0plc384p^vnQ-x5m>Eq!5aOjR|ssy_0TeL}g
zb%(BMrg>!AYnx@KL8i?|HUi7;%iMx8BXGXK3K{qwm1OFUWy)O2zVDr@@ox
zC49ng@lKG+kTf1QoQu!>bI@YoW;Tu)E$K@{Vv`2n!r{_&;RX${Zk`ys*A58UfRwYU
z6dVRM8L0BsLjV2*I-r-!r2-`@buu4x|0B8r8?II+Vtt^UThF@?qAq-Hj(rB^ite2!
zk!DE^N;*uTw(UXuZKWZAq6V8gK0{sX&YukKANP^^U-xtZAGc9p<=Ctln#(H2ED~;<
zwhh6*x#ole;s1mwlk;4P6f6(eZzd04p*7ZHXdSKyo&s0U{n0n8O(Sc*y}?|+cfJf7
zD~vPay11DH6@ZB-#2R-%AV!qG8BLfOU&ScVg^yeBSUS4WvIgbwQ*s!Gi4AL;$+sXC
zA2d_XMXS;W>>3OOEavmseJIX4fl6T_
zkD|g}5K2F=s&XzqmEN2lcY3CNNK^_MRrtCUVD-z(z#E-rN;XFzX2%#I@q_*lieB~u
zfuwkZCs0Qtdrmq60OgIP>r{-tskLsHlC^%#gPL}Vn`nCro9~_;LdV2^4I;g3OOKC_
zMpR?+qwDYnE5XEO3)JC|`U^Mp^oyHj>A~b)#7os1MGgI&Fy&k_C`4{vYctRBe%bsY
zlu<=dN~??*ZJQcRe;3zb+4m6*B4}fjd>5so3~`KBuk>i$o*iWpGbTlTWxsbgC#m3|
zh!vt89fd`4MDE8(A!bYG@@<_*OtIJgc-9aAq#
z7b0AGkeLwK*iYj1_`6+w0LuTt4iXegLMtO**mUK;h8#*6mJ}sw@W34J>C9IV)Noyc
zjT|ffB}&JHFo}WQem?YXYAc02a_RHPaL_v%mie)&6^#J0W_x+N2z`3O->SWG@4%yT
z>A>vNhg{a_yU@1pP?K400GXy8Rw--Yl93z*o9fj~q91xVL}23kj|=6(sBnb$Xr6
z^Qr-7X(9Q*WLFcn_W8QHMJOkc7999SY+)ks)Pto{nSw?x=$Cb508KSR#8bhQfgv8d
zF`|Ly5{tQ0U1KImCE_@@@ittG&4BBAooQuEQB3ASnIbNh3&Ck(Ea5LHOyU>xQ_=m^0ER~Pv_ncZu$HP0rmCefpJ!k
zhP6}ceC$|)qzisUR|^SDMo_jGw9a~c-_*X-iOQNV2t8uRoykCK2)hTG4dz8TE4L
zhcsCT_L=8%i#rr`0>$o*eZq?_>kjK2Kke{qI(J=-H%#1QW%_*4kGorG<0vP>9}G9%
zHfbuQG1RI*L~gayL8If;lSOntA(SgbcU1Yr?1UnTAlffCN{Snl=T7S=7L3Z*7dQ#y
zS-tp;WxjKDR1(3mEAyn!@ljV39F`DiuH+fZAxepdPU8Ox#ZpA9;uIP~p_N+0kC-F5
zGi1EIl0c{quz=aZKN~H$0JD0o`L5QYSkQV
z8;lk5kk4uoK4)Y1F8ve#gNHR_jv|O%saVNZZHysux`RBv0x3{z4ySJkMZ@bmwugEsd_L3~=cYv`4me;qi)D9E`Quu)dzgJ!MM+Zwmoec~K90jaL$DV|rW#u}
zYZDiDCdW?ANX}L}X$25j$2jrZ05@;CjBtpM!E@#*nxN%|vY>1|CWJBe(C87rq+3;hCSnMVXX)uWf%
z6>*l%S(f~Pd&allL?6^-mqV3(2s@F;JcT=-$vlH=X|!PrmAh4Q6_exXw|(>2vU~A+
z|H9{A4MVH9*{;6|^r4GMh;%(t`L{Wss+tIN(XD5IKm50rGt7ihfhx(cZ=Y=YY1Y|Y
z^)F`2qaU;s{j^)D#cw%7+JrArWAYeGb%aV_wwqwdXtHY*D!?!<-14yc97stxofF`~
zPpL*mZ?rNm|K0=wCk?O&Y%XfpscJ7-xLhviRZ-)&E7lviBNvj%1TIP%cmWpI2wTH(
zX{ZipjQ`28&Z6f%`x_Y2k8sy*MlL+>&JUL;qRpIUkZ-B(9qOkCIu~bW4};D(+(zUI
z?D-cK>#{;X*>Y>G3~fKB-F}+{#`ib%fXq7Yv;bK_@zD+s4
zN1FatG-6~sMH}_X{g(m@g&VMQKB;`9j6ZLGPm4lMyg)v$Dgl+By)*o(-1By@Bx+F9MK)uOmOCG~4Q(RN5haQjJ=KS-Or;hDLRVvN6J$iQT
zNBAr?qXzb*2GC+r4|(g5o^+Q#OSa0l8VV*6Di{hPz6)UrtrGlGl>5dEQQnm77x4>J
zpyzemu$v1Yvr&K++8`-f82(ZCza}%5dsKa;`SuCD`~ups=?Jdwf`)wcBb}*#LK8Kr
zAcvez%Y9U-2>J|4)h|DCG_->ATZT%bwP7UyK%gg{A{h^93-Bqn1>isFc7C5w>$Z7L
zHO+4OY1l%LZ1ioundGIls*TldiAW0UT!Bj<(&N5%Kq8h~LDO=xdobPKDNzQS<{uV2l2zBHrR%X>0wvBSkeNDWdvFoYhux{(#
zvvvL0y?eHfsK0&w;<=f0i8JVUo(1{aut)S#5Dk+EM5GKr-R)M0Yn0c>feqV^IL2(LG^
z#cWY$wVHUod!PSD`h2h)Ec!PJ(=)0jcxO6lDe7O%s)wMH03?h=g7i`(P}TPU!bYZ`
zRX`^Sr%JAGl3S(hS;~QO=^8QZ+VGpF$sxBM^61#KT}lJRjI~$H!R49c%?zm+D{#J^
z#>MJ!nwkfpX#eL>L4ku1&-C0$Q0VNU%bEE;pk`BBl)l#1S;OjEvU43iWw@*(=*Le>
z9zi!p^gs@d3Tzd1Np`pglR-17g*Z`Yl11dI+i+B=;-cWA0>AQJV;uNc{9B@m~~WkHdgZ!#m3m3
znSzoBy(Wn&TI=CL4>}GaVeKtQMASPGbT+KI4y$&1xhjcd^gc}N4Y5IV@Lv15SxOJ~
zPra8ndHqklG+rm^^5Mon1&N=@{1j63fipMtq{86Phr1qmj~?IN3(#9k)b+)iUXP5_
zpwI3lFqAxb;IBFPo3dY0b|BQDD;+qWc(CM+s8f@F+FYq~vbLnDwM7F1I996l*V-Ec
zH(_6^wYRmgdPyFoeRP+O4vhEE`XP8Ngt)-!_INaW@z|A=0nK2kujzp^<6j9#X6t;=)OZ-!lIs~tQL&~!jva)V&o}fo_DK9IqkSyH{7nBi7LfvZ=^&E
zc3U`e&RCZf>tezEvP@WJHc8*am}Uc1>nUShCoF5mG&9oVQWF?Mv+pKJOWA6DYQGXg
z=@j}N^vpwVKwHeesMbJKIbyY;4Gf1ufW6j2jTx?5+$iDJsgi)f-Hpb+LqAd@b+^DO
zqKJE3bbS}+WOZoEvS2-(a5|lFIvaqhua|j2S$d6B*$ZyB4PU>0#n&&N@%8JBX`XPq
zy`$`#)rsso%Dz=&+Jn?oA23&s3JCPWf1%jOP9IBJHCQjF%CNf4`LeEfdAWF_U3;8P
zE0$%!G|gtX!L35Pv3;fNDEmG*A=W^;X8XoWx_)PW8y#w&@E|-SI)WdvI<01btb_T+
zhj?x{y$||)eV9J{9big6X44E0bKI5Cl4O1%y|?IKIo)bwG&Z`X@l8U*&1T?Hn>*Lq
zCIA}&U_lPTC5095?rOH-MUB;|4uq|ZM77RkMnbmiHz&(MassD}XbXBSI4;G+n&qExOUzSiVENShdwwo
zk}Tb8!1pIT*kCDwgi}ALKl+{|s0PkgTw3S$-rDOHTQIK&6MGNBC?--VB*kj2iXSuk
z&sV7*QJ3zTDx-p?(bBl}bkZR0yIVi0fuI^(4Q3J_rL9!Q`bmm)rP}MKq7^zUTi-3f
zV~?7+5c}~{MW=^Mdj*mVQ0i{3#Bx^_(+cMUz^M|%psX8uHKIDtB{JtZre^#B);`}t
z?sSyG=g+STZj^u_L%qb|h<%f3iI{-CE?Z&I#EZm#?OrT*;vPW-l
ze{{n2g;sGk$ae5zSXHx16)D8Cze$58KWa44vw!68W>TNWP|y9b51$6qHMd-Dv1!Wgq_ZNcPFL*Npx
zIZs$->z$LmmgjaWD7Ota7Hr#sZ7;zxSCKM-Gg)t*Gyx3k`;PN%$L3CW+ve_XyXkYw
zCe2+&wKir{3x4Mr^PDj+R-Z6ABb8^y^DeO;#&(zFcs*ycnQ58zZ%zPgs{V=+7<4+qMm#KYzyOe_gSD{)~B^
z@cuU8dVR<3dd2k}P}!j+aYZm&wd2|uNj@@Nga{`%x=`fy+>He
z$n%V8DLC!s==Zk1UO>1?du!9zQ5zE5~zq<
zzSnHGn=Pc|K+OujyFI!yiPWvHp&xjhaIDSX@Fe%1A|E6lz6(f^^lPSKPF}vJwal^}
zg8${Au%^5M4}@h
z9X$D~0Y(4L;!K=0R;nH!1pwC)%g3h4$JQQ^y>tt9u9oGFzbVx3Hv4RE$AQ`q+k0O1
zenj0#6`&5F5-f>OO?x|O*PB99?~51{-$%QXsOr!f%tKJXNMbTd0V8?4Uv4($-h*tGprF@|VO3FQ;8uuuS%p7$PWf)2t*dP1KMLlHNY%~iko9egWTWl5EpAG&H&Mtu(+6!}4w+#(GEonHt{K|I
z^nCc8xNz1W&;nC#S;|_7pOd+_&C`r|UaZl2xmiZCY(UxEBtm;JH#6pw0jQiZ<~3uQ
zQ)6>P%+0D4lZZ$KQH^@`B4+PZ#7vl{j5H@qQ?j7H0Vp;wH4*Y`(mD$WJUf`2DUcnz
z1F30V{Btb}SnVP2pWTduWs2
z%!wPA`Vna#-PINyGmZ$_d~a}^67oD_Syr4*CtNNUoX#f)Q|HFskT<=n8sD<-Sl0#f
zJYkv>I03mzkQ`WS8wLCSkumx+1!tuNMPR=L^m!bLgAq1)L@%2U1;fC#`wszV9RZR~wnO!rpPSiC~KEYACx4QYY6!
zrTbkI???4#-3qvwLO!w{rbEzqXy%B>aECzaj{sB}vxYIz;~PWG9%^qv70{lILKT
z(Hdz|D^9P{i_S3PDCBRwku3PJAc53sfD(5%C1nB&2d6t@r3O3TA)t2|o~OK%&w<5j
z6-p^smc^1&^E*~Iv9RwhDBFg25XiI1Wq0SlvV)jVX6t39MWA!G>~2mTr#BZvbB9OhtAntI@fu|Ixm=~1?w^+dE0}?O1;0o<6p_L?YiwK*UkFRh}D)Bw^}Dv
zBxF#kJ%tE|&cpUBy>*MlL*xN_$d-!qIuv)KsuU%W-ibVa2uN|y{8Q}xq~1iALy!#}
zoO!!jFNXkbG*qdE1*#TYiFbxl^m^dhjLC?b-c%EnfvSHFCL&$E(V1C|yb$RSjP>PU
zRJgRnLbYwCLtB)V{dd4q!IU#&M4{H!Wrh@8YZ{QOa^;ZFt?$jDkt$oe3s>)o2?%L`?fA9Ex(!La?CpxS8Gf<*~U
zvOWlWj~U%ZEYx?@G7O7AT*NSX_A!NKl_@pL@p+!H%nL}(we5B**xuia15Rb&zLMB*1TxwV<0**
zGgT5qi9soWb24|OFrWVL5eo5B%3dS3;(WGdw5QV73;cM4OO06)~1w#WnC~$
z8Pk+P(_~x3x7!t0w!UF^VCvfyP#*?F)a$6-&bkNJ8rckKJ^PPtw)f2H*PJwFCGf{l5LYQAwUC}#Qs
z6LIGR_&@>l2oyam>cw|BhF
zE8dqkye*5l00P)=8*cAcYv)-Ma$?MuN#fl|2sG
zd{-P;{j0fhu8TJaqo^6Nh|7Q3
zqpT#|z@oZ3v}V37=n>p?DUNa=?lmX1SW2T=!#7P*erqlCT8OITAgfg)l9uJQ8YZ-Q
z&|r2w0@q6GRLx9KH%Kk{As2VI#;VnYu1m11#0*-Xx;vN}<-(7peFz9ayDX0mu7aQg
zGu0cPc9t&4`bgTdVzgn`)*|&zYY-7Y2P_9>s(luB;AYIOI56@SA-?7(N+6n`#1PsI
zfRYZicu^&c=9EpKOr4F2$$E_dsj`4xNiN-Xqq-fppL2&@4(1XG450*}?8z7nvNA#X
zoorNhxx}5sw6y)8y7U!~tBNN912grYUt6tlZncME_o|TAvd4ju>NU{^>vA2{PD8Il
z?fuxK*m@zif^sufinAKt1-a?ZXiCyu9)speJE&^4{vb$`advYTV~*I`T2R776DFE4
zrx_DZnDdOpj7@J~B_w5VVHDb14@ooPNul1>lw1`^vDYyUvOJ}pw5)7Yb*lFL`?sWh
z9L^&Jm6O#OcSfq4wp3#pZkp+;DAFC!qFqM)w*XSdX-Yo$y|f*rWQCjafWoszusGonovsuLCK61{5iKJz=(*!aOD9nUR-dOw^pZm8;hjl~TZ|
z_Q+85?(E&@p=Y;353jiKsUM3JfNC$6&Pvrz+D1LwO(~)%qFtgF^yz~O)mT30fm{{X
zT3tnx_x*@v>UaKVqqyCvuhvLebJ`2m(+P_MsAX9(ud6eYRtq&xGeo2pLRwc!uiF)<
z6ztm#x9hv#+@-Ox_892hZ$|sw!Hv9GwCm3kH=_VVt+(_BrS-E^)!8VwJHMQpr0;UE
z+NgD1F)s^NokDBvXCnY6PKMhDpllIj>y-O&*cTJ(#v>D%jJ-_)G-{p5PHU<8H
zRMPi-G11ZTI(%wvU}!y(9CR_8q_0UhLOlY7785e4QG9cum}Cybyf*Uy6@_qV+4L$A
zYPPU!n@{4(NbT+Z2X)KX%GAP}S#z>S;?#8awQm^#R)>_#@MNAcGBczU+^#q5w;SHS
zE-2gPh2kbPY%<|epoM_Kkllb$Cuz*w0N~!Yh&K%S-
zeD%VS8sH{payiJht)9i@RG7e+!J4hE#DPt5wmVAuDDJ6b+P5*+686ENV^C
z*Bhw|Gn8!}S^vNZ6Hl0U#>^9DUN9#EREZOkSbtAeOLB|sZf3!)F0wIF+`$w`iHaPF*AOJ~3K~%OxP#RlDNGA%k
zP;997Hvm$xlW~|hTEsRqQrl>?+E>#ztX#i8_IaG6{>Wxp9~O3xz-uhP?jWM~36YR}
zs5aRafP3RbNMn7j*DJ2qg!!6qeP_%IW17s%zk(=c@ocIuVB!)Av8Jvh0@4Du*3yQh
zgAqo~#!_W>?aNC~F-hw{LEy>2*0g5K=LuymSWXk>b;7hvnC1-f8iQKESK7i;>s+K|
z2dic1F4km|l3QImBLxO>@g~EGkvR6Sh>N1OVs`@_3>0hNNYyeO3|0K$g9b}zm+EnE
z!aIBasHxQmrPf!W7Wy(LaL!-=%X-53d}~y)F;X?jEP(8IzrK6(-hxuLwy?JzBe?fR
zQH+ja%vCySE{jvn)+9AqfT}gaR=c=Y&=We*pVtMK^NN@A2``s3USADF8GxGI7C02a
z`E_2TJ|qUDSqUB|LcrHfw>-&Dvwwv{bN4&;4=D%7^4P+yCgzAyDzFk
zzdfaHntg@L(^Y*R-g4j*%O^RNx
zYM`ltDCX?X+!{5rJGSL1VV1eiYbH0cu;dSC!s)!W+NrmpUGAryI|j)jlJ#0
zgn`Y?a9K?%4T7l4`;vp%q9`|$a?`fEd^vfth=UY9)cp;;$vT
zWORRPn44i~Ol4`n&iRGfnA8m<5*xQrc?b$_V#Fp#CwH=|0eO{<+HrDk%2^%7T4V3N
z7ui0Md`}K03^*pD^kCp1=(gyBd*hx8o+20gC;0e
zCnLpd{<#OIEjh#DPJkK>(L$%{dI~Fu+Fg3tec8m%J6FFpa(0sN18SlXO!be}FyXbk
zV1~_86-^amvC45TLQq1*8ov>g>2JiP^bwJOuGSn9XmZ$7sLQME5&Cn9$I)IlE6!q(
zn&jD93owubn!LZAs{LJb_H#KuIMpE1l1MybrWsROG4qU>7po0TGp00Q=D8(mNvQSL
z6x!{|HjD9|pkjhEQtYz6-7Lq|8PnR}xF*o61E4!ks>(pDj2w$MJ4at1A*J_wtNWy?
zyYLk^m0)W&tPGN5fJ#i`sI|$Kd{l1nSCsZ?Fn_2;cxLy}n|cvfKN-67$A5@_?zKAy
zCSLEFY7TW1hZ?9mn_wR_rZT2w0z5-o*&h)reyq$w-_T$Tm54J^yJeZ`4eEMX5k)S7lI+1;Mzv9JDYHhN
z)QXO6k?`icqE;Xa{1BkJQ-lr)K^-^F)||FB;hiuq3+8pjG|$M>1Ww6y?Nu&hflx{r
z#DwLvLc{{VY9rsAG*tj8nBFpOx2wxS2^-zqGWxz2S%*!bOEpuX*8gd4wH+1^CeP@%
zj_sv%U=_gB>9pc}I^l9LyYtuA7rb1~IGs+I=NUPtUdIs|3m!n!Adr==3AcV)rPWH^
zJNU))#Yb85gMDYC-RZzgK4P!+upa0`GIKb`gRD;;Bwjw+T7M#gB?@}LfYDujXHUMH
zY?v*n!re-nf>eWi-6aXmaH^)UM>h+Q2Bf%&l8C7F@=*2`qYk2k6(Zbfe)eK!IXP{;
z#=_T*#zLjsdi7Km%4A9Yb0FktG9Z==ROKmKK$+l%IT@GB8J|CY!Iv*z@cFZ~oY?m*
zDhfTxWYRtZQ4Dsl!D{OXCvQ4Vj5+1j*nC|V2T=Z4jfF~7Szfkmr+6L1W&mp6Hk8{2
z6a^Dw&J)hdYRUrZF^p{&>|_$iQWT}^wrAIUBWIH^Co|5d>kw6UwQ7yth)46aVkWOO
zDQ)kfz*I+*1pE+?6702*esI|2CW&jwCzl$bvNi#b7>UhhnvlHTR;eZ~#?sq+MX60s
z-XNTZn7~xEpEYSnz|#+_^Ex%I#|t9KQX9xQrObJpw?R+Fy<4pZH>8Jg`O8KfD|1QoU|MhjF5r1iRT
z5XG_0i?LQD#qWNyuZ7;|y2`^^%~su$s^Lj&H0zD33{1HVw-SKV$|bZWXov@^|L+Z2
zdutS5J~@~`1owIkAgbvJ&NGy=+|qE
zl&g~b=TKEjr0(KeJdscxehLF++0(iy|eDPhAlJt`3^6R^cAA3X(BTykJf%W?rzQ
z)m=(wdrlJ))ph}dqzOqf3hjf?PT|RYVkAb3QEeAVAT0o85Q^21I@A_bx8rY`cLy{y
zN^od|sBO=yQfKM%VIi@5(jA&AP}2&cgaeRr0^1bKO=|tO15&YwzXnRwQyD^|ZG;Bq+Nr*F2Y-tIQ!X|aqpf;;Artldaft7Xhw@vhr3;QVPNvcvYB5F(w&|ga}Lm{v7(C(1QuuVs6eSz0L7$e2l%Wm#X
zdBav1RIJ~UdJ@E#qTJTSR+P+c9z)fHg^(zN7{(Au!89cc!V|SQeLkIhj%Q3cAv18j
z-P{$jH0d#FNxK?5Wk8C+Wa|g1g?5=4OUgJ+vrBl5iMpK6IIkz1PiHLi94v~xAi7i_
zRjgj?x?$gTs3;N$Q<`y_orNljZCA``$HqHKvCQf2EdH(vo)U5v@RX4!n`4~VuTc%=
zorS{KFRYeqOUR&ii|?Ng1a>LdQMyF7?niK=3r`HiM!j^j)hFsLZ9EAMHBioMcF#oK
z$hJ1eM6$o5ba1EZeXA_tWHQ(0w%xj~QoARrK}~B;TxE*t^0`{;RgGAPyhUk=Z#P
zffNy;ICvr@W22PX+N>TF-u0|ICh)50rPZ*=kPxX7DIj;_D8T%$xZKvF6l-=lJx6gHU65uAw?wzH`
zj^`&BsoHO-iS}9@SeK3}Xh=y(ONQHb-a{NCI<`Qm<_9#yj|a9%4=T=Dk`fdJAIV9yKP48;BxJ#^Ba-EFIv=@5fQUPuh3v-kU(7HpqjF2l75{u0-rEb(Vco2=)SCQ-BNRE93o*vCQOs
zo`6JxluXu_S&)3ON!6sHx|&D<8jzAfJDog3kaI$IAe9%wv{;U*)a+leAha@4^C@FK
z&me^Xsme&r6NJpz0e6g4koE-zse&kWo6eM6?v@yw8M$PXRIC<(ckEHt7dMlpIaZ+f
z=AB&D2)pTrIrfLj`)HK*rHoYjg)LTa&MCJ|wHEb*YVDRu`|>oIHFQo`mlMK(d&`C#?G#{$@u0uTOG$~vF51jikuU+ZF3_Lv%0rn
zQ@bs{2+k*$_Pw0(`f|qW>nkoV7pwnTOv*?UYItf5__5=<9<1zJQ^P9lD=r7q#z=jC
zFR8ztv~Mg7{uXG`KjveHM$t6bS0A&rEkET!gLtnjZJprJeMhP}Kyg-TD?9SOIU8?s*z62_%9dO(AmywSCkIi~CDB!`q^VU&
zN}0X3bi(<3!sTVf%a=1wrwQwlF=s|*#oO06Y`d9HY}vRk^7C9{Y6>VU_a
zr?2yj^Sa`4GA8zNz8H`?pRMk3FJ?#V^$PGlFXgsjzi!yx&0q*cnzS!@EnC?Yw_R{c
z8?J1}r@Nx;8?M`qTiI=$nFYB>ll0`Az>{CUgeH#C(BDa;E8}2u->s0T)d&699xJM?
z#sj-|DTlYU|?t@hu(Rhjn1KU
zUlI)scbk)JXrkLXTgUdX+E`xwj+pUSW8#TiqwRuB{0I-WFxQ^t-R
zn@%XS_d46=Sg39`5u3eBV=T$%o@fE-j-q?Jo`TwhTe?{W_es}#?Nv33Cz=3Hz#t?H
zH5jHjJ1m|YY;06+4Ga&UKyP&x5>m(xXAVZJ&Bn`HQ4+_41vxxlBauvM&
z$Cig<51&U}wwo=mYG+^E_B*@Qvjz^(Y?8g&-$kpH<>s|88D+m=-`I7>0#rtdtA#wd
zg{f^x>OiVKsAhA$p1{DAB4sn7ND1IvHS;EkQ_o_hiIJC#`7~j@neKHtEtt;}rX?dS
zwop@|-Wfee!KmGg+s$y2xWJ(v6ncvG~5pDSh$#XrBF0ah-mv
zR*Hg4V>WqTHdd~qh59Ia|I^xk$7FtH@th|s
z{;w9K5+trw7SVL(!}bySbx3uZEJO3sk;(e&4tnHOa`2bAlZZ5F+&v4XA#>TX^*;!v
z26r{;W$y8QPzr&VAQ%&pG?M|8w`Q>#^iEw%`It
zMx?ZR-v2SDPq$rCMv%by0e}QkD!Wu&p6+S;3#Y2QFH&X%$%Sx)RTOG|xtBwu#EAVcNYGA@n=l4p+`u3|rn)>=Kr?tnIPbK3
zFIOa6lEkx5U!9@fN%3a6dkBNO9gt_oSXpW$$Q<1WajIeX86&k-WxPnf*S_fhR?>c?
z%bFpY?Hgam^7Omodive5erEk?17t}Um}@2?ffakWg-TMUB)3*DQ4#@GUSJDWmqFY8
zqg2+v7h~8MY_aCtmS=Ci{p8ZV1FalrY**Zeb%vVo+%Mu`vU8wr%~b@Q{{2Fj`LHVx
zXD0q&Z@qHQldjO5NG??>FJ>HPH>tQ+DzA+zwu}PxyP}751t*a>xwE>(xAv9)Rv`7y
zP7M=zi!3sg_8T_nQ2laK{~18H_cr_eTlDVd^@+m^d5NuTw^a*~
zK=MEFEH|{x-wp%aZhqesm{Ph-ss~=iSZ&x8|u=f$K$|nzdi7H9C&{_(R<%}E`}^d)a7^pP;2!PW=l9bj%NOj(?HPU
z@jxkr=dcFpG56HMc=qM!kI4eHA
z?E3Sw;`yx06Oyhd+O5@?=Gq$G9|wLtPP^V7cz?5mzVkTHs`aNBBVmWXDA&F38=ucl
z{Px>#_}hQ}4S)aJ-*MmF%E;ZjAgU(c8&;Bg9tXHq`yL{|C1h~q`pWKL>>iM#Lc3ig
z1KK^{hLO6b{43$wPJqu%9Li!UhwJ+WJEl3eh=^K3o9_64T{WdOwCXj^weqsD8Oaf8zTc@PN&1AAX
z_ruo)j^T%<_fvGr1dR<%gA;%MM`#68X)-HK{}WdNYe7LpqXUgj9Q;5lhmGgORdxc9
zkcCa~VfP!#VsbR>N*+=oioJt7QW=+&HnT#@^Thi;o1~*r{LSkVm_@x`8p&xD%?I7B
zR5p+Eik11Ag)n99{`0YrE6e;iTGru``F|jgT75RVP9^u;(2G-nGF$W-u}btJ#rZwo
zjBefr+8&~P?(nK9pOL{~r4;j;-2KLNU3fk(P%9?UV?r$zh1f+heIBrw)r;v|C#oF~
z(`!(+Sw-aNH0ryILvrb0M8<}J*O;&prL``l$c6W!b`X+EC}+Vj4wLyYVLSw_zM;hd
zx_JUv!JKqxDoy2JYFfh(xKyY3Lh@BdoD-aKAD8-=Y`&)U@T{Fy51$b6gev?QB=vG{
zk!~fm+pkJKsW8aTj;LT*{4gb|en4v|^~PQA`7;u;d#4z!9XQ|K@ZK+o+#ra@jyDG|
z4Ihsa{k|bl%shsLMdZ>z*GEU7#JbEw@;I$L^gPd$nB!!E42(f9#lsGy9%i&Z&jZKN
zEMTJ0><$}!V5p+I=c9*2Vswj-P3eUpKM?E;l79Ezur-O-B#Sg102e#;1&2)_^|RI2ZpGaFlO}kCRDQYcdrFhw*|YOH
zf!F=s@KYl^TW;I+piTx?tP`*Hds^{(5&P-vD=$MnzYY08|&ap3A1LJ(iRq#d^?)
z?SQ#~Cs|(I{Y{hbWy|bJ+>chNisrHP-JgOB2q)gt*TbZ!93L=i8HVEL^`4oauLaE=j1){43XND)UJ`?8jy2>@W6YS?yrQw1A(lzK
zv&uX(k+e6bals2mAl8Si3WXo(Ime;XE5ktG=W~*ob`(!7SZl
z$r^R~o!C2VOs73W(7UXk!QfvG2RzA%vPI0+-YZGWVE4o+kqs7yC26Vn(Mn?|z{@N@6J@=;
z6ZZZP1WXp#vpS=6UPOyQicUfcKgldq0dcTY0i?bbW(A0nED3}$Sdu338<2{FDRwXw
z0UL#yc~ZEd@`1)D3Y}>DK&1m+EEP}8cxTc3m*BpD=jXuZv*Y=^fO3P10V%4sk8=t4
zngeY1i05QkU=rIO$oy<7G`%~T=wf~GpUnio8DeDNQxQ0@z?4-s%t-xhBbFb5r#wWb
ztWwk^jg#TZ5q}-w+0~UXh$MLa)-mhB8^0Tfw0T?C8lK8#=AnNX!Z8`C;c@1aR0iZ~
zt?6rH^7!t|1icneM~DhTt3`Q}uqErEqLTGOooxQ6)%7wMNHOArhzJSMtc&9*Ga$EtsqMr3$DWMuihJ|Ik)?fC
zzj56cp3f(0u`)?2=39B3)fp6p^lmJ`MB0{C0_i4Lmgq?0H7yG7!4RMi50}XD4L8O4
z7ffJOltMr&kjDWX0HJ|!0B8Ue&=^zNr<@q6P1<*u2f#a(
zXzK*S7nX-o0o?EsPJu$k(g<6>NJh|#KYC#v<
z_`4`Xt!#O8!951fJqp#x5VI)e<@_MV{9IXbA4RUbshxj%o)A@x4RDz(!RR
zc>07(Grz21peaqujM0yF>tu8%D#kd@=flA*TZyM7dlH~e3zcz+ud1ZSFLgAzTnXXR
zbD6h-ml0g-UL=q6whl1sLgVJIo|5MQ{ZLnZSg3+{*P0{=DNjJfGa)bmssL5ISgF~J
zn^MA*R-g+<%=ZLQG<}b=fYd!13fkp;{wJ6p$`8Dyp|Z)aDm_s7fkr2Y1cHvP9T)`Z
z9sOopm*Vp|@bPiu<9UJ04Xzzj2Y4uo0HuZzF{bdr-q0Lw4Pv|3gr*&P5~x8?*sC|FY8R0fwk{~jQQE||eQmV%i+T2A-8_p1HR^hnJ(=NFVYk-WNG
zf!5dfvcvvic1+V_OV~f~uT;hQcZ$@5H36SJpk2)Oaa|WaTg7oOq?x>?RNJfpe~`=4
zlpYb;Tl|n7-%~sg!vWH8aM2aogGrOkM90G0MyJ~iBuf&POn4LF7t^DIEiPS!0BwLa
zi&I4bSMP!;E0Hc=g~Kx>u05x%&t}OYUpvBRdZlDalys-+{lKa(o~18Ru76Qb!K)U2
zrY8C3x(uOXGA(0t^8)pb-fy%T6B17=gIZXHT3<(H%xKVB8LbkoqvGCx)Ma0fF*|z#
zNVV2*oQr)rjsv9{6V^N+Lq+VLYr$QM#YvML^lNnclrnu)=g}<0q7>I1(~L-$K_=51
zCMOwV;C6jCdr$yz@nJyO!Z;>~ngEKnV2@WuXA|Di?|Il(a$-a=&vY_c9);Ium;(r<
z)W`^j$iNCO5&Iln7_>UPCXk{XNS){Tqx-FZl5sbuqfwgG>PtwLj~ZxIIIJIEXUI(W
zFV}(?gBxAV@v**WIvD6WaM3JIWn;iWUUciR;;_^&(>`}-_v`MEdPCS+?zI%C*EoQ8
zc~dZJulK;j0KC?3TVpNMhfqOiP;DS;01hw>lnT_Fa92XDgu55i+R%>E3U&|cg?oED
z@pv;PHr8j3^_j`bP**5H$YcXhs~ZSpt_9-KB9rUYtIc7E`Nb543C)uN<;@sfcECP{
zcT#naX&=r$+|lcm!qZ%al1W6OxN
z?Y!_@At0^>gM=|qAj3SmItUWwn9iQmLZHUpNMjPm2u$I`J{gEIiJ>9@RfsYrCIT_+Mpr&l43}pU+vhE5+0%eR9Bi|d}x&mDj%q9tU;p#l@1g%kXDd*
z+Fd7%TXA0l_jBO+d2iR})2s$u2a1Y^KDjrHy*5OBAVLu@k?UV%mXbztB#4|{rIc{D
z4yrUkkgVyyzRaPvv}Brc@zLwU%P7O8vz5q0ANzUX9r-w^4YB>DPy~0*dQdmwvzywr&3i(1SFXBwbvzQ-XPFFv#z+j{ZZQ!aN*SIhK
zyxGx4nHj`nt&JjywL^R_j6!hv9}1-Ygi>(B2tP5)ejx*w(G_1zg5O!y=_>*4#`}==
zwd;xVDtJ5X1LI=-O&q_~8W%xmcsvT;9+nby)QZa9v!G^~8Xo;XdJh*bFwe80%SW|H3vs8cNVP{yzc
zVZSkYo0Bi1V1Nja8!8RZV|q)76qN}_VcbW0$#{|NOQrItmy$l61%p$3(+IeAtpo
z>p^J35HVoI6=5u<5-H2UCr_*2pH&Evu-2l_a
zh0otU@!Q7}A0HP!KA!k|UikdH@cg{+ye`RR1|EyVOh9@XD;avLANm07_;=w2U>K
z7p2^7o*gmC52MqnfatoKpw-n8B#VNlEa{V<-Xhc%g}S2QUbd=G%^eoReK
z>Dw7VcTd!7h{kK+AZ3>n?b60|K3eRyx=cDGp`4Ti2$|nYcj1z8mpN!Bg@f{n#s-|u
za^m!R$^s`0q;uX!wv^00{V-HQL^KEA5ri=%`l$UclH3bG^ZOcp2NEImM@BbmXxosKxZf
zyhH;u7&Outl6XZz61iYHCT(nGeh#Lf)qqqnAQd2rxq-NY%UI=_?$1!xEgCiMbpsT%
zEeuoROevi?4299h6^80Kv_>m;`s9CnB*?nzO=JB^@25#m>XjnqovKe=Vv*!GoRzXf
zKR*<^reCF@iplIqty8KnFbdpYYjftRS;6p{#@7kNB!wBMr|iYVeX(#Z-%vXogg1zb2ak9e?SAetn|lK@EVd6>pmOt%WP%Pl7Q2Eo^zwP!YAA&8DdnEmNQLp9TQ?&vQ8<7f#h))K1&dC}eV@@1Go#LR)`sN4sPaaVy@X_?
zfU^88GWI0C6sYx@2)~7~o_cvnO1L0#8
zkcae^Ng%|jcQy@3ofGEyS7w;&BSb2Ge>1@1;Y{$g!Arpd;3Zc3uNv0>>9C27j--EH
zoV(yt7|>`->-ZW-PXAPnus(b@jU{6^Z=n)2J6Gbyl}L^ICV|c(ZHpsv&
zH+exJ@Apw#*ile;ppf9GWLG1!N~lbW!cE3c_~sGo{r+D1^>Tg-2OaoCmDpVU(8mKa;R*&_r#!8h2E*;F>h2Emg-vEICzq=z|;ucxdW-;MUQjZOtg7s;=Y#zqO8EzOy4n0KhDDr
za_4E2^;(UQa%zYDP^eRH9;Kum{T?3O%0M+AOD)ag_)X#`s`J-}v@aG)o^reZ53lX<
z`XY%AFdYFP?rCrj+v+Ey`MN-b%VE>KKY>&j%7baj)PDw29)n-KbhgGMoN3Db;_v(^
zBUQP)tan-5wgyDx9t@S
z#l1T`578#`#eG8qKoMgU`^>agXssr3i_aa)ccaYyeN*c3lF3Hdf;QY6YSTq!MJZN7
zZvK~wQWb5Gv!)f@(;{nasK<%(d~+amqSnK%R#7UWaCpvS72Jxwhb8q{9Y@9K$(g4!
zQulq&;Q?-V=aeVB-WNVUKdspEJbjNaAZ0@V03zNmcwyX6A7*&^)QGIJb(;(YIDEFn
zQ~W}{me$+6M*L{WOwzr1kfu~e!AahenBR%==5UwXe^Eiy{IF+W!un~dB`Z?Lt=*-J
z9pe&=0gPZ3hMmKifZ;%js`;8K3K}<0S7k_q8aR*=9T<9p#LGYBb_a!fLA;Mg#41bF
zJs1%`4T5zmI#VCf?kbU3*H=HDqz*O*Q5uYtRxp{Ne;#^Eb0(@y<^Ujc88ex0+yl?g
z%K+5J2R?rLz~BG=X>8Qh@%i*}=i7R$h%Eq$E2D-t%g;K6;q@a6L|OZR?DE911~eaPI8-Sk-H0zQmvMkkKmPhk{H<0?PWqn@>s~<6AInsQ}aB`tr
z0#fl|P%eVkT_6~VG<^`Ro4~BN?h`}C#z?ipBs+}+wFoL($UtL7D=?V|t#Y_vpsciX
zX_R$DgL95vR&Z=CvPWrhgfh6i&(f%?dkj`-oIB@1TmV#vG=QtOJXhbCze@Y|t9XI1
z^;PK9awH9~>et9?Q1Zs41&T1eFC2$uFRP6eYVX&ff6aaOCH8bh-6DaxIXRj#0tZvW
zP4-c)%GNW5TIUHvRDeR>QeQ!GfV4v$K}?P8?HT%eb)=`~6=
zZOZyh0LT_gzxR6?Q)E|sy_fO@G0PjnxVp_&WIg3{F2CO>G9OJhrBP!PrRp(Sin_n#u{}6
zLoo)=GU(y!YHbHvGf7`CQgbaUbpe{VZ**yYvb0ZX!BLBcyV$d2KE>+86laB0A!2Jn
zFjA#fa4{e4aW=fYz2P_ys0#Xh;raZ)^Ya7OXSMI&8G|ty-IX9~&qa69|
z^QEwe+`Fy`NNuJD+PObl!^3Y^?Aa|3m`_WOx9VGD8I{UKG>FX)OveJR*9im&PR{Ul=}NF|W+J{Bi^69q3
zY={TCExAH4FaVUHpr+trT6!*Oaj1S@QbdwOFNe>kK^0udlj<63JzNHMpw!y{&dNpq
zQZD|xA`x1CKW4T~CrScIUrC6UAu7F;sQxAL)o(!5GImp@x_DhdG4{SSJ~Zo0Th1X`e{^{Q<1JCufN7Pd??DF-u}BLLszTfTMXB
zLRN1cmm(OlC3QAi!WnAz4zkFB))~&h%BH9KNoQ|6sC5NI>WNVnocu#9ErA
zsg>qwYIa|m-|H}-Qx}jb{^zevUbhJ$Xa!Yp;T6r#D_M+|4#WXeK}$f&fY*}B1duw~
z@iS>(>YZ8}&WofkDJ_(~(d1`a7`@hIEg7vVW}Qi!`7m3#u0_sk(vEOS8gQdMhwO)@
zI!33Ic~2d*K|;Z3&6wID9w5Q1gC(TUgsoxcCcliv=Ta
zF+kQT<2VY=qvCN|g6I3&f$Q_cxF6<490PaAdR1pQ0xLjXRzzae8kWB5nxEpQBhi&g
zr9boZo4lsXubsbHp(wepY$|pl$F6{>q7u2J$PMpgp$Y~Z+Ps(rPlBy44xDWb8IZ1C
zqae76QOiKT26`#z)w;xl6cqxGur-8n~VoCh_t4
z#7D^bK0feyb=*U6-_ex$d13wFUK6(tAzGey_d
z7fn`$={cKwioPguUIr)n0-*2$#5DQPl7tbF(CqgrDK^k!UMbnZx?~BSvCR|th$axM
zn~!dOk3O6-f4?VvQj+KM>IPHD|0W=%UnT*$c+eYcwunkin()Ge#_SQ{EmskaKPf7KKXz{n|zu!DK)*OxjBd}0sd$y$VRXQy3DSG9zG5jT++eoB`xEn3viXJdbohvmaEFp
z$_^j^77OJOw)bupmWRoGf~B!gpdTQhJd?^quu7$Shx7rxO(w>0FjX0kRsoU8_8GXZ
zCjEO@MElV!p|e#&t(Fy03(y+jI0_zT!~6RKqu;oD$9;E{(b4aYAv!~NR+5|prjSuk
z4ji)Rbc*}thyP(%`qvs{t?sCm?zrTS_Lwa%R3(9MnMq-0fVYWwqn3cw0nk8WL8Sqv
z?#Z7~@@i{p9|NTm`pvj2qZS{Iy4<&yj>-%vY$dclFtMx!@5NzH1yf_Kz9gh+dcLOg
zc=h?LV%6l4_j3bNfpUeqq)^xvhlgsu#Z+v`S
z_;_CCi+Vos@%hBv2fAPeDr|D#J@?}!YGZ*Y3*+e1u0QuIn8c9?HVuA>=~{y+XS9nY
z0JAcxnmmd&&tZ~&y~@_0KUy<6MZNYEsq*E|nm{W#Zs~>L3H6UGYsCigbZ6x`4|`t%
zCXfIFy(n-vGy9p1B7AvoH@}f!_Nz`1=RqaLJQ)jhaKq6KJkEEV#~Y6JKr1H-HxP>T
zrgs79?H!)pvE@}en&o&|s<%nBmZc$NefkAR$=1U)hpXSDCM=KJ0#oKgjDd`I&9S7w
z`WU_rS1=HO?weEUZVL~*CoIQ+3?5o!F7kK=;+`t={r$dS?7tC6=|A~r`sVff$nqr9
zA(|f?Q?`@ulob(7H(ksmTnk&0jUQ-@nzuK}9i8TR20!F>GIjL67{HaLFIxs0HJwK5R39bTsqvwc|iL
zt1(jt<7gJiUL8c00<_9_92M8wfpMRZapSr=u5zPv#uMy5Tr5vn0+;r=xtC{%8K{ab
zAoVBuP%zJcijBs!Aqh0}nJB4k{{WMwi>?
zU_-4NofP+N?_Y5_Zed1e#wZ1&3}`h6AQxUePUI$bHXj{JTkDgj?pa!53SD48M;PLh
z*Nk}2LX4Z0%kqDOG?&@@QGPKzE!A1}=KEdSe7z8o9KnRiVpp*U+J*N>&oIS=r^oLhCzK?o@}C3`>&
z=;A{OrQ>AcbGM;`D2erzo$+O3qAC~ET5;3^M{PLkfwP(i^k@$QQuRck1f+C82Bq_g
zq~Mg;7hmEEUoDmhELX<>03ZNKL_t)V&w29lz$V6>?(hgraoTrAX3A@;<1oK_-;&
z^A5+LolKsmJy`;1FTqatw0XNM>4Ip$vDHjx)ko>|0(h3x_g$s^)UT`-%sCN
zMALvf1wwHsV$6rWR+Hq|Wiid!h&UhUTxjD@3cmy*SO*Bt^rjMrt5(u$u^8QuTzTQ0
zj<|*$pu6C@2A-FJspsWDYHmoj5S9Y8W>*9^6bDhkMvd-2KN~Y;hX|nFjCzPqf(L+V
zS^HGLJj@rRx=Kai8%@EKQPpXs
z;gQ-W$=*^QU|nc>6Q*>J_Ggahg`$e*g^>~uzk;bbxYDH0+uk+zXt6)*tWdTtADiIH
zN_QNx4kl{Pcx?ux`rW-xDLp%rp)BjOqE~CXeqpC3BjqWMN@}4CWQQRVkLwk&HB&I9
zn+uTo<~p@KW>A&^NcS#3_oCWbV4f?H0qr+P1_}hN5RPWar-dy+uvNz02bc?Li-A;`
zopHfPO(4}6=UMS~ACP`Rb>LZVW6>`AzPQs$AoBx{=E=6CQ<>~w>N{`89|BS__M(v4
z@GP|r0o{uFj^g(ce*j2flX9{!g)`VGGEY-=Uy$-Z<$+2Ag>Mjbmz1u`k={F)fX0rIko!L_{rNMWZ)^poo-ui9G_Yaz}
z=t&mmEGb}tAV8UM=e0Bony8kbCeDHk-xuyX&ITQy&yJ4?raboi^NEkoPfI&=*`@kG
zLSLGjL|01OOlTvQHDxg6N}2F{i8FHy-lFv5pL1}PZ!a#{&iv5Qm6B#ZO|TyF?fLMZ
zt(gl*<(U6(4vpJjZZSL9dXpU64c!3MChgM{ytB!VmjZMPTX@AHDg767Xf^4sxq)O$
zgXRwr*%L)GCAZav^EmNnCmzR%vp#UNH`D~A7}eKg!cc+Az4qu4gIq~*Z~LA?c#(B#
zB=+eHo(PZRI!j|kzGF!-HHCx_?dL>GZB3txhf=Asof8v-IN5g{OF04Ypp#W9KhH+N
zL`j^Fy0cQEYPCbq;a`WhL}zq3W;z^@
z4wR(sfxwi76lb+o8k)=ds>%9lD<~Kq7824v1%{X`#4ouAo|oYHwAasT;J$Th+g4am
zz$u>C<%jIyQvERkNMZJ5NF+UxqIEWQ1%d`UXcS`829^dY{vgou?wGU>;9`=T<_Dn~
z-Z17ZK3r2$*tgQYe7NH^8)8-v%beETX=Rl`=-E&oGHy
zmM~4Ola-3;wob{`79Sk4($xD!Iv~Cy5&L~sc*Y$sMqDw(4pvj@l@^Lv=tyNH$`QQ7
zh3)ff2BM}^Eh5Epx@8op`qhS-C!s`augf)G(qUGYhlfmf5$R;Q{6K5J&Pz9)#h}R~
z+Xh)(+IO5M9&Zo4zrW*rd$a5Dj`Mure4GZj;#mx7AMMh<$v(#rB};qL#=p9%c!H@`
zZTdHXR4q%gl35=B_w)M0K+1ezJpXPEk9BeNVY$2y
z>mH)bT$?A^eh-LR2$=xF)=sBGb_O;vQtHW#?L5&OL`mus%hEn(o)PC?AM8BpO@|M^
zw^c0hFNz*DEQk@5;ovV0d=p5;VL(krsvbElT`etbD5RVGdZgfrA;C!X<&&z`om1TV
zbhqO509MFdngv>v!e}AyD~pwCbxHeloPdhUQ~i-2x{W@;=td!Dq_~9W)B(Deo)d60
zqudUne2utEL2q@c1=yz>q>O?(`)C+yzE3CyZro0JDF%k7ULwu_74y$BhRbDwL=M$M5ln0^%pQnw9RgO#$rsCa$i^ZE3s^9u6LF(h_7t6C&6
zQb}Ww`nCpS0x6jRQL{QlO`%|C1c7TOU6va@Jo`L?p~_yE&99`C_KBb0ms#4US(;7X
zcc3xKK-v7W*y?jy
zH96mTKJa$F;j9m|`amfsT796F6HqeCh_7K*CE%)&F(#gi&`V-6&2@9Kmm=He
zIm>pPedFaK{CbAV(ms5Z_L2H9Q~azL%$(EUI)mc6|5;^vTcr|YOxk|s3(y`KpnJpk
zk{S`<@){OG5r!zEM7bJe5c_*nZe}pruTS!$FJ354
za*>%rl%jaOv~?C->k6C5<0{gZ@fFd_DMeEWcVw&k;i^PgYOzC}UzD!dMK?n}BqeGn
z_6`+I-6|IX)B*-pFp_OVGam@w@Oct%FPC~jks4{jsuKVtgenH(Q
z)ry7n9IZgK&Y_aM2LYn&XFkr;82+=G%#RhLSeolm38R_^XcS{!YI=sJAJu&mtk|T~
zR?Xrd$_s^AWtC6;=wLwGkuldmkPlhd56_G&AGIRch_bI-XE>34E8@~nUENrz115%f
z5pbcAN;(7`9fb%NmI@%AA*Ka{+HvAgLUmtM3F(g4GuVKs&MrCYbYeeTxsP`;xz_bG
z6CSBfZ>2>v`@W_rxFYs%&97elz#i*NF;IdrRuhYnc@Vi;jQ~}cT8G|CfVuwTl(zgtaD;65pIg|yXuc+q
z_eZYF{pgVt7ojiwX;oQ{Zum#LN!
zu`@^j=hvjy)jP&>&!A+3SOvTe%;tLS_7sZJU7a5J9O9Hg+Lv>kCKK?Mp5$CSN#q}6
zs29G&QjmiQo+nbdoo()3p`mGT=oXj6Fy}~o{R?o+e;G*W57ukKtF&v;s+99)wusdg
zeMnC4ZKyh;z9;UmLp&8rJwl*nNkyPRNjgq3;AHZ;;&I0Yq`XUp%be`*Obu36{Sat3
zHcD>m;L)q2Ujv`dfzM~h=hGOe>-GZMv=Fz(eh3X#%0N|2&Vw>jRg3}L1Ln6HmfojK
z;HD@I5V6B|ap(AwDI^eu6~%6G*EjmGOPms2hh6fb=oN*Usr!&mGUenw-;^D|MzLqf
zU~5Xol5Ap@`2BS^WWMiWvA6fw<$bd^siWWbTy&z;ca`j|pEn}H4S;oA*KLP^=!kMK
zg)#h)Gr%4;Axqk8hY^ZMR9&G4N|0g)t?I8*N(4(mV4fO8nonU=FXWBbaGr-V>k%T_
zlf9IZgo123T%47ub)XhTFE(T#7jvsRO&ke!D<9R&OuTBnj(dB1~M%92)#&`=N{(q
z=kYaVqvCE(NTDz^|E@Ktc7Rkpn!Z(kZQxqd6@bcpD@@xEXGvRhGeIxVkUIc*+YgX?
z=`f6poLEYb8_Ucm!Shv>JUKP16`T$6vhDls?v=a%=^zph7b^2-nbGPZY~I4PXGV;De5O{gOn+*WxgBxl+?jwu=IJV>0vKBaa>xUjt?w#oKPHe$ud@s^`C}%d8xr-+`JKiTX84;v7;pt~
z1-Jp_29;)CN63r6mk3$i)_Rb^)RL;@P9<^`&Sa}J=cncpXH5*Y_x-S^T-w-gS`k$D
z$a~e~Oq!CoNRMYc8oH*KvR!`J=qI$3dqk7!F%N6}z
z$*O_1zKQ#oW%&a)2{p~Lf~kjizu7{Xj~9l@yXW$H9CiWBXU3j0P6fl;*+1<+N=f=Y
zYl(&jlWI5~=9zYd!l-$uJX``^9v`1$q@WtWD{Bh*7yjo@0jXD}R)6Q~e~$R^^1IkJ
z_|VU*q*%&aJV~=_38_D7$^D#>aoL^7N*W{DH)1BBcsfui$>8eJrs~twi*C?Fo2h
z#AVX0!)JEaXaT=6JShMmrTg~(&-KK<$Y2&$
zngm6<%*3thdExndx*qw0j2k*UM%m4eG4zsU;+DY-PK=Vm?E9!!x-p8A!3B$}P>VqE$gJ!z8S=ppN14Kc5nf?hAHiExjPO
zR`X_^1IHLHUu4`DO|z@Ll{rj7-S^aR9tR$e2i||ZXh?-UDHs^F-i)_q|`uA;qAM!LeN>VbN2aVA$OYFRF
zJg?g&z9w5`3{-#rYqL}~1Ke$bC$8&foxZ~Oj8xHCakHW^(4C#?MNvsnNaq%BEL0Rw
zuEPw?3GM<=Y&upS()@|Tfu>HN>6=VlZA245MVAfK%YogaKr!oemBtu8IC20*B;c`z
zQwBzs%qDKyIr#AVGKPH8OQuxjVbOpTPcJYP_xrE~fRs|5MJO1hSpm8nP?z$POZh3f
z;;2qQ#Z;C=77>(^0%LSd)WxNVn*5n@AI`yBq%JZ5E%rQe1xmFp$9e!&K&1l2z7wpy
zDl@f|K-=;dF90>%heg>VE{nZdlPuH5nx7Thwbtl#R!P8dmN0KY4>X@eyKjxPCDClg
z`FYp`>nu`dGAZhSJu?CYt1(llV#sf87jcm!&s5Hm`)Va2-NPp+z>g2!C%firBs+kL
z()cpv;F5q;gx}B_#M8plSpsLisD@T8Xz4uOa0VMyPBf1?w`aBh!;`49UwejotYl)_
z_xb|Dl!%Iu;TMy3GRwDTDoG$pR9vCG(hEQWb_G-^e%5xCS3dAlwu^++FAkXTAq-F2
zb01&y?b&0**+~|(0A{KC95rLIPz>GEdvjW${x?ed^qYVB`oF)!4(#)=q_
z(QdND7U5MCh?o}gKFOK@u%q`IyV7Fqqq2>hq+O#S^Sfc^Y_fC`V4%k-F*zI!G?r35X}#Ypx0wj#ynZHvP&
zUxiAavOWtH7+`a@+nkm$lfzVa9tb(^aSx!7VjyZrK^5^Cuh*CQ0k}nZs9)uUiHJZI
zPTlpv%prhcDF8bYHRHb-qx0+uv4u*P{4)-}ey>I+S!)C5u_Rm?4BqINCcw;g`tx<1
z2hPWdx3@REzrW+{*E`<-?HAtP-|={VU&m@n>%~i~u?-p=NO>Pve8yu4ga>HQWWfx?
z)%mq0cGlEW^$Mcw{<%x}(Onrq55capFZ~q5eXK}&CQB6J08SAeQM}H;zvjYIex7lM
zY+81#19?27ktXGb^Vz|XWqsQM8;_$8tsFF$xH2R44J$Q)I=hZuBt_-!=Iy!9=ecl5
zsz}%TRu{`Q-+*Z&)IIEl)DLBr_NQPoV5f<#^|##^_l5hqa6O;q#qswtBG5}^>u5U<
zoTtb12Y~W>h`D%2i_XKx7=q4<=FHTc7^&(#guzOwKdS+zHXnx{V520NA;vO#IwAc~
z*|%_7N*$Z`SNNb(A!v>xP?wBr3@v@gARaY5hE+L0FcKrRJQO^Q>T@tT1Va^IAMz1l
z1oIUKWz{7QmRY(T;3>&bgON(frCE)!
zsxFW`nXVLxd(0VAa74-bCRfyS_@%j@c>vs?RG`Gu>#1ZNlbNO6JyKa>zq3-Jt8Yq#
zW)Ik6cHRFJz_I|PP
zyu#W+)C^Sk-2p4TCGhB2YEy=g;4S@FWr2EOw`5}Ra~}ky`eiz3oh2X68$TF
z7y&6nEr{a271vWys^UH^)lUXPq;g~YZpzH93Jq2_-NZOHz+O;SX#!*(C;ce-(3_w1rEVvxOq
zE5+5r4@NHSyKZNsR6x=Jy-laWyQp6C}+Q%j*tK#JBv
z---dLqZJ&rnmJ!(J1K2gRTq$2G2jIhQmxvpN}v>s%Bx{~PPzJGw5I1R{80yvq1(nQ
zBA&*@5YjD`jbf@-D-LI*-rnEv_Wp)n@9%j3^~>3*_mnZGEA&EES+dTl_Lx4lAt?=v
z2L`VIgOFbZkP7K;$oe7`aF!3Qyjzr?0x8um=$?J?Uul}oN%~q=3W2IWKP6Z{2U6h!
z-GG$NDRxZpB2)#@`Ql%b=nO6+)pT&+Wvq`v%p^V`&<2zut!`e~XcJVHJvYtA8&m|V^$YxW*MlMF^?x$bZ<)yIW?KTVpON*oI@
zS}S;aG&~*$-rgQ~yuEEeid|AW*Hpjvx`<2q(iKcpx;XLi6ktmAi;83V-5f-jbn(Z6
z`}kWnIp_ozHDpvv-ZZE4QOJv8J`I=q&i55ScP5G<7&I_k9isEYD=UpI1;nTfvpM*7
z(9@Jb`LQOe=IygEN)ll8&Zc-z~;56$x`6FjZp7EVA)Cj+ID
zpS9)lx_U>ajy{Yz%XI~Ff=ihOGKrtML~%*`5{SwvZ^^iyeNju|7yq1-WC>f;b?i56
z(qIdpGHIVW)uxhT)|0$ecljc1a?KRBAest(|9C!sR3^m1x;UUMehF+ZG4ZBAvu0^!
zf;bDon@@+lH*ZPxG^O;VITF1DP!O$-x7bIbRwOQ$-%mfLl0Ba+UP
z(*FXG$~QyC_L?TInpup01fY`knmwIoF+vNptb=-D6yhKS1rIB)<(raxh`#MV&Wx)O>Gun=fOfvWkLIJ!Hq3U>xG%`{gxq#%zkqRD6svX%q6se%d>x_a1M&mPly=hU%W`HVA6R~LYi4OR
z7DE(8yo&)EtI&Q5&veES$Mjo=F-epYQiwo<&q-iUo9uj1
zNVVaZHvf5w*B+Er(o2;5fDu5Fq@QN{;i(E(>#^^-e$`f&Y%IJws!18L9!~N5tgt1d
z5;4x>#MvI()%ZZ=hC+2Rzj~%35m