From d975c1f0fb06d92695349b3e2d8362b135e5c217 Mon Sep 17 00:00:00 2001 From: knikolaou <> Date: Tue, 3 Oct 2023 15:21:20 +0200 Subject: [PATCH] Implement regularization schedules. A schedule is a function that depends on the current epoch and rescales the regularization factor. This function can also be defined by the user. --- examples/trace_regularization.ipynb | 144 +++++++++--------- znnl/regularizers/__init__.py | 4 +- .../regularizers/grad_variance_regularizer.py | 79 ---------- znnl/regularizers/norm_regularizer.py | 29 ++-- znnl/regularizers/regularizer.py | 94 +++++++++++- znnl/regularizers/trace_regularizer.py | 26 ++-- .../loss_aware_reservoir.py | 2 + .../partitioned_training.py | 2 + znnl/training_strategies/simple_training.py | 4 + 9 files changed, 200 insertions(+), 184 deletions(-) delete mode 100644 znnl/regularizers/grad_variance_regularizer.py diff --git a/examples/trace_regularization.ipynb b/examples/trace_regularization.ipynb index b3e38d7..c39583d 100644 --- a/examples/trace_regularization.ipynb +++ b/examples/trace_regularization.ipynb @@ -10,9 +10,9 @@ "name": "stderr", "output_type": "stream", "text": [ - "2023-09-27 18:36:45.795748: E external/xla/xla/stream_executor/cuda/cuda_driver.cc:268] failed call to cuInit: CUDA_ERROR_NO_DEVICE: no CUDA-capable device is detected\n", + "2023-10-03 14:12:49.333240: E external/xla/xla/stream_executor/cuda/cuda_driver.cc:268] failed call to cuInit: CUDA_ERROR_NO_DEVICE: no CUDA-capable device is detected\n", "No GPU/TPU found, falling back to CPU. (Set TF_CPP_MIN_LOG_LEVEL=0 and rerun for more info.)\n", - "2023-09-27 18:36:47.976976: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n" + "2023-10-03 14:12:51.662550: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Could not find TensorRT\n" ] }, { @@ -87,7 +87,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 2, "id": "afc55b14", "metadata": {}, "outputs": [], @@ -111,7 +111,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 12, "id": "19f5363d", "metadata": {}, "outputs": [], @@ -445,7 +445,7 @@ }, { "cell_type": "code", - "execution_count": 2, + "execution_count": 3, "id": "af442d14-0791-48cc-a9e8-aa0c5ee9f9c4", "metadata": {}, "outputs": [ @@ -453,7 +453,7 @@ "name": "stderr", "output_type": "stream", "text": [ - "2023-09-27 18:36:51.950437: E tensorflow/compiler/xla/stream_executor/cuda/cuda_driver.cc:266] failed call to cuInit: CUDA_ERROR_NO_DEVICE: no CUDA-capable device is detected\n" + "2023-10-03 14:12:55.737167: E tensorflow/compiler/xla/stream_executor/cuda/cuda_driver.cc:266] failed call to cuInit: CUDA_ERROR_NO_DEVICE: no CUDA-capable device is detected\n" ] } ], @@ -463,7 +463,7 @@ }, { "cell_type": "code", - "execution_count": 3, + "execution_count": 4, "id": "11123b2a-b981-4218-98bf-47b0a2bfc271", "metadata": {}, "outputs": [], @@ -487,7 +487,7 @@ }, { "cell_type": "code", - "execution_count": 4, + "execution_count": 5, "id": "7936f03f-ee9b-46cb-a399-ba916cad09c2", "metadata": {}, "outputs": [], @@ -536,48 +536,50 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 6, "id": "4ce72747", "metadata": {}, - "outputs": [ - { - "ename": "ImportError", - "evalue": "cannot import name 'GradVarianceRegularizer' from 'znnl.regularizers' (/tikhome/knikolaou/work/Repositories/ZnRND/znnl/regularizers/__init__.py)", - "output_type": "error", - "traceback": [ - "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", - "\u001b[0;31mImportError\u001b[0m Traceback (most recent call last)", - "Cell \u001b[0;32mIn[45], line 1\u001b[0m\n\u001b[0;32m----> 1\u001b[0m \u001b[39mfrom\u001b[39;00m \u001b[39mznnl\u001b[39;00m\u001b[39m.\u001b[39;00m\u001b[39mregularizers\u001b[39;00m \u001b[39mimport\u001b[39;00m TraceRegularizer, NormRegularizer, GradVarianceRegularizer\n", - "\u001b[0;31mImportError\u001b[0m: cannot import name 'GradVarianceRegularizer' from 'znnl.regularizers' (/tikhome/knikolaou/work/Repositories/ZnRND/znnl/regularizers/__init__.py)" - ] - } - ], + "outputs": [], "source": [ - "from znnl.regularizers import TraceRegularizer, NormRegularizer, GradVarianceRegularizer" + "from znnl.regularizers import TraceRegularizer, NormRegularizer\n", + "from znnl.training_strategies import SimpleTraining" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 7, + "id": "6b6bece6", + "metadata": {}, + "outputs": [], + "source": [ + "loss_fn = znnl.loss_functions.CrossEntropyLoss(),\n", + "\n", + "def reg_schedule_fn(epoch, reg_factor):\n", + " return reg_factor * 0.9 ** epoch\n", + "\n", + "regularizer = NormRegularizer(reg_factor=1e-2, reg_schedule_fn=reg_schedule_fn)" + ] + }, + { + "cell_type": "code", + "execution_count": 8, "id": "05e60cd9", "metadata": {}, "outputs": [], "source": [ - "trainer = RegularizedTraining(\n", + "trainer = SimpleTraining(\n", " model=model, \n", " loss_fn=znnl.loss_functions.CrossEntropyLoss(),\n", " accuracy_fn=znnl.accuracy_functions.LabelAccuracy(), \n", " recorders=[train_recorder, test_recorder], \n", - " regulizer=TraceRegularizer(0.1),\n", - " # regularization=1e-2, \n", - " # # regularization=0.0,\n", + " regularizer=regularizer,\n", " seed=0\n", ")" ] }, { "cell_type": "code", - "execution_count": null, + "execution_count": 9, "id": "da9ecc3f-dab4-4bc6-bd3a-35a3e5b6f855", "metadata": {}, "outputs": [ @@ -585,7 +587,14 @@ "name": "stderr", "output_type": "stream", "text": [ - "Epoch: 100: 100%|███████████████████████████████| 100/100 [01:34<00:00, 1.06batch/s, accuracy=0.58]\n" + " 0%| | 0/100 [00:00" ] @@ -639,13 +648,13 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 12, "id": "a6fd3a3c", "metadata": {}, "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -664,13 +673,13 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 13, "id": "f8eea5f3", "metadata": {}, "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ "
" ] @@ -698,18 +707,18 @@ }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 16, "id": "8eec2bb3", "metadata": {}, "outputs": [], "source": [ - "from znnl.regularizers import TraceRegularizer, NormRegularizer, GradVarianceRegularizer, Regularizer\n", + "from znnl.regularizers import TraceRegularizer, NormRegularizer, Regularizer\n", "from znnl.training_strategies import SimpleTraining" ] }, { "cell_type": "code", - "execution_count": 8, + "execution_count": 31, "id": "1e35c600", "metadata": {}, "outputs": [], @@ -741,6 +750,7 @@ " test_recorder = znnl.training_recording.JaxRecorder(\n", " name=\"test_recorder\",\n", " loss=True,\n", + " accuracy=True,\n", " update_rate=1,\n", " chunk_size=1000\n", " )\n", @@ -772,21 +782,20 @@ }, { "cell_type": "code", - "execution_count": 17, + "execution_count": 32, "id": "d17182d1", "metadata": {}, "outputs": [], "source": [ "regularizers = [\n", " NormRegularizer(reg_factor=1e1),\n", - " GradVarianceRegularizer(reg_factor=1e-1),\n", - " TraceRegularizer(reg_factor=1e-1),\n", + " TraceRegularizer(reg_factor=5e-1),\n", "]" ] }, { "cell_type": "code", - "execution_count": 9, + "execution_count": 33, "id": "d1ecc3d3", "metadata": {}, "outputs": [ @@ -794,48 +803,35 @@ "name": "stderr", "output_type": "stream", "text": [ - "Epoch: 100: 100%|████████████████████████████████| 100/100 [00:21<00:00, 4.56batch/s, accuracy=0.6]\n", - "Epoch: 100: 100%|███████████████████████████████| 100/100 [01:30<00:00, 1.11batch/s, accuracy=0.58]\n", - "Epoch: 100: 100%|███████████████████████████████| 100/100 [01:39<00:00, 1.01batch/s, accuracy=0.58]\n" + " 0%| | 0/100 [00:00" + "
" ] }, "metadata": {}, @@ -846,25 +842,23 @@ "fig, axs = plt.subplots(1, 3, figsize=(15, 5), tight_layout=True)\n", "\n", "axs[0].plot(train_report_norm.loss, 'o', mfc='None', label=\"Train Norm\")\n", - "axs[0].plot(train_report_grad.loss, 'o', mfc='None', label=\"Train Var\")\n", "axs[0].plot(train_report_trace.loss, 'o', mfc='None', label=\"Train Trace\")\n", "\n", - "# axs[0].plot(test_report_norm.loss, '-', mfc='None', label=\"Test Norm\")\n", - "# axs[0].plot(test_report_grad.loss, '-', mfc='None', label=\"Test Var\")\n", - "# axs[0].plot(test_report_trace.loss, '-', mfc='None', label=\"Test Trace\")\n", - "\n", "axs[0].set_xlabel(\"Epoch\")\n", - "axs[0].set_ylabel(\"Loss\")\n", + "axs[0].set_ylabel(\"Train Loss\")\n", "axs[0].set_yscale(\"log\")\n", "\n", + "axs_twinx = axs[0].twinx()\n", + "axs_twinx.plot(test_report_norm.loss, '-', mfc='None', label=\"Test Norm\")\n", + "axs_twinx.plot(test_report_trace.loss, '-', mfc='None', label=\"Test Trace\")\n", + "axs_twinx.set_ylabel(\"Test Loss\")\n", + "\n", "axs[1].plot(train_report_norm.entropy, 'o', mfc='None', label=\"Norm\")\n", - "axs[1].plot(train_report_grad.entropy, 'o', mfc='None', label=\"Var\")\n", "axs[1].plot(train_report_trace.entropy, 'o', mfc='None', label=\"Trace\")\n", "axs[1].set_xlabel(\"Epoch\")\n", "axs[1].set_ylabel(\"Entropy\")\n", "\n", "axs[2].plot(train_report_norm.trace, 'o', mfc='None', label=\"Norm\")\n", - "axs[2].plot(train_report_grad.trace, 'o', mfc='None', label=\"Var\")\n", "axs[2].plot(train_report_trace.trace, 'o', mfc='None', label=\"Trace\")\n", "axs[2].set_xlabel(\"Epoch\")\n", "axs[2].set_ylabel(\"Trace\")\n", diff --git a/znnl/regularizers/__init__.py b/znnl/regularizers/__init__.py index 461a5df..3c787e4 100644 --- a/znnl/regularizers/__init__.py +++ b/znnl/regularizers/__init__.py @@ -24,14 +24,12 @@ Summary ------- """ -from znnl.regularizers.regularizer import Regularizer from znnl.regularizers.norm_regularizer import NormRegularizer +from znnl.regularizers.regularizer import Regularizer from znnl.regularizers.trace_regularizer import TraceRegularizer -from znnl.regularizers.grad_variance_regularizer import GradVarianceRegularizer __all__ = [ Regularizer.__name__, NormRegularizer.__name__, TraceRegularizer.__name__, - GradVarianceRegularizer.__name__, ] diff --git a/znnl/regularizers/grad_variance_regularizer.py b/znnl/regularizers/grad_variance_regularizer.py deleted file mode 100644 index be9f0b9..0000000 --- a/znnl/regularizers/grad_variance_regularizer.py +++ /dev/null @@ -1,79 +0,0 @@ -""" -ZnNL: A Zincwarecode package. - -License -------- -This program and the accompanying materials are made available under the terms -of the Eclipse Public License v2.0 which accompanies this distribution, and is -available at https://www.eclipse.org/legal/epl-v20.html - -SPDX-License-Identifier: EPL-2.0 - -Copyright Contributors to the Zincwarecode Project. - -Contact Information -------------------- -email: zincwarecode@gmail.com -github: https://github.com/zincware -web: https://zincwarecode.com/ - -Citation --------- -If you use this module please cite us with: - -Summary -------- -Module containing the trace regularizer class. -""" -from znnl.regularizers.regularizer import Regularizer -from typing import Callable -import jax.flatten_util -import jax.tree_util -import jax.numpy as np - - -class GradVarianceRegularizer(Regularizer): - """ - Regularizer class to regularize on the variance of the gradients. - - Regularizing the loss of gradient based learning proportional to the variance of the - gradients, as: - Var(grad) = E[(grad - E[grad])^2] - """ - - def __init__(self, reg_factor: float = 1e-1) -> None: - """ - Constructor of the gradient variance regularizer class. - - Parameters - ---------- - reg_factor : float - Regularization factor. - """ - super().__init__(reg_factor) - - def __call__(self, apply_fn: Callable, params: dict, batch: dict) -> float: - """ - Call function of the trace regularizer class. - - Parameters - ---------- - apply_fn : Callable - Function to apply the model to inputs. - params : dict - Parameters of the model. - batch : dict - Batch of data. - - Returns - ------- - reg_loss : float - Loss contribution from the regularizer. - """ - # Compute squared gradient of shape=(batch_size, n_outputs, params) - grads = jax.jacrev(apply_fn)(params, batch["inputs"]) - # Square the gradients and take the mean over the batch - grad_variance = jax.tree_util.tree_map(lambda x: np.var(x, axis=(0, 1)), grads) - raveled_grad_variance = jax.flatten_util.ravel_pytree(grad_variance)[0] - reg_loss = self.reg_factor * raveled_grad_variance.mean() - return reg_loss diff --git a/znnl/regularizers/norm_regularizer.py b/znnl/regularizers/norm_regularizer.py index 0e3c63c..c5eec10 100644 --- a/znnl/regularizers/norm_regularizer.py +++ b/znnl/regularizers/norm_regularizer.py @@ -24,13 +24,15 @@ Summary ------- """ -from znnl.regularizers.regularizer import Regularizer +from functools import partial from typing import Callable, Optional + import jax.flatten_util -import jax.tree_util import jax.numpy as np +import jax.tree_util from jax import jit -from functools import partial + +from znnl.regularizers.regularizer import Regularizer class NormRegularizer(Regularizer): @@ -38,13 +40,16 @@ class NormRegularizer(Regularizer): Class to regularize on the norm of the parameters. Regularizing training using the norm of the parameters. - Any function can be used as norm, as long as it takes the parameters as input + Any function can be used as norm, as long as it takes the parameters as input and returns a scalar. - The function is applied to each parameter + The function is applied to each parameter """ def __init__( - self, reg_factor: float = 1e-2, norm_fn: Optional[Callable] = None + self, + reg_factor: float = 1e-2, + reg_schedule_fn: Optional[Callable] = None, + norm_fn: Optional[Callable] = None, ) -> None: """ Constructor of the regularizer class. @@ -57,22 +62,22 @@ def __init__( Function to compute the norm of the parameters. If None, the default norm is the mean squared error. """ - super().__init__(reg_factor) + super().__init__(reg_factor, reg_schedule_fn) self.norm_fn = norm_fn if self.norm_fn is None: self.norm_fn = lambda x: np.mean(x**2) - - def __call__(self, params: dict, **kwargs: dict) -> float: + + def _calculate_regularization(self, params: dict, **kwargs: dict) -> float: """ - Call function of the trace regularizer class. + Calculate the regularization contribution to the loss using the norm of the Parameters ---------- params : dict Parameters of the model. kwargs : dict - Additional arguments. + Additional arguments. Individual regularizers can define their own arguments. Returns @@ -80,8 +85,6 @@ def __call__(self, params: dict, **kwargs: dict) -> float: reg_loss : float Loss contribution from the regularizer. """ - param_vector = jax.flatten_util.ravel_pytree(params)[0] reg_loss = self.reg_factor * self.norm_fn(param_vector) return reg_loss - \ No newline at end of file diff --git a/znnl/regularizers/regularizer.py b/znnl/regularizers/regularizer.py index e4158c2..c912bc8 100644 --- a/znnl/regularizers/regularizer.py +++ b/znnl/regularizers/regularizer.py @@ -24,7 +24,11 @@ Summary ------- """ +import logging from abc import ABC +from typing import Callable, Optional + +logger = logging.getLogger(__name__) class Regularizer(ABC): @@ -32,7 +36,9 @@ class Regularizer(ABC): Parent class for a regularizer. All regularizers should inherit from this class. """ - def __init__(self, reg_factor) -> None: + def __init__( + self, reg_factor: float, reg_schedule_fn: Optional[Callable] = None + ) -> None: """ Constructor of the regularizer class. @@ -40,20 +46,71 @@ def __init__(self, reg_factor) -> None: ---------- reg_factor : float Regularization factor. + reg_schedule_fn : Optional[Callable] + Function to schedule the regularization factor. + The function takes the current epoch and the regularization factor + as input and returns the scheduled regularization factor (float). + An example function is: + + def reg_schedule(epoch: int, reg_factor: float) -> float: + return reg_factor * 0.99 ** epoch + + where the regularization factor is reduced by 1% each epoch. + The default is None, which means no scheduling is applied: + + def reg_schedule(epoch: int, reg_factor: float) -> float: + return reg_factor """ self.reg_factor = reg_factor + self.reg_schedule_fn = reg_schedule_fn - def __call__(self, params: dict, **kwargs: dict) -> float: + if self.reg_schedule_fn: + logger.info( + "Setting a regularization schedule." + "The set regularization factor will be overwritten." + ) + if not callable(self.reg_schedule_fn): + raise TypeError("Regularization schedule must be a Callable.") + + if self.reg_schedule_fn is None: + self.reg_schedule_fn = self._schedule_fn_default + + @staticmethod + def _schedule_fn_default(epoch: int, reg_factor: float) -> float: """ - Call function of the regularizer class. + Default function for the regularization factor. + + Parameters + ---------- + epoch : int + Current epoch. + reg_factor : float + Regularization factor. + + Returns + ------- + scheduled_reg_factor : float + Scheduled regularization factor. + """ + return reg_factor + + def _calculate_regularization(self, params: dict, **kwargs: dict) -> float: + """ + Calculate the regularization contribution to the loss. Parameters ---------- params : dict Parameters of the model. kwargs : dict - Additional arguments. - Individual regularizers can define their own arguments. + Additional arguments. + Individual regularizers can utilize arguments from the set: + apply_fn : Callable + Function to apply the model to inputs. + batch : dict + Batch of data. + epoch : int + Current epoch. Returns ------- @@ -61,3 +118,30 @@ def __call__(self, params: dict, **kwargs: dict) -> float: Loss contribution from the regularizer. """ raise NotImplementedError + + def __call__( + self, apply_fn: Callable, params: dict, batch: dict, epoch: int + ) -> float: + """ + Call function of the regularizer class. + + Parameters + ---------- + apply_fn : Callable + Function to apply the model to inputs. + params : dict + Parameters of the model. + batch : dict + Batch of data. + epoch : int + Current epoch. + + Returns + ------- + scaled_reg_loss : float + Scaled loss contribution from the regularizer. + """ + self.reg_factor = self.reg_schedule_fn(epoch, self.reg_factor) + return self.reg_factor * self._calculate_regularization( + apply_fn=apply_fn, params=params, batch=batch, epoch=epoch + ) diff --git a/znnl/regularizers/trace_regularizer.py b/znnl/regularizers/trace_regularizer.py index b5de53b..840fa2d 100644 --- a/znnl/regularizers/trace_regularizer.py +++ b/znnl/regularizers/trace_regularizer.py @@ -25,25 +25,29 @@ ------- Module containing the trace regularizer class. """ -from znnl.regularizers.regularizer import Regularizer from typing import Callable + import jax.flatten_util import jax.tree_util +from znnl.regularizers.regularizer import Regularizer + class TraceRegularizer(Regularizer): """ Trace regularizer class. - Regularizing the loss of gradient based learning proportional to the trace of the + Regularizing the loss of gradient based learning proportional to the trace of the NTK. As: Trace(NTK) = sum_i (d f(x_i)/d theta)^2 - the trace of the NTK is the sum of the squared gradients of the model, the trace - regularizer is equivalent to regularizing on the sum of the squared gradients of + the trace of the NTK is the sum of the squared gradients of the model, the trace + regularizer is equivalent to regularizing on the sum of the squared gradients of the model. """ - def __init__(self, reg_factor: float = 1e-1) -> None: + def __init__( + self, reg_factor: float = 1e-1, reg_schedule_fn: Callable = None + ) -> None: """ Constructor of the trace regularizer class. @@ -51,17 +55,21 @@ def __init__(self, reg_factor: float = 1e-1) -> None: ---------- reg_factor : float Regularization factor. + reg_schedule_fn : Callable + """ - super().__init__(reg_factor) - - def __call__(self, apply_fn: Callable, params: dict, batch: dict) -> float: + super().__init__(reg_factor, reg_schedule_fn) + + def _calculate_regularization( + self, apply_fn: Callable, params: dict, batch: dict, epoch: int + ) -> float: """ Call function of the trace regularizer class. Parameters ---------- apply_fn : Callable - Function to apply the model to inputs. + Function to apply the model to inputs. params : dict Parameters of the model. batch : dict diff --git a/znnl/training_strategies/loss_aware_reservoir.py b/znnl/training_strategies/loss_aware_reservoir.py index d759163..4fc6cd4 100644 --- a/znnl/training_strategies/loss_aware_reservoir.py +++ b/znnl/training_strategies/loss_aware_reservoir.py @@ -418,6 +418,8 @@ def train_model( train_losses = [] train_accuracy = [] for i in loading_bar: + self.epoch = i + # Update the recorder properties if self.recorders is not None: for item in self.recorders: diff --git a/znnl/training_strategies/partitioned_training.py b/znnl/training_strategies/partitioned_training.py index a08632b..f1b6b39 100644 --- a/znnl/training_strategies/partitioned_training.py +++ b/znnl/training_strategies/partitioned_training.py @@ -278,6 +278,8 @@ def train_model( ) for i in loading_bar: + self.epoch = i + # Update the recorder properties if self.recorders is not None: for item in self.recorders: diff --git a/znnl/training_strategies/simple_training.py b/znnl/training_strategies/simple_training.py index 70803f9..1f8001c 100644 --- a/znnl/training_strategies/simple_training.py +++ b/znnl/training_strategies/simple_training.py @@ -112,6 +112,7 @@ def __init__( self.rng = PRNGKey(seed) self.review_metric = None + self.epoch = 0 # Add the loss and accuracy function to the recorders and re-instantiate them if self.recorders is not None: @@ -219,6 +220,7 @@ def loss_fn(params): apply_fn=self.model.apply, params=params, batch=batch, + epoch=self.epoch ) loss += reg_loss return loss, inner_predictions @@ -381,6 +383,8 @@ def train_model( train_losses = [] train_accuracy = [] for i in loading_bar: + self.epoch = i + # Update the recorder properties if self.recorders is not None: for item in self.recorders: