added readme to resnet and efficient, also made minor fix to unet (#87)

* added readme to resnet and efficient, also made minor fix to unet

* changed RESNET to ResNet

ivy_models/efficientnet/README.rst

@@ -0,0 +1,91 @@
+.. image:: https://github.com/unifyai/unifyai.github.io/blob/main/img/externally_linked/logo.png?raw=true#gh-light-mode-only
+   :width: 100%
+   :class: only-light
+
+.. image:: https://github.com/unifyai/unifyai.github.io/blob/main/img/externally_linked/logo_dark.png?raw=true#gh-dark-mode-only
+   :width: 100%
+   :class: only-dark
+
+
+.. raw:: html
+
+    <br/>
+    <a href="https://pypi.org/project/ivy-models">
+        <img class="dark-light" style="float: left; padding-right: 4px; padding-bottom: 4px;" src="https://badge.fury.io/py/ivy-models.svg">
+    </a>
+    <a href="https://github.com/unifyai/models/actions?query=workflow%3Adocs">
+        <img class="dark-light" style="float: left; padding-right: 4px; padding-bottom: 4px;" src="https://github.com/unifyai/models/actions/workflows/docs.yml/badge.svg">
+    </a>
+    <a href="https://github.com/unifyai/models/actions?query=workflow%3Anightly-tests">
+        <img class="dark-light" style="float: left; padding-right: 4px; padding-bottom: 4px;" src="https://github.com/unifyai/models/actions/workflows/nightly-tests.yml/badge.svg">
+    </a>
+    <a href="https://discord.gg/G4aR9Q7DTN">
+        <img class="dark-light" style="float: left; padding-right: 4px; padding-bottom: 4px;" src="https://img.shields.io/discord/799879767196958751?color=blue&label=%20&logo=discord&logoColor=white">
+    </a>
+    <br clear="all" />
+
+EfficientNet
+===========
+
+`EfficientNet <https://arxiv.org/abs/1905.11946>`_ represents a convolutional neural network structure and scaling approach that consistently adjusts depth, 
+width, and resolution using a combined coefficient. Diverging from traditional methods that unevenly modify these aspects, 
+the EfficientNet approach maintains uniform adjustments to network dimensions by utilizing a predefined set of scaling coefficients. 
+
+This method employs a compound coefficient to ensure consistent modifications in network attributes, resulting in a more systematic approach.
+The rationale behind the compound scaling method is intuitive: as the input image size increases, 
+the network requires additional layers to enhance its receptive field, along with increased channels to capture intricate patterns specific to larger images.
+
+Getting started
+-----------------
+
+.. code-block:: python
+
+    import ivy
+    from ivy_models.efficientnet import efficientnet_b0
+    ivy.set_backend("torch")
+
+    # Instantiate efficientnet_b0 model
+    ivy_efficientnet_b0 = efficientnet_b0(pretrained=True)
+
+    # Convert the Torch image tensor to an Ivy tensor and adjust dimensions
+    img = ivy.asarray(torch_img.permute((0, 2, 3, 1)), dtype="float32", device="gpu:0")
+
+    # Compile the Ivy efficientnet_b0 model with the Ivy image tensor
+    ivy_efficientnet_b0.compile(args=(img,))
+
+    # Pass the Ivy image tensor through the Ivy efficientnet_b0 model and apply softmax
+    output = ivy.softmax(ivy_efficientnet_b0(img))
+
+    # Get the indices of the top 3 classes from the output probabilities
+    classes = ivy.argsort(output[0], descending=True)[:3] 
+
+    # Retrieve the logits corresponding to the top 3 classes
+    logits = ivy.gather(output[0], classes) 
+
+    print("Indices of the top 3 classes are:", classes)
+    print("Logits of the top 3 classes are:", logits)
+    print("Categories of the top 3 classes are:", [categories[i] for i in classes.to_list()])
+
+    `Indices of the top 3 classes are: ivy.array([282, 281, 285], dev=gpu:0)`
+    `Logits of the top 3 classes are: ivy.array([0.60317987, 0.18620452, 0.07535177], dev=gpu:0)`
+    `Categories of the top 3 classes are: ['tiger cat', 'tabby', 'Egyptian cat']`
+
+Citation
+--------
+
+::
+
+    @article{
+      title={EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks},
+      author={Mingxing Tan and Quoc V. Le},
+      journal={arXiv preprint arXiv:1905.11946},
+      year={2020}
+    }
+
+
+    @article{lenton2021ivy,
+      title={Ivy: Templated deep learning for inter-framework portability},
+      author={Lenton, Daniel and Pardo, Fabio and Falck, Fabian and James, Stephen and Clark, Ronald},
+      journal={arXiv preprint arXiv:2102.02886},
+      year={2021}
+    }

ivy_models/resnet/README.rst

@@ -0,0 +1,96 @@
+.. image:: https://github.com/unifyai/unifyai.github.io/blob/main/img/externally_linked/logo.png?raw=true#gh-light-mode-only
+   :width: 100%
+   :class: only-light
+
+.. image:: https://github.com/unifyai/unifyai.github.io/blob/main/img/externally_linked/logo_dark.png?raw=true#gh-dark-mode-only
+   :width: 100%
+   :class: only-dark
+
+
+.. raw:: html
+
+    <br/>
+    <a href="https://pypi.org/project/ivy-models">
+        <img class="dark-light" style="float: left; padding-right: 4px; padding-bottom: 4px;" src="https://badge.fury.io/py/ivy-models.svg">
+    </a>
+    <a href="https://github.com/unifyai/models/actions?query=workflow%3Adocs">
+        <img class="dark-light" style="float: left; padding-right: 4px; padding-bottom: 4px;" src="https://github.com/unifyai/models/actions/workflows/docs.yml/badge.svg">
+    </a>
+    <a href="https://github.com/unifyai/models/actions?query=workflow%3Anightly-tests">
+        <img class="dark-light" style="float: left; padding-right: 4px; padding-bottom: 4px;" src="https://github.com/unifyai/models/actions/workflows/nightly-tests.yml/badge.svg">
+    </a>
+    <a href="https://discord.gg/G4aR9Q7DTN">
+        <img class="dark-light" style="float: left; padding-right: 4px; padding-bottom: 4px;" src="https://img.shields.io/discord/799879767196958751?color=blue&label=%20&logo=discord&logoColor=white">
+    </a>
+    <br clear="all" />
+
+ResNet 
+===========
+
+`ResNet <https://arxiv.org/abs/1512.03385>`_ also known as Residual Network is a deep learning model utilized in computer vision tasks. 
+It's a type of Convolutional Neural Network (CNN) architecture designed to accommodate a large number of convolutional layers, 
+possibly ranging from hundreds to thousands. ResNet addresses the challenge of diminishing gradients, a problem encountered during training, 
+by introducing a concept called "skip connections." These connections enable the network to skip over several initial layers, 
+which consist of identity mappings (convolutional layers with no immediate impact). 
+This results in faster initial training by condensing the network into fewer layers.
+
+During subsequent retraining, the network expands to its full depth, and the sections of the network that weren't initially utilized, 
+called "residual parts," are given the opportunity to explore the features present in the input image more comprehensively.
+
+Getting started
+-----------------
+
+.. code-block:: python
+
+    import ivy
+    from ivy_models.resnet import resnet_34
+    ivy.set_backend("torch")
+
+    # Instantiate Resnet_34 model
+    ivy_resnet_34 = resnet_34(pretrained=True)
+
+    # Convert the Torch image tensor to an Ivy tensor and adjust dimensions
+    img = ivy.asarray(torch_img.permute((0, 2, 3, 1)), dtype="float32", device="gpu:0")
+
+    # Compile the Ivy Resnet_34 model with the Ivy image tensor
+    ivy_resnet_34.compile(args=(img,))
+
+    # Pass the Ivy image tensor through the Ivy Resnet_34 model and apply softmax
+    output = ivy.softmax(ivy_resnet_34(img))
+
+    # Get the indices of the top 3 classes from the output probabilities
+    classes = ivy.argsort(output[0], descending=True)[:3] 
+
+    # Retrieve the logits corresponding to the top 3 classes
+    logits = ivy.gather(output[0], classes) 
+
+
+    print("Indices of the top 3 classes are:", classes)
+    print("Logits of the top 3 classes are:", logits)
+    print("Categories of the top 3 classes are:", [categories[i] for i in classes.to_list()])
+
+    `Indices of the top 3 classes are: ivy.array([282, 281, 285], dev=gpu:0)`
+    `Logits of the top 3 classes are: ivy.array([0.85072654, 0.13506058, 0.00688287], dev=gpu:0)`
+    `Categories of the top 3 classes are: ['tiger cat', 'tabby', 'Egyptian cat']`
+
+See `this demo <https://unify.ai/demos/examples_and_demos/resnet_demo.html>`_ for more usage example.
+
+Citation
+--------
+
+::
+
+    @article{
+      title={Deep Residual Learning for Image Recognition},
+      author={Kaiming He, Xiangyu Zhang, Shaoqing Ren and Jian Sun},
+      journal={arXiv preprint arXiv:1512.03385},
+      year={2015}
+    }
+
+
+    @article{lenton2021ivy,
+      title={Ivy: Templated deep learning for inter-framework portability},
+      author={Lenton, Daniel and Pardo, Fabio and Falck, Fabian and James, Stephen and Clark, Ronald},
+      journal={arXiv preprint arXiv:2102.02886},
+      year={2021}
+    }

ivy_models/unet/README.rst

@@ -27,7 +27,7 @@
 U-Net
 ===========
 
-`Unet <https://arxiv.org/abs/1505.04597>`_ The UNET architecture and training approach effectively leverage data augmentation to make the most of 
+The `U-Net <https://arxiv.org/abs/1505.04597>`_  architecture and training approach effectively leverage data augmentation to make the most of 
 available annotated samples, even with limited data. The design features a contracting path for context capture and a symmetric expanding path 
 for precise localization. Notably, this UNET network achieves superior performance with minimal images during end-to-end training. 
 It surpasses previous methods, including a sliding-window convolutional network, in the ISBI challenge for segmenting neuronal structures in