Deep Learning Models

A collection of various deep learning architectures, models, & their implementations (PyTorch or TensorFlow) in Jupyter Notebooks.

Convolutional Neural Networks (CNN)

LeNet

Title	Dataset	Notebooks
LeNet-5	MNIST
LeNet-5	CIFAR-10

AlexNet

Title	Dataset	Notebooks
AlexNet	CIFAR-10

VGG-16

Title	Dataset	Notebooks
VGG-16	CIFAR-10

Network-in-Network (NiN)

Title	Dataset	Notebooks
NiN	CIFAR-10

GoogLeNet

Title	Dataset	Notebooks
GoogLeNet	CIFAR-10

ResNet

Title	Dataset	Notebooks
ResNet-18	CIFAR-10
ResNet-34	CIFAR-10
ResNet-50	CIFAR-10
ResNet-101	CIFAR-10
ResNet-152	CIFAR-10

DenseNet

Title	Dataset	Notebooks
DenseNet-121	CIFAR-10

MobileNet

Title	Dataset	Notebooks
MobileNet	CIFAR-10
MobileNet-V2	CIFAR-10
MobileNet-V3-Small	CIFAR-10
MobileNet-V3-Large	CIFAR-10

ShuffleNet

Title	Dataset	Notebooks
ShuffleNet	CIFAR-10
ShuffleNet-V2	CIFAR-10

EfficientNet

Title	Dataset	Notebooks
EfficientNet-B0	CIFAR-10

Recurrent Neural Networks (RNN)

Many-to-one: Sentiment Analysis / Classification

Title	Dataset	Notebooks
Simple single-layer RNN (Elman)	IMDB
Simple single-layer RNN (Elman)	AG News
Simple single-layer LSTM	IMDB
Simple single-layer LSTM	AG News
Simple single-layer GRU	IMDB
Simple single-layer GRU	AG News
Bidirectional GRU	AG News
Stacked GRU	AG News
Stacked Bidirectional GRU	AG News

Many-to-Many / Sequence-to-Sequence: Text Generation

Title	Dataset	Description	Notebooks
Character-level RNN (GRU)	Tiny Shakespeare
Character-level RNN (GRU)	Wikitext
RNN Encoder-Decoder (Seq2Seq)	Multi30k	Implementation of Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation (2014)
Encoder-Decoder with Bahdanau Attention	Multi30k	Implementation of Neural Machine Translation by Jointly Learning to Align and Translate (2016)

Many-to-Many: Named Entity Recognition (NER)

Title	Dataset	Description	Notebooks
Bi-LSTM + CRF	CoNLL-2003	Implementation of Bidirectional LSTM-CRF Models for Sequence Tagging (2015)
CNN-Bi-LSTM + CRF	CoNLL-2003	Implementation of End-to-end Sequence Labeling via Bi-directional LSTM-CNNs-CRF (2016)

Many-to-Many: Video Action Recognition (Synchronous)

Title	Dataset	Description	Notebooks
Spatial CNN + Bi-LSTM	UCF-101	Synchronous Many-to-Many video tagging using a frozen ResNet-18 backbone and a Bidirectional LSTM for frame-level action segmentation.

Currently ongoing ...

---