Deep Learning Projects

• Hyperparameter Tuning: In this notebook, an Artificial Neural Network was trained to perform Prediction of Body Fat. The hyperparameters of the network were tuned and the corresponding accuracy and loss for each combination was plottted and analyzed.
• LeNet CIFAR10: In this notebook, the LeNet architecture was trained and tested on the CIFAR10 dataset.
• LeNet MNIST: In this notebook, the LeNet architecture was trained and tested on the MNNIST dataset.
• AlexNet Cats & Dogs: In this notebook, the AlexNet architecture was trained and tested on the Cats & Dogs Dataset.
• Transfer Learning: In this notebook, the Keras Applications API was used to implement Transfer Learning of the VGG-16/ InceptionNet/ ResNet/ EfficientNet/ MobileNet on the Horse2Zebra dataset.
• Time Series Modelling: In this notebook, a Vanilla RNN was used to predict stock price using 50 timesteps.
• LSTM & Bidirectional RNN: In this notebook, a Long-Short-Term-Memory Network and a Bidirectional Recurrent Neural Network was trained and tested on the IMDB Reviews Dataset for Sentiment Analysis.
• Functional API: In this notebook, the Functional API was used to train a branched Neural Network and its performance was compared with that of the Sequential API.
• Wide & Deep and Deep & Cross Networks: In this notebook, the wide part is a linear model using sparse representation, the deep part is a feed-forward network using dense representation, and the cross part applies explicit feature crossing.