DNN from Scratch
Implementing core deep learning layers and training mechanics from first principles, with each project ready to link to its GitHub repository as the code is published.
Implementing core deep learning layers and training mechanics from first principles, with each project ready to link to its GitHub repository as the code is published.
ANN Fully connected networks, backpropagation, activations, and optimizers in NumPy
CNN Convolution, pooling, im2col, batch norm, and image-classification training loops
RNN Sequence modeling, truncated BPTT, embeddings, and recurrent hidden state dynamics
LSTM Gates, cell state, gradient flow, and sequence prediction experiments