Neural Networks from Scratch Pinned

Project summary

A from-scratch tensor operation and automatic differentiation library in pure Python, featuring feed-forward mechanisms, backpropagation, loss functions (MSE, cross-entropy, etc.), extensible optimizers (SDG, Adam, etc.), extensible modules (linear, sequential, etc.), activations (ReLU, Tanh, etc.), functional formulations (like torch.nn.functional), persistent and temporary buffers, and saving state dicts.