Projects &
experiments.

Building image-captioning models, from CNN–LSTM pipelines to attention and Transformer decoders.

Implementing a Vision Transformer step by step, training on CIFAR-10, and inspecting where the model attends.