This is a jupyterlab workspace exploring various deep learning techniques for NLP and computer vision.
A majority of these notebooks are sourced from various tutorials, edited for experiments, as well as to allow it to be interchangeable between diff frameworks, containerization, and cloud deployment.
This is strictly a study and proof of concept repository for education purposes only.
- Visual Transformer in Pytorch and Jax
- A Vanilla visual transformer using Pytorch for understanding
- A reimplementation using jax, flax, and optmax (no huge increment in performance)
- A DETR sample for object classification