🔬 The most atomic GPT-2 implementation in 265 lines of pure Python & CUDA. A bilingual "Rosetta Stone" for understanding LLM internals from scratch. No dependencies, just math and kernels.

Profine automatically profiles and optimizes PyTorch training jobs on real GPUs, delivering measurable speedups and lower GPU costs before teams waste days tuning configs by hand.

A TF re-implementation of the Karpathy's minGPT (Generative Pretrained Transformer) training

Teaching Transformer Models facts about the world throughpretraining, and accessing that knowledge through finetuning.

A minimal PyTorch re-implementation of the OpenAI GPT (Generative Pretrained Transformer) training

A python package to experiment with GPT-like transformer models

This repository chronicles my journey through fundamental and advanced deep learning concepts. Each project is a battle-tested module combining rigorous theory with practical implementation.