Here are some similar smaller projects, minimizing research implementations related to machine learning/deep learning, allowing for quick learning of the underlying principles of this field: - minimalRL: minimalRL is a collection of minimal reinforcement learning (RL) algorithms implemented in PyTorch https://((((((github.com))))))/seungeunrho/minimalRL - micrograd: A purely educational tool, implementing scalar automatic differentiation (autograd) in approximately 100 lines of code, plus about 50 lines to create a mini neural network library. https://(((((((github.com))))))/karpathy/micrograd - nanoGPT: The actual working principle of GPT training https://(((((((github.com))))))/karpathy/nanoGPT - minGPT: A minimalist GPT implementation written by Andrej Karpathy, reimplementing OpenAI GPT (Generative Pretrained Transformer) training and inference in PyTorch https://(((((((github.com))))))/karpathy/minGPT - tinygrad: tinygrad is a minimalist yet complete deep learning framework led by George Hotz. https://(((((((github.com))))))/tinygrad/tinygrad - llm.c: llm.c is another minimalist project by Andrej Karpathy, implementing Large Language Model (LLM) training in pure C/CUDA, completely independent of PyTorch and Python. https://(((((((github.com))))))/karpathy/llm.c