Toy implementations of whatever papers I happen to be reading.

I try to prioritize papers where either no open-source implementations exist, or open-source implementations aren't very lean or usable.

Example

pip install -r llllvvuu/expert_iteration/requirements.txt
python -m llllvvuu.expert_iteration

Implemented

• Vanilla Monte-Carlo Tree Search (MCTS)
• Expert Iteration (ExIt), from Thinking Fast and Slow with Deep Learning and Tree Search (Anthony, Tian, Barber 2017)

Weights

https://huggingface.co/llllvvuu/ml_papers