Concluding my quest into reinforcement learning, building a RL game env and doing low-level GPU optimizations.
Publishing my RL Game Env as OSS
Posts
Scaling RL: GPU multithreaded batching with Puffer
A series of macro and micro optimizations to make Puffer RL training faster and scale better.
Pixel Platformer Demo: Explore RL and Self-Play
Using a toy but playable pixel platformer to create agents that can play along-side you.
Deconstructing a 1-bit LSTM
(Part 2/2) Understanding the internals of an LSTM by (de)constructing an 1-bit LSTM.
Draw with an LSTM
(Part 1/2) A creative coding experiment to explore LSTMs with your touch
