Popular repositories Loading

1. hlb-CIFAR10 hlb-CIFAR10 Public

   Train to 94% on CIFAR-10 in <6.3 seconds on a single A100. Or ~95.79% in ~110 seconds (or less!)

   Python 1.3k 81

2. hlb-gpt hlb-gpt Public

   Minimalistic, extremely fast, and hackable researcher's toolbench for GPT models in 307 lines of code. Reaches <3.8 validation loss on wikitext-103 on a single A100 in <100 seconds. Scales to large…

   Python 360 27

3. fileshare fileshare Public

   A misc space to hold (smallish) files to share with others online

   1

4. modded-nanogpt modded-nanogpt Public

   Forked from KellerJordan/modded-nanogpt

   NanoGPT (124M) in 5 minutes

   Python 1

5. HeavyBall HeavyBall Public

   Forked from HomebrewML/HeavyBall

   Efficient optimizers

   Python