ReaLLMASIC / nanoGPT

The simplest, fastest repository for training/finetuning medium-sized GPTs.
MIT License
24 stars 18 forks source link

Add hw efficient and learned GELU variations #281

Closed gkielian closed 3 weeks ago

klei22 commented 3 weeks ago

This PR is apparently a subset of the one just merged, so closing since these are already merged in.