kingoflolz/mesh-transformer-jax

Model parallel transformers in JAX and Haiku

636 best points · 146 comments · 1 Hacker News post · 6,389 GitHub stars · View on GitHub

Hacker News discussions

• GPT-J-6B – A 6 billion parameter, autoregressive text generation model (636 points, 146 comments, by canada_dry, 2021-07-04)