Appgithub-com-eleutherai-gpt-neox

Download Pinokio

Sign in Create accountJoin

Sign in Create accountJoin

gpt-neox

github.com/eleutherai/gpt-neoxUpdated 12d ago

An implementation of model parallel autoregressive transformers on GPUs, based on the Megatron and DeepSpeed libraries

Loading community details…

Loading project owner…

Loading launchers…

Loading check-ins…

Community

Sort byCreate post

Ask or share about gpt-neox…

Loading...

Loading references…

Loading project activity…

Pinokio

0

Projects using this repository

No references listed yet.

Check-ins (0)

Recent commits

Remove unused `sparse_attention` argument (#1429)

Aflah2 weeks ago28c81ed

Move adoption summary to top of README and add USERS.md (#1439)

Quentin Anthony2 weeks ago66d247f

Fix gradient clipping configuration and remove `clip_grad` (#1433)

Aflah2 weeks agod8ef803

Fix cosine LR climbing back up after lr_decay_iters (#1438)

Mohammad Hijjawi2 weeks ago5fae199

Fix port use (#1426)

Aflah2 weeks agoc9202af

Cleanup (#1427)

Aflah2 weeks ago2d66d33

Add ability to log loss/PPL per validation/test set (#1392)

Aflah2 months agocf40e42

OLMo Dense Model Training Configs (#1407)

Aflah2 months ago3aa3ac0

fix bug in EvalHarnessAdapter when pipe parallel size is set to 1 (#1356)

Marisa Hudspeth2 months agoe2a2613

Align Args in .md with .py files (#1389)

Aflah2 months ago606baa2