google/paxml

View on GitHub

Pax is a Jax-based machine learning framework for training large scale models. Pax allows for advanced and fully configurable experimentation and parallelization, and has demonstrated industry leading model flop utilization rates.

Stars: 560
Forks: 72
Open beginner issues: 0
Indexed issues: 17

Dominant language: Python
License: Apache-2.0
Last GitHub push: Sep 9, 2026
Latest indexed: Sep 16, 2026
Contributing guide: Contributing guide
Code of conduct: Code of conduct
Beginner labels: No beginner labels indexed
PR merge metrics: No merged PRs in 30d

Open All

17 open issues indexed

Unexpected Overheads with Activation Checkpointing with Pipeline Parallelism Open

google/paxml#17 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
ARM64 Build Open

google/paxml#30 · 4 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Int8 checkpoint Open

google/paxml#35 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
How to continue training from a checkpoint? Open

google/paxml#37 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Remaining tutorials Open

google/paxml#58 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[Feature Request] Need ZeRo-1/2 to cooperate with PP+TP+DP. Which may more faster than FSDP sometimes. Open

google/paxml#64 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
[Question] Very low MFU(30%~35%) when train bf16 Llama2 and GPT model with single SXM4 A100 machine. Open

google/paxml#65 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Use bfloat16 for eval Open

google/paxml#66 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Jax + tpu and AQT int8 train model loss is abnormal Open

google/paxml#71 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
DEADLINE_EXCEEDED on 1024 GPUs. Open

google/paxml#77 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
lingvo issue while installing paxml in vscode Open

google/paxml#89 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Error running tutorial notebook Open

google/paxml#95 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
LoRA support in transformers Open

google/paxml#96 · 1 comment · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Question about Integration Timeline for PR #99 Open

google/paxml#102 · 8 comments · 9 reactions · 0 assignees · Indexed Sep 16, 2026
How to do gradient accumulation? Open

google/paxml#104 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Hard dependencies causing conflict with other packages Open

google/paxml#105 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026
Remove tensorflow dependencies Open

google/paxml#106 · 0 comments · 0 reactions · 0 assignees · Indexed Sep 16, 2026

google/paxml

Get new issues in your inbox