Skip to content

Pull requests: AI-Hypercomputer/maxtext

New pull request New
178 Open 3,015 Closed
178 Open 3,015 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Skip Tokamax RaggedDotGroupSizes for FP8
#3420 opened Mar 14, 2026 by BirdsOfAFthr Loading…
4 tasks done
2
Add option to start test_batch in train_rl from a specific index, also add default tokenizer_path for default model
#3419 opened Mar 14, 2026 by A9isha Loading…
4 tasks done
1
log lr hyperparams; add exact match reward
#3417 opened Mar 14, 2026 by andytwigg Loading…
4 tasks done
2
add qwen3-base variants and qwen3-1.7b
#3416 opened Mar 14, 2026 by andytwigg Loading…
4 tasks done
2
Enable training strategy for Indexer gemini-review
#3415 opened Mar 14, 2026 by RissyRan Loading…
4 tasks done
@shuningjin @Rohan-Bierneni
7
Add MoE and MLA remat policies
#3414 opened Mar 13, 2026 by abhinavgoel95 Loading…
1
adding new inference documentation.
#3413 opened Mar 13, 2026 by NicoGrande Loading…
4 tasks done
19
Remove additional weight prefetching all gathers
#3412 opened Mar 13, 2026 by NuojCheng • Draft
4 tasks
1
Shuwen check vma
#3411 opened Mar 13, 2026 by Shuwen-Fang Loading…
4 tasks
Migrate train_tokenizer off TF
#3410 opened Mar 13, 2026 by aireenmei Loading…
4 tasks done
1
Adding single-host integration test for GRPO and GSPO.
#3409 opened Mar 13, 2026 by NicoGrande Loading…
4 tasks done
11
[TEST] DS v3.2 optimization from agents
#3408 opened Mar 13, 2026 by RissyRan • Draft
4 tasks
1
Introduce a batch-split config for DSv3
#3407 opened Mar 13, 2026 by copybara-service bot Loading…
@jesselu-google
1
Make weight all-gathers explicit for DSv3 batch-split
#3406 opened Mar 13, 2026 by copybara-service bot Loading…
@jesselu-google
1
Explicitly pass qwix config for deepseek batch split pull ready
#3405 opened Mar 13, 2026 by shuningjin • Draft
4 tasks done
3
[DECOUPLED-MODE] Add ICI parallelism decoupled logic to new tests
#3404 opened Mar 12, 2026 by gulsumgudukbay Loading…
4 tasks done
@gulsumgudukbay
1
Update generated requirements for JAX 0.9.0
#3403 opened Mar 12, 2026 by SurbhiJainUSC • Draft
4 tasks done
1
add unit tests for trail_rl's get_dataset()
#3401 opened Mar 12, 2026 by A9isha Loading…
4 tasks done
1
Simplify checkpoint storage flags for Pathways workloads pull ready
#3400 opened Mar 12, 2026 by igorts-git Loading…
4 tasks done
3
Refactor rl_train and add unit test for prompt filtering
#3399 opened Mar 12, 2026 by igorts-git Loading…
4 tasks done
1
Add post training tests to CI
#3394 opened Mar 12, 2026 by dipannita08 Loading…
4 tasks done
10
/s/representative_value/representative_value_or_total_size.
#3389 opened Mar 11, 2026 by copybara-service bot Loading…
1
remove variable names
#3384 opened Mar 11, 2026 by copybara-service bot Loading…
1
Implementation for soft offline distillation using saved top-k teacher logits
#3382 opened Mar 11, 2026 by ajkv-google Loading…
4 tasks done
14
feat: convert MaxText-trained LoRA adapters to HF format
#3376 opened Mar 11, 2026 by chiajunglien • Draft
4 tasks
ProTip! Updated in the last three days: updated:>2026-03-12.