Skip to content

Pull requests: AI-Hypercomputer/maxtext

New pull request New
178 Open 2,791 Closed
178 Open 2,791 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

Remove unnecessary sharding logic for configuring Orbax single replica restore and broadcast.
#3187 opened Feb 18, 2026 by copybara-service bot Loading…
1
[ DON'T MERGE ] Experimenting with make_dynamic_splash_mha
#3186 opened Feb 18, 2026 by parambole • Draft
4 tasks
1
Support checkpointing TE quantizations with new remat policies
#3185 opened Feb 18, 2026 by hx89 Loading…
Checkpoint conversion tool: Optimize to_maxtext & Onboard deepseek2/3/3.2
#3184 opened Feb 18, 2026 by shuningjin • Draft
4 tasks
Integrate Engram into custom model gemini-review
#3183 opened Feb 18, 2026 by RissyRan Loading…
4 tasks done
@shuningjin @gagika
2
Refactor execution of deepseek sparse layers in Decoder to use pure JAX.
#3182 opened Feb 18, 2026 by copybara-service bot Loading…
1
Refactor Github Action per b/485167538
#3181 opened Feb 18, 2026 by google-admin Loading…
1
Move RL code from src/MaxText/rl/ to src/maxtext/trainers/post_train/rl/ pull ready
#3180 opened Feb 18, 2026 by A9isha Loading…
4 tasks done
3
chore: Migrate gsutil usage to gcloud storage
#3171 opened Feb 18, 2026 by gurusai-voleti Loading…
chore: Migrate gsutil usage to gcloud storage
#3170 opened Feb 18, 2026 by gurusai-voleti Loading…
Add external model bringup guide gemini-review
#3151 opened Feb 17, 2026 by RissyRan Loading…
4 tasks done
@parambole @shuningjin @Rohan-Bierneni
10
Reformat multi_token_prediction with pyink
#3150 opened Feb 17, 2026 by bvandermoon Loading…
4 tasks done
1
Migrate PlaceHolderDataIterator from tf to np gemini-review
#3149 opened Feb 17, 2026 by aireenmei Loading…
4 tasks done
3
Add a full SFT GPU demo for Llama 3_1
#3146 opened Feb 17, 2026 by katjasrz Loading…
4 tasks done
3
Mattdavidow image revert
#3142 opened Feb 14, 2026 by gobbleturk Loading…
4 tasks
1
support attn_dp_expert parallelism
#3141 opened Feb 13, 2026 by khatwanimohit Loading…
4 tasks done
2
Tool to inspect checkpoint structure gemini-review
#3139 opened Feb 13, 2026 by shuningjin Loading…
4 tasks done
@parambole @RissyRan @Rohan-Bierneni
14
#v1 Add use_load_and_broadcast option.
#3137 opened Feb 13, 2026 by copybara-service bot Loading…
1
Allow restore+broadcast logic to not require a single-replica sharding parameter, which is always constructed as a sharding over replica-local devices anyway.
#3132 opened Feb 13, 2026 by copybara-service bot Loading…
1
Disable/set tpu_only for tests required tpu backend
#3128 opened Feb 12, 2026 by NuojCheng • Draft
4 tasks
1
add cosine loss to train_distill
#3127 opened Feb 12, 2026 by entrpn Loading…
4 tasks done
1
[Draft] Engram integration
#3125 opened Feb 12, 2026 by RissyRan • Draft
1
RL: Add chips_per_vm hyperparameter and fix an issue of missing config pull ready
#3123 opened Feb 11, 2026 by charlesli640 Loading…
4 tasks done
2
DOC: add a pre-commit hook to spell-check markdown files in docs/
#3122 opened Feb 11, 2026 by ev-br Loading…
4 tasks done
Feat/migrate decoder and tests to nnx
#3114 opened Feb 9, 2026 by hsuan-lun-chiang • Draft
4 tasks done
1
ProTip! Updated in the last three days: updated:>2026-02-15.