Skip to content

Pull requests: AI-Hypercomputer/maxtext

New pull request New
238 Open 2,245 Closed
238 Open 2,245 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

Add checkpoint conversion parameter mapping for Qwen3 Omni
#2678 opened Nov 13, 2025 by eitanporat Loading…
4 tasks done
Support GCS file pattern in grain
#2677 opened Nov 13, 2025 by aireenmei Loading…
4 tasks done
Install tpu requirements by default in pypi
#2676 opened Nov 13, 2025 by SurbhiJainUSC Loading…
4 tasks done
9
[DECOUPLED-MODE] Adding necessary files gemini-review
#2673 opened Nov 12, 2025 by gulsumgudukbay Loading…
4 tasks done
5
Add conversion script for Qwen3 Next and Readme gemini-review
#2672 opened Nov 12, 2025 by Rohan-Bierneni Loading…
4 tasks done
@Rohan-Bierneni
14
[WIP] quantization with tokamax ragged_dot
#2671 opened Nov 12, 2025 by khatwanimohit • Draft
4 tasks
Async Data Loading & Accurate Step Timing to Exclude Profiler Overhead draft Draft PR gemini-review
#2668 opened Nov 12, 2025 by NuojCheng • Draft
3 of 4 tasks
3
Add SFT documentation for Pathways
#2667 opened Nov 11, 2025 by SurbhiJainUSC Loading…
4 tasks done
[WIP] Update multimodal doc
#2663 opened Nov 11, 2025 by hengtaoguo • Draft
4 tasks done
2
Qwen3 deepstack
#2660 opened Nov 11, 2025 by eitanporat Loading…
4 tasks
Add MRoPE support for Qwen3-Omni [WIP]
#2659 opened Nov 11, 2025 by eitanporat Loading…
4 tasks
feat: migrate deepseek models to nnx
#2658 opened Nov 11, 2025 by mesakhcienet Loading…
4 tasks done
Fix deepseek tp sharding error draft Draft PR
#2657 opened Nov 11, 2025 by NuojCheng • Draft
4 tasks
Rampup batch size checkpoint support draft Draft PR gemini-review
#2656 opened Nov 11, 2025 by NuojCheng • Draft
4 tasks
2
tunix import fix pull ready
#2653 opened Nov 10, 2025 by mydatascience Loading…
4 tasks done
1
Reverts 69ed0c5d29aa25c61fd4c31a666ef35cf345d30e
#2649 opened Nov 10, 2025 by copybara-service bot Loading…
Checkpoint conversion utility: gpt-oss orbax scan to hf, many-to-one transform, inhomogeneous scan block gemini-review
#2647 opened Nov 10, 2025 by shuningjin Loading…
4 tasks done
@RissyRan @hengtaoguo
11
Add Deepseek vLLM weight fallback mappings, and correct existing deepseek mapping
#2646 opened Nov 10, 2025 by copybara-service bot Loading…
@abhinavclemson
fix: revert deepseek linen version
#2641 opened Nov 10, 2025 by mesakhcienet • Draft
4 tasks
Reverts bfdb7edb1cdb5c3d4679a034360ad744ea197790
#2639 opened Nov 8, 2025 by copybara-service bot Loading…
Fix rl and integration imports
#2637 opened Nov 8, 2025 by copybara-service bot Loading…
@gobbleturk
Internal only
#2634 opened Nov 7, 2025 by copybara-service bot Loading…
@abhinavclemson
Fixes for linter and DeepSeek
#2633 opened Nov 7, 2025 by mydatascience Loading…
4 tasks done
Updates to bring parity with the train_rl with demo scripts
#2632 opened Nov 7, 2025 by A9isha Loading…
4 tasks done
9
Fix vllm weight mapping import issue.
#2629 opened Nov 7, 2025 by abhinavclemson Loading…
4 tasks done
3
ProTip! Updated in the last three days: updated:>2025-11-10.