Skip to content

Pull requests: AI-Hypercomputer/maxtext

New pull request New
285 Open 1,768 Closed
285 Open 1,768 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 an option to save a checkpoint on completion
#2222 opened Aug 21, 2025 by xuefgu Loading…
4 tasks done
@aireenmei @gobbleturk
Provide MaxText axes to cudnn_flash_te to correctly perform dbias reduction if required
#2221 opened Aug 21, 2025 by jberchtold-nvidia Loading…
4 tasks done
1
Update RESTRUCTURE.md to include latest changes
#2220 opened Aug 21, 2025 by bvandermoon Loading…
4 tasks done
fix grpo name pull ready
#2219 opened Aug 21, 2025 by Taishi-N324 Loading…
2 tasks done
Demo Decode Colab in MaxText
#2218 opened Aug 21, 2025 by A9isha Loading…
4 tasks done
1
Update remat rules
#2217 opened Aug 21, 2025 by RissyRan Loading…
4 tasks done
1
@gobbleturk @khatwanimohit
Add stochastic decoding sampling strategy, allowing run-time sampling parameters change.
#2215 opened Aug 20, 2025 by babyplutokurt Loading…
4 tasks done
8
Integrative RAG Agent
#2214 opened Aug 20, 2025 by wei879-100 Loading…
4 tasks done
21
Add Zero1+GA to improve performance
#2208 opened Aug 20, 2025 by hx89 Loading…
[Qwix] Correctly plumb quantization_rule to kernel
#2206 opened Aug 19, 2025 by khatwanimohit Loading…
4 tasks done
Fix [late-directive] error reported by pytype static analyzer in max_utils.py
#2205 opened Aug 19, 2025 by copybara-service bot Loading…
Disable depth scaling in query projection if qk_norm or query_pre_attn_scalar are set
#2204 opened Aug 19, 2025 by gagika Loading…
4 tasks done
@shralex @RissyRan
@gobbleturk @khatwanimohit
Support multimodal in logit checker + match gemma3 logits with HF
#2203 opened Aug 19, 2025 by aireenmei Loading…
4 tasks done
@hengtaoguo @gagika
2
feat: migrate decoders.py to nnx
#2202 opened Aug 19, 2025 by mesakhcienet • Draft
4 tasks
Migrate DotProductAttention to NNX
#2198 opened Aug 18, 2025 by hsuan-lun-chiang Loading…
4 tasks done
5
Migrate QwenDecoderLayer to NNX
#2196 opened Aug 17, 2025 by gagika • Draft
4 tasks done
Refactor MaxText's checkpointing.py to support sharded loading of SafeTensor's checkpoints.
#2192 opened Aug 15, 2025 by copybara-service bot Loading…
Add remote_bytes_transferred to Pallas CostEstimate dataclass.
#2191 opened Aug 15, 2025 by copybara-service bot Loading…
Add new flag --enable_rich_metrics
#2189 opened Aug 15, 2025 by ycchenzheng • Draft
4 tasks
XLA flag xla_gpu_graph_level has been depreciated, and should use
#2186 opened Aug 15, 2025 by copybara-service bot Loading…
XLA flag xla_gpu_graph_level has been depreciated, and should use
#2184 opened Aug 15, 2025 by copybara-service bot Loading…
Use pl.ANY/pl.MemorySpace.ANY instead of the pltpu variant in preparation for its deprecation.
#2180 opened Aug 15, 2025 by copybara-service bot Loading…
Migrate LlamaDecoderLayer to NNX
#2178 opened Aug 14, 2025 by bvandermoon Loading…
4 tasks done
15
Enhance Checkpoint Converter for MoE Models
#2177 opened Aug 14, 2025 by parambole Loading…
4 tasks done
@shralex @hengtaoguo
@shuningjin @YixuanWang-99
6
Add valid test cases list for sharding tests
#2175 opened Aug 14, 2025 by hsuan-lun-chiang Loading…
4 tasks done
6
ProTip! Adding no:label will show everything without a label.