-
Notifications
You must be signed in to change notification settings - Fork 540
Pull requests: AI-Hypercomputer/maxtext
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Use resolved mesh size for context parallel sharding
#4211
opened Jun 21, 2026 by
huytransformer
Collaborator
Loading…
4 tasks done
Fix compile_cache_test to assert single jit_train_step cache file
#4210
opened Jun 20, 2026 by
Liauuu
Loading…
Fix explicit-mesh sharding assert in deepseek batch-split scan
#4208
opened Jun 19, 2026 by
ecnal-cienet
Collaborator
•
Draft
4 tasks done
Fix logical sharding resolution in NNX
#4205
opened Jun 19, 2026 by
xibinliu
Collaborator
Loading…
4 tasks done
Adding myself to codeowners to unblock my team on PRs.
#4201
opened Jun 18, 2026 by
entrpn
Collaborator
Loading…
4 tasks done
feat(layers): implement custom shape-aligned attention and MoE primit…
#4200
opened Jun 18, 2026 by
katyaoussar
Loading…
[ROCm]: test: update HLO references after tmem optimizations (PR4)
#4194
opened Jun 17, 2026 by
cj401-amd
Collaborator
Loading…
1 task
[ROCm]: fix: reduce MoE temp memory — embedding cap, weight sum default, skip trivial specs (PR3)
#4193
opened Jun 17, 2026 by
cj401-amd
Collaborator
Loading…
2 tasks
[ROCm]: fix: reduce pipeline temp memory — replace ppermute collectives with lax.slice/pad (PR2)
#4192
opened Jun 17, 2026 by
cj401-amd
Collaborator
Loading…
3 tasks
[ROCm]: fix: JAX/TE sharding compatibility and tmem reduction foundations (PR1)
#4191
opened Jun 17, 2026 by
cj401-amd
Collaborator
Loading…
2 tasks
Dsv4 load balancing
gemini-review
#4190
opened Jun 17, 2026 by
dipakg-lang
Collaborator
Loading…
4 tasks
Add ragged sort kernel fallback mechanism and version guard
#4187
opened Jun 17, 2026 by
NuojCheng
Collaborator
Loading…
3 of 4 tasks
[WIP] Add E2E test scripts for qwen3-30b model
#4185
opened Jun 17, 2026 by
YixuanWang-99
Collaborator
Loading…
4 tasks
Fix post-training Docker build for new vllm commit
#4184
opened Jun 17, 2026 by
khatwanimohit
Collaborator
Loading…
4 tasks done
Make MoE dispatch/MLP expert-axis batch sharding configurable (fix Mixtral EP throughput)
gemini-review
#4179
opened Jun 16, 2026 by
gulsumgudukbay
Collaborator
Loading…
4 tasks done
Load balancing changes for Deepseek v4
#4178
opened Jun 16, 2026 by
dipakg-lang
Collaborator
Loading…
4 tasks
[Deepseek V4] Add caching support and verify decoding
#4176
opened Jun 16, 2026 by
Rohan-Bierneni
Collaborator
Loading…
4 tasks done
[RL] Fix GPT-OSS 20B dimension mismatch error in vLLM adapter by resolving intermediate_size fallback
#4175
opened Jun 16, 2026 by
susanbao
Collaborator
Loading…
2 of 4 tasks
Add layer by layer hidden state testing support to forward_pass_logit_checker.py
#4173
opened Jun 16, 2026 by
snehalv2002
Collaborator
•
Draft
4 tasks
Introduce SubBatchCheckpointManager interface.
#4171
opened Jun 15, 2026 by
copybara-service
Bot
Loading…
Refactor moe.p: gmm and a2a unsort
#4170
opened Jun 15, 2026 by
Shuwen-Fang
Collaborator
Loading…
4 tasks done
Add support for
keep_every_nth_step in checkpointing options.
#4169
opened Jun 15, 2026 by
copybara-service
Bot
Loading…
Previous Next
ProTip!
no:milestone will show everything without a milestone.