Repository navigation

All pull requests

Fix npu fa
#425SweetStanleyL opened Apr 16, 2025
feat(moe): support fused permute when capacity is not none
#424blankde opened Apr 9, 2025Collaborator6 tasks
add tp overlap feature
#416sallyjunjun opened Feb 18, 2025Collaborator6 tasks
feat(ampipe): impl ampipe
#406blankde opened Jan 17, 2025Collaborator6 tasks
Feat/loong train mla
#400huangting4201 opened Jan 2, 2025Collaborator6 tasks
feat(checkpoint): support universal checkpoint
#394li126com opened Dec 23, 2024Collaborator9 tasks done
add use_fp32_logits flag
#393KimmiShi opened Dec 20, 2024Collaborator
feat(pipeline): add non-p2p-comm support
#385mwiacx opened Dec 4, 2024Contributor
fix(mlp): enhance mlp_layer_fusion
#382yingtongxiong opened Dec 3, 2024Collaborator1 task done
check CUDA_DEVICE_MAX_CONNECTIONS and TORCH_NCCL_AVOID_RECORD_STREAMS
#379sallyjunjun opened Dec 2, 2024Collaborator6 tasks
2
[Feat] Heterogeneous Code Part 1: Add Model and Module Code for Chameleon Lumina
#377zhhsplendid opened Nov 28, 20245 of 6 tasks
12
fix(monitor): send exception when feishu alert is enable && remove light monitoring address
#373JiaoPL opened Nov 20, 2024Collaborator6 tasks
fix llava model device bugs
#359hellozmz opened Oct 28, 20246 tasks
feat(moe): add gshard token rearrange optim
#352blankde opened Oct 21, 2024Collaborator6 tasks
feat(moe): support moe zero1 setting
#350blankde opened Oct 16, 2024Collaborator6 tasks
feat(moe): add moe async param handler
#332blankde opened Sep 13, 2024Collaborator6 tasks
1
feat(checkpoint): TP recomputation communication optimization
#275li126com opened Jul 8, 2024Collaborator6 tasks
7
Feat/add zeropp
#256chrysantd opened Jun 20, 20246 tasks
feat(simulator): support parallel cost simulator for internevo
#243SolenoidWGT opened Jun 5, 2024Contributor
add gradient sharding
#87ChenQiaoling00 opened Mar 18, 2024Contributor6 tasks

Pull requests