Skip to content

WIP: feature(pu): adapt to npu device#39

Open
puyuan1996 wants to merge 29 commits intomainfrom
dev-npu
Open

WIP: feature(pu): adapt to npu device#39
puyuan1996 wants to merge 29 commits intomainfrom
dev-npu

Conversation

@puyuan1996
Copy link
Collaborator

@puyuan1996 puyuan1996 commented Feb 9, 2026

only for reference now

puyuan1996 and others added 29 commits February 9, 2026 17:39
@puyuan1996
feature(pu): adapt to npu device
f49cf58
fix(pu): fix indent type
0918623
@puyuan1996
fix(pu): fix import compatibility when no sglang
4ab9eb8
@puyuan1996
fix(pu): fix npu compatibility in strategy_base.py
ce5fd16
@puyuan1996
fix(pu): fix npu compatibility in the whole repo
877a5c2
@puyuan1996
fix(pu): fix npu backend
179b50e
@puyuan1996
fix(pu): fix torch.cuda hardcoded to adapt to npu
1b868e3
@puyuan1996
fix(pu): fix cuda hardcoded to adapt to npu
ccda92d
@puyuan1996
fix(pu): fix vllm init bug
61efe44
@puyuan1996
tmp
fac17fd
@puyuan1996
fix(pu): adapt vllm WorkerWrap to npu
bf02648
@puyuan1996
test(pu): try to fix vllm inference
3676fee
@puyuan1996
Merge branch 'dev-npu' of https://github.com/opendilab/LightRFT into …
55af18a 
…dev-npu
@puyuan1996
fix(pu): fix torch.cuda hardcoded to adapt to current_device
5f11088
@puyuan1996
test(pu): try to fix gen_time.item() runtime error
76f79ac
@puyuan1996
fix(pu): fix get_current_device
0997809
@puyuan1996
test(pu): try Synchronize before CPU transfer
65c283a
@puyuan1996
test(pu): try to fix all_reduce and all_gather in npu setting
bde52b2
@puyuan1996
test(pu): try to fix all_reduce and all_gather in npu setting
fa07849
@puyuan1996
test(pu): try to fix all_reduce and all_gather in npu setting
fefa608
@puyuan1996
test(pu): try to fix all_reduce and all_gather in npu setting
0189c9b
@puyuan1996
polish(pu): add compilation_config in vllm_ascend
6dd1a44
@puyuan1996
fix(pu): fix npu device compatibility in reward_models_utils.py
2cb6724
@puyuan1996
test(pu): add safework_t1 test
f279ca0
@puyuan1996
Merge branch 'dev-npu' of https://github.com/opendilab/LightRFT into …
0cd9540 
…dev-npu
@puyuan1996
polish(pu): adapt safework_t1 code to npu device
7cc96cc
@puyuan1996
sync code
e2982f1
@puyuan1996
fix(pu): fix rm use_engine cfg bug
a735d32
@puyuan1996
fix(pu): fix get_tokenizer_processor_vl args
8a96ab2
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@puyuan1996