[Fix] Fix device mismatch for ascend npu when loss_cfg.loss_reduction != "token"#1365
Open
nil0x9 wants to merge 1 commit intoInternLM:mainfrom
Open
[Fix] Fix device mismatch for ascend npu when loss_cfg.loss_reduction != "token"#1365nil0x9 wants to merge 1 commit intoInternLM:mainfrom
nil0x9 wants to merge 1 commit intoInternLM:mainfrom
Conversation
nil0x9
force-pushed
the
linty/fix-npu-loss-weight-device-mismatch
branch
from
8613e2d to
f1294d3
Compare
nil0x9
force-pushed
the
linty/fix-npu-loss-weight-device-mismatch
branch
from
f1294d3 to
dacb7ef
Compare
HAOCHENYE
approved these changes
Dec 24, 2025
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Currently when one passes
loss_cfg.loss_reductionother than "token" on a ascend/npu device, a Runtime Error (device mismatch) is expected in this line:xtuner/xtuner/v1/loss/ce_loss.py
Line 191 in d339718
The root cause of this error is that, in ascend npu device,
cu_seq_lenstensors are required to be oncpu. In funcbuild_batches_loss_kwargs, the devuce ofloss_weightis inherited fromnum_grad_tokens -> boundaries -> cu_seq_lens-- and hence the problem.