[Bugfix] Fix test fused quant layernorm tests #27865
Merged
Conversation
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
Signed-off-by: ElizaWszola <[email protected]>
Signed-off-by: yewentao256 <[email protected]>
Signed-off-by: yewentao256 <[email protected]>
yewentao256
requested review from
WoosukKwon,
mgoin,
tlrmchlsmth and
yewentao256
as code owners
yewentao256
reviewed
Nov 3, 2025
Member
yewentao256
left a comment
yewentao256
left a comment
There was a problem hiding this comment.
After we fix the scale_ub case in unit test, we now face three known issues.
- precision error for int8
- precision error for fp8
- illegal memory access without setting scale_ub
And I fix all of them in my commits
reserve core]$ pytest test_fused_quant_layernorm.py -x
============= test session starts ==============
platform linux -- Python 3.12.11, pytest-8.4.2, pluggy-1.6.0
rootdir: /home/yewentao256/vllm-source
configfile: pyproject.toml
plugins: anyio-4.11.0
collected 576 items
test_fused_quant_layernorm.py .......... [ 1%]
........................................ [ 8%]
........................................ [ 15%]
........................................ [ 22%]
........................................ [ 29%]
........................................ [ 36%]
........................................ [ 43%]
........................................ [ 50%]
........................................ [ 57%]
........................................ [ 64%]
........................................ [ 71%]
........................................ [ 78%]
........................................ [ 85%]
........................................ [ 92%]
........................................ [ 98%]
...... [100%]
= 576 passed, 2 warnings in 160.03s (0:02:40) ==Signed-off-by: ElizaWszola <[email protected]>
ElizaWszola
added a commit
to neuralmagic/vllm
that referenced
this pull request
Nov 4, 2025
…or int8 Signed-off-by: ElizaWszola <[email protected]>
Signed-off-by: ElizaWszola <[email protected]>
yewentao256
approved these changes
Nov 7, 2025
mgoin
approved these changes
Nov 8, 2025
xuebwang-amd
pushed a commit
to xuebwang-amd/vllm
that referenced
this pull request
Nov 13, 2025
Signed-off-by: ElizaWszola <[email protected]> Signed-off-by: yewentao256 <[email protected]> Co-authored-by: yewentao256 <[email protected]> Co-authored-by: Wentao Ye <[email protected]> Signed-off-by: xuebwang-amd <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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.
Fix optional vs. bool type issue in
test_fused_quant_layernorm.py. This allows int8 tests to run correctly.TODO: fix unit test failures