Skip to content

Conversation

@Potabk
Copy link
Collaborator

@Potabk Potabk commented Aug 3, 2025

What this PR does / why we need it?

This pr fix broken CI:

  1. Fix the vllm-project/vllm@ee2eb6e changes, in this commit, they fused the gate and up projections in the vision MLP, This can improve performance by reducing one matrix multiplication. so, this pr do the following things:
    • Specify that the two linear layers are fused as mlp.gate_up_proj when loading the weights.
    • Use a SiluAndMul activation function.
  2. Fix vllm-project/vllm@aefeea0, Update ModelRunnerOutput parameters to adapt to its changes
  3. Fix vllm-commit, fix qwen moe

Does this PR introduce any user-facing change?

How was this patch tested?

@github-actions
Copy link

github-actions bot commented Aug 3, 2025

👋 Hi! Thank you for contributing to the vLLM Ascend project. The following points will speed up your PR merge:‌‌

  • A PR should do only one thing, smaller PRs enable faster reviews.
  • Every PR should include unit tests and end-to-end tests ‌to ensure it works and is not broken by other future PRs.
  • Write the commit message by fulfilling the PR description to help reviewer and future developers understand.

If CI fails, you can run linting and testing checks locally according Contributing and Testing.

@codecov
Copy link

codecov bot commented Aug 3, 2025

Codecov Report

❌ Patch coverage is 70.73171% with 12 lines in your changes missing coverage. Please review.
✅ Project coverage is 76.21%. Comparing base (ddaded1) to head (1921229).
⚠️ Report is 620 commits behind head on main.

Files with missing lines Patch % Lines
vllm_ascend/models/qwen2_5_vl.py 55.55% 4 Missing ⚠️
...ts/ut/kv_connector/test_remote_decode_lifecycle.py 77.77% 2 Missing ⚠️
...s/ut/kv_connector/test_remote_prefill_lifecycle.py 80.00% 2 Missing ⚠️
vllm_ascend/models/qwen3_moe.py 0.00% 2 Missing ⚠️
tests/ut/kv_connector/utils.py 83.33% 1 Missing ⚠️
vllm_ascend/models/qwen2_5_vl_without_padding.py 80.00% 1 Missing ⚠️
Additional details and impacted files
@@            Coverage Diff             @@
##             main    #2181      +/-   ##
==========================================
- Coverage   76.24%   76.21%   -0.03%     
==========================================
  Files         113      113              
  Lines       12492    12523      +31     
==========================================
+ Hits         9524     9545      +21     
- Misses       2968     2978      +10     
Flag Coverage Δ
unittests 76.21% <70.73%> (-0.03%) ⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@Yikun Yikun added accuracy-test enable all accuracy test for PR ready-for-test start test by label for PR labels Aug 3, 2025
Potabk added 5 commits August 4, 2025 10:09
Signed-off-by: wangli <wangli858794774@gmail.com>
Signed-off-by: wangli <wangli858794774@gmail.com>
Signed-off-by: wangli <wangli858794774@gmail.com>
Signed-off-by: wangli <wangli858794774@gmail.com>
Signed-off-by: wangli <wangli858794774@gmail.com>
Potabk added 2 commits August 4, 2025 11:41
Signed-off-by: wangli <wangli858794774@gmail.com>
Signed-off-by: wangli <wangli858794774@gmail.com>
@Potabk Potabk changed the title [Bugfix] Fix qwen2.5-vl mlp [Bugfix] Fix broken CI Aug 4, 2025
Potabk added 2 commits August 4, 2025 12:29
Signed-off-by: wangli <wangli858794774@gmail.com>
Signed-off-by: wangli <wangli858794774@gmail.com>
Signed-off-by: wangli <wangli858794774@gmail.com>
Potabk added 3 commits August 4, 2025 14:32
Signed-off-by: wangli <wangli858794774@gmail.com>
Signed-off-by: wangli <wangli858794774@gmail.com>
Signed-off-by: wangli <wangli858794774@gmail.com>
@Yikun Yikun changed the title [Bugfix] Fix broken CI [Bugfix] Follow vLLM Qwen-Moe/VL and KV Connector change to fix broken CI Aug 4, 2025
@Yikun Yikun merged commit ad366bf into vllm-project:main Aug 4, 2025
30 of 32 checks passed
@Potabk Potabk deleted the fix_vl branch August 5, 2025 01:09
zzhx1 pushed a commit to lidenghui1110/vllm-ascend that referenced this pull request Aug 11, 2025
…n CI (vllm-project#2181)

### What this PR does / why we need it?
This pr fix broken CI:
1. Fix the
vllm-project/vllm@ee2eb6e
changes, in this commit, they fused the gate and up projections in the
vision MLP, This can improve performance by reducing one matrix
multiplication. so, this pr do the following things:
- Specify that the two linear layers are fused as `mlp.gate_up_proj`
when loading the weights.
    - Use a SiluAndMul activation function.
2. Fix
vllm-project/vllm@aefeea0,
Update ModelRunnerOutput parameters to adapt to its changes
3. Fix
[vllm-commit](https://github.com/vllm-project/vllm/pull/20815/files#diff-3ffb829a39ab2b3e4706aa28f5e476815f36c3a87b98d6a66514ebedc8f3ffb4R354-R356),
fix qwen moe
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?


- vLLM version: v0.10.0
- vLLM main:
vllm-project/vllm@fed5849

---------

Signed-off-by: wangli <wangli858794774@gmail.com>
zzhx1 pushed a commit to lidenghui1110/vllm-ascend that referenced this pull request Aug 11, 2025
…n CI (vllm-project#2181)

### What this PR does / why we need it?
This pr fix broken CI:
1. Fix the
vllm-project/vllm@ee2eb6e
changes, in this commit, they fused the gate and up projections in the
vision MLP, This can improve performance by reducing one matrix
multiplication. so, this pr do the following things:
- Specify that the two linear layers are fused as `mlp.gate_up_proj`
when loading the weights.
    - Use a SiluAndMul activation function.
2. Fix
vllm-project/vllm@aefeea0,
Update ModelRunnerOutput parameters to adapt to its changes
3. Fix
[vllm-commit](https://github.com/vllm-project/vllm/pull/20815/files#diff-3ffb829a39ab2b3e4706aa28f5e476815f36c3a87b98d6a66514ebedc8f3ffb4R354-R356),
fix qwen moe
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?


- vLLM version: v0.10.0
- vLLM main:
vllm-project/vllm@fed5849

---------

Signed-off-by: wangli <wangli858794774@gmail.com>
chopper0126 pushed a commit to chopper0126/vllm-ascend that referenced this pull request Sep 26, 2025
…n CI (vllm-project#2181)

### What this PR does / why we need it?
This pr fix broken CI:
1. Fix the
vllm-project/vllm@ee2eb6e
changes, in this commit, they fused the gate and up projections in the
vision MLP, This can improve performance by reducing one matrix
multiplication. so, this pr do the following things:
- Specify that the two linear layers are fused as `mlp.gate_up_proj`
when loading the weights.
    - Use a SiluAndMul activation function.
2. Fix
vllm-project/vllm@aefeea0,
Update ModelRunnerOutput parameters to adapt to its changes
3. Fix
[vllm-commit](https://github.com/vllm-project/vllm/pull/20815/files#diff-3ffb829a39ab2b3e4706aa28f5e476815f36c3a87b98d6a66514ebedc8f3ffb4R354-R356),
fix qwen moe
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?


- vLLM version: v0.10.0
- vLLM main:
vllm-project/vllm@fed5849

---------

Signed-off-by: wangli <wangli858794774@gmail.com>
Angazenn pushed a commit to Angazenn/vllm-ascend that referenced this pull request Oct 21, 2025
…n CI (vllm-project#2181)

### What this PR does / why we need it?
This pr fix broken CI:
1. Fix the
vllm-project/vllm@ee2eb6e
changes, in this commit, they fused the gate and up projections in the
vision MLP, This can improve performance by reducing one matrix
multiplication. so, this pr do the following things:
- Specify that the two linear layers are fused as `mlp.gate_up_proj`
when loading the weights.
    - Use a SiluAndMul activation function.
2. Fix
vllm-project/vllm@aefeea0,
Update ModelRunnerOutput parameters to adapt to its changes
3. Fix
[vllm-commit](https://github.com/vllm-project/vllm/pull/20815/files#diff-3ffb829a39ab2b3e4706aa28f5e476815f36c3a87b98d6a66514ebedc8f3ffb4R354-R356),
fix qwen moe
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?


- vLLM version: v0.10.0
- vLLM main:
vllm-project/vllm@fed5849

---------

Signed-off-by: wangli <wangli858794774@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

accuracy-test enable all accuracy test for PR module:tests ready-for-test start test by label for PR vllm-break

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants