Dynamic Expert Load Balance with Zero-like-overhead #12241
Triggered via pull request
September 12, 2025 11:36
offline893
Status
Failure
Total duration
5m 46s
Artifacts
–
vllm_ascend_test.yaml
on: pull_request
Matrix: singlecard e2e test - light
Matrix: unit test
Matrix: multicard e2e test - light
Annotations
10 errors
|
lint / pre-commit:
vllm_ascend/ops/common_fused_moe.py#L290
Cannot determine type of "global_num_experts" [has-type]
|
|
lint / pre-commit:
vllm_ascend/eplb/core/eplb_worker.py#L401
Need type annotation for "block_update_q" [var-annotated]
|
|
lint / pre-commit:
vllm_ascend/eplb/core/eplb_worker.py#L400
Need type annotation for "planner_q" [var-annotated]
|
|
lint / pre-commit:
vllm_ascend/ops/common_fused_moe.py#L291
Cannot determine type of "local_num_experts" [has-type]
|
|
lint / pre-commit:
vllm_ascend/ops/common_fused_moe.py#L290
Cannot determine type of "global_num_experts" [has-type]
|
|
lint / pre-commit:
vllm_ascend/eplb/core/eplb_worker.py#L401
Need type annotation for "block_update_q" [var-annotated]
|
|
lint / pre-commit:
vllm_ascend/eplb/core/eplb_worker.py#L400
Need type annotation for "planner_q" [var-annotated]
|
|
lint / pre-commit:
vllm_ascend/ops/common_fused_moe.py#L291
Cannot determine type of "local_num_experts" [has-type]
|
|
lint / pre-commit:
vllm_ascend/ops/common_fused_moe.py#L290
Cannot determine type of "global_num_experts" [has-type]
|
|
Ruff (F841):
vllm_ascend/eplb/adaptor/vllm_adaptor.py#L210
vllm_ascend/eplb/adaptor/vllm_adaptor.py:210:13: F841 Local variable `expert_tensor` is assigned to but never used
|