EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[Misc] Remove some parts of metrics patch (#603)

Browse Source 
### What this PR does / why we need it?
Remove some parts of metrics patch, since the `cuda` hard code has been
fixed by https://github.com/vllm-project/vllm/pull/14411.

Signed-off-by: shen-shanshan <467638484@qq.com>
This commit is contained in:
Shanshan Shen
2025-04-22 18:45:21 +08:00
committed by GitHub
parent cf6ab42ee2
commit 4a0ce3660e
4 changed files with 69 additions and 40 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
vllm_ascend/patch/__init__.py
View File

@@ -90,7 +90,7 @@
#
# * Worker Patch:
# ===============
# ** File: worker/patch_common/patch_metrics.py **
# ** File: worker/patch_0_8_4/patch_metrics.py **
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# 1. `vllm.spec_decode.metrics.AsyncMetricsCollector.init_tensors` and
# `vllm.spec_decode.metrics.AsyncMetricsCollector._copy_rejsample_metrics_async`
@@ -104,7 +104,9 @@
# Future Plan:
# Revert it when the related pr is merged in vllm.
#
# 2. `vllm.spec_decode.metrics.AsyncMetricsCollector.maybe_collect_rejsample_metrics`
# ** File: worker/patch_common/patch_metrics.py **
# ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
# 1. `vllm.spec_decode.metrics.AsyncMetricsCollector.maybe_collect_rejsample_metrics`
# Why:
# There are cuda hard code (current_platform.is_cuda_alike()) in
# `AsyncMetricsCollector.maybe_collect_rejsample_metrics`