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

[TEST]Add initial multi modal cases of Qwen2.5-VL-32B-Instruct for nightly test (#3707)

Browse Source 
### What this PR does / why we need it?
This PR adds the initial multi modal model for nightly test, including 2
cases for Qwen2.5-vl-32b acc/perf test on A3, we need test them daily.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
by running the test

vLLM version: v0.11.0rc3
vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0

- vLLM version: v0.11.0rc3
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0

---------

Signed-off-by: wangyu31577 <wangyu31577@hundsun.com>
Co-authored-by: wangyu31577 <wangyu31577@hundsun.com>
This commit is contained in:
wangyu
2025-10-24 17:12:06 +08:00
committed by GitHub
parent 9b0baa1182
commit d301c56d1a
2 changed files with 115 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
tools/aisbench.py
View File

@@ -144,17 +144,17 @@ class AisbenchRunner:
"temperature = 0.6,\n ignore_eos = False,", content)
if self.temperature:
content = re.sub(r"temperature.*",
f"temperature = {self.temperature}", content)
f"temperature = {self.temperature},", content)
if self.top_p:
content = re.sub(r"#?top_p.*", f"top_p = {self.top_p}", content)
content = re.sub(r"#?top_p.*", f"top_p = {self.top_p},", content)
if self.top_k:
content = re.sub(r"#top_k.*", f"top_k = {self.top_k}", content)
content = re.sub(r"#top_k.*", f"top_k = {self.top_k},", content)
if self.seed:
content = re.sub(r"#seed.*", f"seed = {self.seed}", content)
content = re.sub(r"#seed.*", f"seed = {self.seed},", content)
if self.repetition_penalty:
content = re.sub(
r"#repetition_penalty.*",
f"repetition_penalty = {self.repetition_penalty}", content)
f"repetition_penalty = {self.repetition_penalty},", content)
conf_path_new = os.path.join(REQUEST_CONF_DIR,
f'{self.request_conf}_custom.py')
with open(conf_path_new, 'w', encoding='utf-8') as f: