biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[CI] Add ci_envs for convenient local testing (#24630)

Browse Source 
Signed-off-by: wang.yuqi <noooop@126.com>
This commit is contained in:
wang.yuqi
2025-09-12 16:52:25 +08:00
committed by GitHub
parent 561a0baee0
commit d21a36f5f9
4 changed files with 98 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
vllm/config/__init__.py
View File

@@ -1775,16 +1775,21 @@ class ModelConfig:
such as the lm_head in a generation model,
or the score or classifier in a classification model.
The default head_dtype based on runner_type.\n
`head_dtype` currently only supports pooling models.\n
- The pooling model defaults to using fp32 head,
you can use --hf-overrides '{"head_dtype": "model"}' to disable it.\n
- The generate model defaults to not using fp32 head,
you can use --hf-overrides '{"head_dtype": "float32"}' to enable it.
you can use --hf-overrides '{"head_dtype": "model"}' to disable it.
"""
head_dtype = _get_head_dtype(config=self.hf_config,
dtype=self.dtype,
runner_type=self.runner_type)
if self.runner_type != "pooling" and head_dtype != self.dtype:
logger.warning_once(
"`head_dtype` currently only supports pooling models."
"fallback to model dtype [%s].", self.dtype)
return self.dtype
if head_dtype not in current_platform.supported_dtypes:
logger.warning_once(
"The current platform does not support [%s] head dtype, "