[Model Runner V2] Migration to support quantized model by default [5/N] (#44446)
Signed-off-by: yewentao256 <zhyanwentao@126.com>
This commit is contained in:
@@ -188,7 +188,7 @@ def test_v2_model_runner_env_tri_state(monkeypatch, env_value, expected):
|
||||
is_moe=False,
|
||||
is_quantized=True,
|
||||
),
|
||||
False,
|
||||
True,
|
||||
),
|
||||
(
|
||||
SimpleNamespace(
|
||||
|
||||
@@ -555,9 +555,6 @@ class VllmConfig:
|
||||
if model_config.runner_type != "generate":
|
||||
return False
|
||||
|
||||
if model_config.is_quantized:
|
||||
return False
|
||||
|
||||
architectures = getattr(model_config, "architectures", [])
|
||||
return any(
|
||||
arch in DEFAULT_V2_MODEL_RUNNER_ARCHITECTURES for arch in architectures
|
||||
|
||||
Reference in New Issue
Block a user