[Bugfix] Set enforce_eager automatically for mllama (#12127)
Signed-off-by: Chen Zhang <zhangch99@outlook.com>
This commit is contained in:
@@ -325,7 +325,6 @@ def run_mllama(question: str, modality: str):
|
||||
model=model_name,
|
||||
max_model_len=4096,
|
||||
max_num_seqs=16,
|
||||
enforce_eager=True,
|
||||
disable_mm_preprocessor_cache=args.disable_mm_preprocessor_cache,
|
||||
)
|
||||
|
||||
|
||||
@@ -186,7 +186,6 @@ def load_mllama(question, image_urls: List[str]) -> ModelRequestData:
|
||||
model=model_name,
|
||||
max_model_len=4096,
|
||||
max_num_seqs=16,
|
||||
enforce_eager=True,
|
||||
limit_mm_per_prompt={"image": len(image_urls)},
|
||||
)
|
||||
|
||||
|
||||
Reference in New Issue
Block a user