[BUG] Exclude .pth files when pulling remote files (#25092)

Signed-off-by: ahao-anyscale <ahao@anyscale.com>
This commit is contained in:
ahao-anyscale 2025-09-17 13:42:09 -07:00 committed by GitHub
parent 883131544f
commit f20c3b0951
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -845,7 +845,8 @@ class ModelConfig:
object_storage_model.pull_files(model, object_storage_model.pull_files(model,
ignore_pattern=[ ignore_pattern=[
"*.pt", "*.safetensors", "*.pt", "*.safetensors",
"*.bin", "*.tensors" "*.bin", "*.tensors",
"*.pth"
]) ])
self.tokenizer = object_storage_model.dir self.tokenizer = object_storage_model.dir
return return
@ -853,9 +854,12 @@ class ModelConfig:
# Only download tokenizer if needed and not already handled # Only download tokenizer if needed and not already handled
if is_runai_obj_uri(tokenizer): if is_runai_obj_uri(tokenizer):
object_storage_tokenizer = ObjectStorageModel() object_storage_tokenizer = ObjectStorageModel()
object_storage_tokenizer.pull_files( object_storage_tokenizer.pull_files(model,
model, ignore_pattern=[
ignore_pattern=["*.pt", "*.safetensors", "*.bin", "*.tensors"]) "*.pt", "*.safetensors",
"*.bin", "*.tensors",
"*.pth"
])
self.tokenizer = object_storage_tokenizer.dir self.tokenizer = object_storage_tokenizer.dir
def _get_encoder_config(self): def _get_encoder_config(self):