A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 05:30 UTC
[BUG] Exclude .pth files when pulling remote files (#25092)
Signed-off-by: ahao-anyscale <ahao@anyscale.com>
ahao-anyscale committed
Sep 17, 2025 at 13:42 UTC
f20c3b095109fd2a016e201d550bcaae6414e9fc
1 file changed
+8
-4
vllm/config/__init__.py
+8
-4
@@ -845,7 +845,8 @@ class ModelConfig:
845
object_storage_model.pull_files(model,
846
ignore_pattern=[
847
"*.pt", "*.safetensors",
848
- "*.bin", "*.tensors"
848
+ "*.bin", "*.tensors",
849
+ "*.pth"
850
])
851
self.tokenizer = object_storage_model.dir
852
return
@@ -853,9 +854,12 @@ class ModelConfig:
854
# Only download tokenizer if needed and not already handled
855
if is_runai_obj_uri(tokenizer):
856
object_storage_tokenizer = ObjectStorageModel()
856
- object_storage_tokenizer.pull_files(
857
- model,
858
- ignore_pattern=["*.pt", "*.safetensors", "*.bin", "*.tensors"])
857
+ object_storage_tokenizer.pull_files(model,
858
+ ignore_pattern=[
859
+ "*.pt", "*.safetensors",
860
+ "*.bin", "*.tensors",
861
+ "*.pth"
862
+ ])
863
self.tokenizer = object_storage_tokenizer.dir
864
865
def _get_encoder_config(self):