huggingface-hub 0.21.2__tar.gz → 0.21.4__tar.gz

{huggingface_hub-0.21.2 → huggingface_hub-0.21.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: huggingface_hub
-Version: 0.21.2
+Version: 0.21.4
 Summary: Client library to download and publish models, datasets and other repos on the huggingface.co hub
 Home-page: https://github.com/huggingface/huggingface_hub
 Author: Hugging Face, Inc.

{huggingface_hub-0.21.2 → huggingface_hub-0.21.4}/src/huggingface_hub/__init__.py

@@ -46,7 +46,7 @@ import sys
 from typing import TYPE_CHECKING
 
 
-__version__ = "0.21.2"
+__version__ = "0.21.4"
 
 # Alphabetical order of definitions is ensured in tests
 # WARNING: any comment added in this dictionary definition will be lost when

{huggingface_hub-0.21.2 → huggingface_hub-0.21.4}/src/huggingface_hub/hub_mixin.py

@@ -27,8 +27,8 @@ if is_torch_available():
     import torch  # type: ignore
 
 if is_safetensors_available():
-    from safetensors import safe_open
-    from safetensors.torch import save_file
+    from safetensors.torch import load_model as load_model_as_safetensor
+    from safetensors.torch import save_model as save_model_as_safetensor
 
 
 logger = logging.get_logger(__name__)
@@ -270,6 +270,9 @@ class ModelHubMixin:
 
                 # Forward config to model initialization
                 model_kwargs["config"] = config
+            elif any(param.kind == inspect.Parameter.VAR_KEYWORD for param in init_parameters.values()):
+                # If __init__ accepts **kwargs, let's forward the config as well (as a dict)
+                model_kwargs["config"] = config
 
         instance = cls._from_pretrained(
             model_id=str(model_id),
@@ -460,7 +463,7 @@ class PyTorchModelHubMixin(ModelHubMixin):
     def _save_pretrained(self, save_directory: Path) -> None:
         """Save weights from a Pytorch model to a local directory."""
         model_to_save = self.module if hasattr(self, "module") else self  # type: ignore
-        save_file(model_to_save.state_dict(), save_directory / SAFETENSORS_SINGLE_FILE)
+        save_model_as_safetensor(model_to_save, str(save_directory / SAFETENSORS_SINGLE_FILE))
 
     @classmethod
     def _from_pretrained(
@@ -521,10 +524,15 @@ class PyTorchModelHubMixin(ModelHubMixin):
 
     @classmethod
     def _load_as_safetensor(cls, model: T, model_file: str, map_location: str, strict: bool) -> T:
-        state_dict = {}
-        with safe_open(model_file, framework="pt", device=map_location) as f:  # type: ignore [attr-defined]
-            for k in f.keys():
-                state_dict[k] = f.get_tensor(k)
-        model.load_state_dict(state_dict, strict=strict)  # type: ignore
-        model.eval()  # type: ignore
+        load_model_as_safetensor(model, model_file, strict=strict)  # type: ignore [arg-type]
+        if map_location != "cpu":
+            # TODO: remove this once https://github.com/huggingface/safetensors/pull/449 is merged.
+            logger.warning(
+                "Loading model weights on other devices than 'cpu' is not supported natively."
+                " This means that the model is loaded on 'cpu' first and then copied to the device."
+                " This leads to a slower loading time."
+                " Support for loading directly on other devices is planned to be added in future releases."
+                " See https://github.com/huggingface/huggingface_hub/pull/2086 for more details."
+            )
+            model.to(map_location)  # type: ignore [attr-defined]
         return model

{huggingface_hub-0.21.2 → huggingface_hub-0.21.4}/src/huggingface_hub.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: huggingface-hub
-Version: 0.21.2
+Version: 0.21.4
 Summary: Client library to download and publish models, datasets and other repos on the huggingface.co hub
 Home-page: https://github.com/huggingface/huggingface_hub
 Author: Hugging Face, Inc.