autogluon.multimodal 1.1.2b20240927__tar.gz → 1.1.2b20241002__tar.gz

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: autogluon.multimodal
-Version: 1.1.2b20240927
+Version: 1.1.2b20241002
 Summary: Fast and Accurate ML in 3 Lines of Code
 Home-page: https://github.com/autogluon/autogluon
 Author: AutoGluon Community

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon/multimodal/learners/base.py

@@ -1477,7 +1477,7 @@ class BaseLearner(ExportMixin, DistillationMixin, RealtimeMixin):
                     "state_dict": {"model." + name: param for name, param in self._model.state_dict().items()}
                 }
 
-        torch.save(checkpoint, os.path.join(save_path, MODEL_CHECKPOINT))
+        torch.save(checkpoint, os.path.join(save_path, MODEL_CHECKPOINT))  # nosec B614
 
         if clean_ckpts:
             # clean old checkpoints + the intermediate files stored
@@ -2112,9 +2112,9 @@ class BaseLearner(ExportMixin, DistillationMixin, RealtimeMixin):
 
                 convert_zero_checkpoint_to_fp32_state_dict(path + "-dir", path)
                 shutil.rmtree(path + "-dir")
-                state_dict = torch.load(path, map_location=torch.device("cpu"))["state_dict"]
+                state_dict = torch.load(path, map_location=torch.device("cpu"))["state_dict"]  # nosec B614
             else:
-                state_dict = torch.load(path, map_location=torch.device("cpu"))["state_dict"]
+                state_dict = torch.load(path, map_location=torch.device("cpu"))["state_dict"]  # nosec B614
         state_dict = {k.partition(prefix)[2]: v for k, v in state_dict.items() if k.startswith(prefix)}
 
         # Some buffers like `position_ids` are registered as persistent=False since transformers 4.31.0
@@ -2222,7 +2222,7 @@ class BaseLearner(ExportMixin, DistillationMixin, RealtimeMixin):
 
         if save_model:
             checkpoint = {"state_dict": {"model." + name: param for name, param in model.state_dict().items()}}
-            torch.save(checkpoint, os.path.join(os.path.abspath(path), MODEL_CHECKPOINT))
+            torch.save(checkpoint, os.path.join(os.path.abspath(path), MODEL_CHECKPOINT))  # nosec B614
 
     @staticmethod
     def _load_metadata(

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon/multimodal/learners/matching.py

@@ -1057,7 +1057,7 @@ class MultiModalMatcher(BaseLearner):
         )
 
         checkpoint = {"state_dict": task.state_dict()}
-        torch.save(checkpoint, os.path.join(save_path, MODEL_CHECKPOINT))
+        torch.save(checkpoint, os.path.join(save_path, MODEL_CHECKPOINT))  # nosec B614
 
         if clean_ckpts:
             # clean old checkpoints + the intermediate files stored
@@ -1841,7 +1841,7 @@ class MultiModalMatcher(BaseLearner):
         response_prefix: str = "response_model.",
     ):
         if state_dict is None:
-            state_dict = torch.load(path, map_location=torch.device("cpu"))["state_dict"]
+            state_dict = torch.load(path, map_location=torch.device("cpu"))["state_dict"]  # nosec B614
         query_state_dict = {
             k.partition(query_prefix)[2]: v for k, v in state_dict.items() if k.startswith(query_prefix)
         }
@@ -1986,7 +1986,7 @@ class MultiModalMatcher(BaseLearner):
                 response_model=self._response_model,
             )
             checkpoint = {"state_dict": task.state_dict()}
-            torch.save(checkpoint, os.path.join(path, MODEL_CHECKPOINT))
+            torch.save(checkpoint, os.path.join(path, MODEL_CHECKPOINT))  # nosec B614
 
     @staticmethod
     def _load_metadata(

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon/multimodal/models/ft_transformer.py

@@ -609,12 +609,12 @@ class FT_Transformer(nn.Module):
 
         if pretrained and checkpoint_name:
             if os.path.exists(checkpoint_name):
-                ckpt = torch.load(checkpoint_name)
+                ckpt = torch.load(checkpoint_name)  # nosec B614
             else:
                 with tempfile.TemporaryDirectory() as tmpdirname:
                     checkpoint_path = os.path.join(tmpdirname, "./ft_transformer_pretrained.ckpt")
                     download(checkpoint_name, checkpoint_path)
-                    ckpt = torch.load(checkpoint_path)
+                    ckpt = torch.load(checkpoint_path)  # nosec B614
             self.transformer.load_state_dict(ckpt["state_dict"])
 
         self.name_to_id = self.get_layer_ids()

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon/multimodal/models/mmdet_image.py

@@ -169,7 +169,7 @@ class MMDetAutoModelForObjectDetection(nn.Module):
         if not save_path:
             save_path = f"./{self.checkpoint_name}_autogluon.pth"
 
-        torch.save({"state_dict": self.model.state_dict(), "meta": {"CLASSES": self.model.CLASSES}}, save_path)
+        torch.save({"state_dict": self.model.state_dict(), "meta": {"CLASSES": self.model.CLASSES}}, save_path)  # nosec B614
 
     def _save_configs(self, save_path=None):
         if not save_path:
@@ -500,7 +500,7 @@ class MMDetAutoModelForObjectDetection(nn.Module):
         """
         sd = source_path
 
-        model_dict = torch.load(sd, map_location=torch.device("cpu"))
+        model_dict = torch.load(sd, map_location=torch.device("cpu"))  # nosec B614
         if "state_dict" in model_dict:
             model_dict = model_dict["state_dict"]
         if "model" in model_dict:
@@ -617,6 +617,6 @@ class MMDetAutoModelForObjectDetection(nn.Module):
         data = {"state_dict": new_dict}
 
         target_directory = os.path.splitext(sd)[0] + f"_cvt.pth"
-        torch.save(data, target_directory)
+        torch.save(data, target_directory)  # nosec B614
 
         return target_directory

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon/multimodal/models/timm_image.py

@@ -294,7 +294,7 @@ class TimmAutoModelForImagePrediction(nn.Module):
 
     def save(self, save_path: str = "./", tokenizers: Optional[dict] = None):
         weights_path = f"{save_path}/pytorch_model.bin"
-        torch.save(self.model.state_dict(), weights_path)
+        torch.save(self.model.state_dict(), weights_path)  # nosec B614
         logger.info(f"Model {self.prefix} weights saved to {weights_path}.")
         config_path = f"{save_path}/config.json"
         self.dump_config(config_path)

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon/multimodal/utils/cache.py

@@ -92,10 +92,10 @@ class DDPPredictionWriter(BasePredictionWriter):
         """
         # this will create N (num processes) files in `cache_dir` each containing
         # the predictions of its respective rank
-        torch.save(predictions, self.get_predictions_cache_dir(trainer.global_rank))
+        torch.save(predictions, self.get_predictions_cache_dir(trainer.global_rank))  # nosec B614
         # here we save `batch_indices` to get the information about the data index
         # from prediction data
-        torch.save(batch_indices, self.get_batch_indices_cache_dir(trainer.global_rank))
+        torch.save(batch_indices, self.get_batch_indices_cache_dir(trainer.global_rank))  # nosec B614
 
     def read_single_gpu_results(self, global_rank: Optional[int]):
         """
@@ -109,8 +109,8 @@ class DDPPredictionWriter(BasePredictionWriter):
         while (not os.path.exists(sample_indices_file)) or (not os.path.exists(predictions_file)):
             logger.info(f"waiting for rank #{global_rank} to finish saving predictions...")
             time.sleep(self.sleep_time)
-        sample_indices = torch.load(sample_indices_file)
-        predictions = torch.load(predictions_file)
+        sample_indices = torch.load(sample_indices_file)  # nosec B614
+        predictions = torch.load(predictions_file)  # nosec B614
 
         return sample_indices, predictions
 

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon/multimodal/utils/checkpoint.py

@@ -40,9 +40,9 @@ def average_checkpoints(
 
                 convert_zero_checkpoint_to_fp32_state_dict(per_path + "-dir", per_path)
                 shutil.rmtree(per_path + "-dir")
-                state_dict = torch.load(per_path, map_location=torch.device("cpu"))["state_dict"]
+                state_dict = torch.load(per_path, map_location=torch.device("cpu"))["state_dict"]  # nosec B614
             else:
-                state_dict = torch.load(per_path, map_location=torch.device("cpu"))["state_dict"]
+                state_dict = torch.load(per_path, map_location=torch.device("cpu"))["state_dict"]  # nosec B614
             for k, v in state_dict.items():
                 if k not in avg_state_dict:
                     avg_state_dict[k] = v.clone().to(dtype=torch.float64)
@@ -60,7 +60,7 @@ def average_checkpoints(
         for k in avg_state_dict:
             avg_state_dict[k].clamp_(float32_info.min, float32_info.max).to(dtype=torch.float32)
     else:
-        avg_state_dict = torch.load(checkpoint_paths[0], map_location=torch.device("cpu"))["state_dict"]
+        avg_state_dict = torch.load(checkpoint_paths[0], map_location=torch.device("cpu"))["state_dict"]  # nosec B614
 
     return avg_state_dict
 

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon/multimodal/utils/cloud_io.py

@@ -48,7 +48,7 @@ def _load(
     """
     if not isinstance(path_or_url, (str, Path)):
         # any sort of BytesIO or similar
-        return torch.load(path_or_url, map_location=map_location)
+        return torch.load(path_or_url, map_location=map_location)  # nosec B614
     if str(path_or_url).startswith("http"):
         return torch.hub.load_state_dict_from_url(
             str(path_or_url),
@@ -56,7 +56,7 @@ def _load(
         )
     fs = get_filesystem(path_or_url)
     with fs.open(path_or_url, "rb") as f:
-        return torch.load(f, map_location=map_location)
+        return torch.load(f, map_location=map_location)  # nosec B614
 
 
 def get_filesystem(path: _PATH, **kwargs: Any) -> AbstractFileSystem:
@@ -75,6 +75,6 @@ def _atomic_save(checkpoint: Dict[str, Any], filepath: Union[str, Path]) -> None
             This points to the file that the checkpoint will be stored in.
     """
     bytesbuffer = io.BytesIO()
-    torch.save(checkpoint, bytesbuffer)
+    torch.save(checkpoint, bytesbuffer)  # nosec B614
     with fsspec.open(filepath, "wb") as f:
         f.write(bytesbuffer.getvalue())

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon/multimodal/version.py

@@ -1,3 +1,3 @@
 """This is the autogluon version file."""
-__version__ = '1.1.2b20240927'
+__version__ = '1.1.2b20241002'
 __lite__ = False

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon.multimodal.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: autogluon.multimodal
-Version: 1.1.2b20240927
+Version: 1.1.2b20241002
 Summary: Fast and Accurate ML in 3 Lines of Code
 Home-page: https://github.com/autogluon/autogluon
 Author: AutoGluon Community

{autogluon.multimodal-1.1.2b20240927 → autogluon.multimodal-1.1.2b20241002}/src/autogluon.multimodal.egg-info/requires.txt

@@ -20,9 +20,9 @@ text-unidecode<1.4,>=1.3
 torchmetrics<1.3.0,>=1.2.0
 nptyping<2.5.0,>=1.4.4
 omegaconf<2.3.0,>=2.1.1
-autogluon.core[raytune]==1.1.2b20240927
-autogluon.features==1.1.2b20240927
-autogluon.common==1.1.2b20240927
+autogluon.core[raytune]==1.1.2b20241002
+autogluon.features==1.1.2b20241002
+autogluon.common==1.1.2b20241002
 pytorch-metric-learning<2.4,>=1.3.0
 nlpaug<1.2.0,>=1.1.10
 nltk<4.0.0,>=3.4.5