fix(pt): set weights_only=True for torch.load

deepmd/pt/entrypoints/main.py

@@ -283,7 +283,9 @@ def train(
     # update init_model or init_frz_model config if necessary
     if (init_model is not None or init_frz_model is not None) and use_pretrain_script:
         if init_model is not None:
-            init_state_dict = torch.load(init_model, map_location=DEVICE)
+            init_state_dict = torch.load(
+                init_model, map_location=DEVICE, weights_only=True
+            )
             if "model" in init_state_dict:
                 init_state_dict = init_state_dict["model"]
             config["model"] = init_state_dict["_extra_state"]["model_params"]
@@ -380,7 +382,9 @@ def change_bias(
     output: Optional[str] = None,
 ):
     if input_file.endswith(".pt"):
-        old_state_dict = torch.load(input_file, map_location=env.DEVICE)
+        old_state_dict = torch.load(
+            input_file, map_location=env.DEVICE, weights_only=True
+        )
         model_state_dict = copy.deepcopy(old_state_dict.get("model", old_state_dict))
         model_params = model_state_dict["_extra_state"]["model_params"]
     elif input_file.endswith(".pth"):

deepmd/pt/infer/deep_eval.py

@@ -103,7 +103,9 @@ def __init__(
         self.output_def = output_def
         self.model_path = model_file
         if str(self.model_path).endswith(".pt"):
-            state_dict = torch.load(model_file, map_location=env.DEVICE)
+            state_dict = torch.load(
+                model_file, map_location=env.DEVICE, weights_only=True
+            )
             if "model" in state_dict:
                 state_dict = state_dict["model"]
             self.input_param = state_dict["_extra_state"]["model_params"]

deepmd/pt/infer/inference.py

@@ -34,7 +34,7 @@ def __init__(
         - config: The Dict-like configuration with training options.
         """
         # Model
-        state_dict = torch.load(model_ckpt, map_location=DEVICE)
+        state_dict = torch.load(model_ckpt, map_location=DEVICE, weights_only=True)
         if "model" in state_dict:
             state_dict = state_dict["model"]
         model_params = state_dict["_extra_state"]["model_params"]

deepmd/pt/train/training.py

@@ -400,7 +400,9 @@ def get_lr(lr_params):
         optimizer_state_dict = None
         if resuming:
             log.info(f"Resuming from {resume_model}.")
-            state_dict = torch.load(resume_model, map_location=DEVICE)
+            state_dict = torch.load(
+                resume_model, map_location=DEVICE, weights_only=True
+            )
             if "model" in state_dict:
                 optimizer_state_dict = (
                     state_dict["optimizer"] if finetune_model is None else None

deepmd/pt/utils/finetune.py

@@ -136,7 +136,7 @@ def get_finetune_rules(
         Fine-tuning rules in a dict format, with `model_branch`: FinetuneRuleItem pairs.
     """
     multi_task = "model_dict" in model_config
-    state_dict = torch.load(finetune_model, map_location=env.DEVICE)
+    state_dict = torch.load(finetune_model, map_location=env.DEVICE, weights_only=True)
     if "model" in state_dict:
         state_dict = state_dict["model"]
     last_model_params = state_dict["_extra_state"]["model_params"]

deepmd/pt/utils/serialization.py

@@ -33,7 +33,7 @@
         model = get_model(model_def_script)
         model.load_state_dict(saved_model.state_dict())
     elif model_file.endswith(".pt"):
-        state_dict = torch.load(model_file, map_location="cpu")
+        state_dict = torch.load(model_file, map_location="cpu", weights_only=True)
         if "model" in state_dict:
             state_dict = state_dict["model"]
         model_def_script = state_dict["_extra_state"]["model_params"]

source/tests/pt/model/test_descriptor_dpa1.py

@@ -245,13 +245,15 @@ def test_descriptor_block(self):
         des = DescrptBlockSeAtten(
             **dparams,
         ).to(env.DEVICE)
-        des.load_state_dict(torch.load(self.file_model_param))
+        des.load_state_dict(torch.load(self.file_model_param, weights_only=True))
         coord = self.coord
         atype = self.atype
         box = self.cell
         # handel type_embedding
         type_embedding = TypeEmbedNet(ntypes, 8, use_tebd_bias=True).to(env.DEVICE)
-        type_embedding.load_state_dict(torch.load(self.file_type_embed))
+        type_embedding.load_state_dict(
+            torch.load(self.file_type_embed, weights_only=True)
+        )
 
         ## to save model parameters
         # torch.save(des.state_dict(), 'model_weights.pth')
@@ -299,8 +301,8 @@ def test_descriptor(self):
             **dparams,
         ).to(env.DEVICE)
         target_dict = des.state_dict()
-        source_dict = torch.load(self.file_model_param)
-        type_embd_dict = torch.load(self.file_type_embed)
+        source_dict = torch.load(self.file_model_param, weights_only=True)
+        type_embd_dict = torch.load(self.file_type_embed, weights_only=True)
         target_dict = translate_se_atten_and_type_embd_dicts_to_dpa1(
             target_dict,
             source_dict,

source/tests/pt/model/test_descriptor_dpa2.py

@@ -123,10 +123,10 @@ def test_descriptor(self):
             **dparams,
         ).to(env.DEVICE)
         target_dict = des.state_dict()
-        source_dict = torch.load(self.file_model_param)
+        source_dict = torch.load(self.file_model_param, weights_only=True)
         # type_embd of repformer is removed
         source_dict.pop("type_embedding.embedding.embedding_net.layers.0.bias")
-        type_embd_dict = torch.load(self.file_type_embed)
+        type_embd_dict = torch.load(self.file_type_embed, weights_only=True)
         target_dict = translate_type_embd_dicts_to_dpa2(
             target_dict,
             source_dict,

source/tests/pt/model/test_saveload_dpa1.py

@@ -85,7 +85,9 @@ def get_model_result(self, read=False, model_file="tmp_model.pt"):
         optimizer = torch.optim.Adam(wrapper.parameters(), lr=self.start_lr)
         optimizer.zero_grad()
         if read:
-            wrapper.load_state_dict(torch.load(model_file, map_location=env.DEVICE))
+            wrapper.load_state_dict(
+                torch.load(model_file, map_location=env.DEVICE, weights_only=True)
+            )
             os.remove(model_file)
         else:
             torch.save(wrapper.state_dict(), model_file)

source/tests/pt/model/test_saveload_se_e2_a.py

@@ -85,7 +85,9 @@ def get_model_result(self, read=False, model_file="tmp_model.pt"):
         optimizer = torch.optim.Adam(wrapper.parameters(), lr=self.start_lr)
         optimizer.zero_grad()
         if read:
-            wrapper.load_state_dict(torch.load(model_file, map_location=env.DEVICE))
+            wrapper.load_state_dict(
+                torch.load(model_file, map_location=env.DEVICE, weights_only=True)
+            )
             os.remove(model_file)
         else:
             torch.save(wrapper.state_dict(), model_file)