added weights_only=True for all torch.load calls

rvc/infer/infer.py

@@ -458,7 +458,7 @@ def load_model(self, weight_root):
             weight_root (str): Path to the model weights.
         """
         self.cpt = (
-            torch.load(weight_root, map_location="cpu")
+            torch.load(weight_root, map_location="cpu", weights_only=True)
             if os.path.isfile(weight_root)
             else None
         )

rvc/lib/predictors/FCPE.py

@@ -728,7 +728,7 @@ def __init__(self, model_path, device=None, dtype=torch.float32):
         if device is None:
             device = "cuda" if torch.cuda.is_available() else "cpu"
         self.device = device
-        ckpt = torch.load(model_path, map_location=torch.device(self.device))
+        ckpt = torch.load(model_path, map_location=torch.device(self.device), weights_only=True)
         self.args = DotDict(ckpt["config"])
         self.dtype = dtype
         model = FCPE(

rvc/lib/predictors/RMVPE.py

@@ -429,7 +429,7 @@ class RMVPE0Predictor:
     def __init__(self, model_path, device=None):
         self.resample_kernel = {}
         model = E2E(4, 1, (2, 2))
-        ckpt = torch.load(model_path, map_location="cpu")
+        ckpt = torch.load(model_path, map_location="cpu", weights_only=True)
         model.load_state_dict(ckpt)
         model.eval()
         self.model = model

rvc/train/data_utils.py

@@ -125,7 +125,7 @@ def get_audio(self, filename):
         spec_filename = filename.replace(".wav", ".spec.pt")
         if os.path.exists(spec_filename):
             try:
-                spec = torch.load(spec_filename)
+                spec = torch.load(spec_filename, weights_only=True)
             except Exception as error:
                 print(f"An error occurred getting spec from {spec_filename}: {error}")
                 spec = spectrogram_torch(

rvc/train/process/change_info.py

@@ -4,7 +4,7 @@
 
 def change_info(path, info, name):
     try:
-        ckpt = torch.load(path, map_location="cpu")
+        ckpt = torch.load(path, map_location="cpu", weights_only=True)
         ckpt["info"] = info
 
         if not name:

rvc/train/process/model_blender.py

@@ -17,8 +17,8 @@ def extract(ckpt):
 def model_blender(name, path1, path2, ratio):
     try:
         message = f"Model {path1} and {path2} are merged with alpha {ratio}."
-        ckpt1 = torch.load(path1, map_location="cpu")
-        ckpt2 = torch.load(path2, map_location="cpu")
+        ckpt1 = torch.load(path1, map_location="cpu", weights_only=True)
+        ckpt2 = torch.load(path2, map_location="cpu", weights_only=True)
 
         if ckpt1["sr"] != ckpt2["sr"]:
             return "The sample rates of the two models are not the same."

rvc/train/process/model_information.py

@@ -13,7 +13,7 @@ def prettify_date(date_str):
 
 
 def model_information(path):
-    model_data = torch.load(path, map_location="cpu")
+    model_data = torch.load(path, map_location="cpu", weights_only=True)
 
     print(f"Loaded model from {path}")
 

rvc/train/train.py

@@ -469,23 +469,23 @@ def run(
                 print(f"Loaded pretrained (G) '{pretrainG}'")
             if hasattr(net_g, "module"):
                 net_g.module.load_state_dict(
-                    torch.load(pretrainG, map_location="cpu")["model"]
+                    torch.load(pretrainG, map_location="cpu", weights_only=True)["model"]
                 )
             else:
                 net_g.load_state_dict(
-                    torch.load(pretrainG, map_location="cpu")["model"]
+                    torch.load(pretrainG, map_location="cpu", weights_only=True)["model"]
                 )
 
         if pretrainD != "" and pretrainD != "None":
             if rank == 0:
                 print(f"Loaded pretrained (D) '{pretrainD}'")
             if hasattr(net_d, "module"):
                 net_d.module.load_state_dict(
-                    torch.load(pretrainD, map_location="cpu")["model"]
+                    torch.load(pretrainD, map_location="cpu", weights_only=True)["model"]
                 )
             else:
                 net_d.load_state_dict(
-                    torch.load(pretrainD, map_location="cpu")["model"]
+                    torch.load(pretrainD, map_location="cpu", weights_only=True)["model"]
                 )
 
     # Initialize schedulers

rvc/train/utils.py

@@ -45,7 +45,7 @@ def load_checkpoint(checkpoint_path, model, optimizer=None, load_opt=1):
         checkpoint_path
     ), f"Checkpoint file not found: {checkpoint_path}"
 
-    checkpoint_dict = torch.load(checkpoint_path, map_location="cpu")
+    checkpoint_dict = torch.load(checkpoint_path, map_location="cpu", weights_only=True)
     checkpoint_dict = replace_keys_in_dict(
         replace_keys_in_dict(
             checkpoint_dict, ".weight_v", ".parametrizations.weight.original1"