Merge branch 'dev' into videodataset

Author: ericspod

docs/source/networks.rst

@@ -750,3 +750,38 @@ Utilities
 
 .. automodule:: monai.apps.reconstruction.networks.nets.utils
   :members:
+
+Noise Schedulers
+----------------
+.. automodule:: monai.networks.schedulers
+.. currentmodule:: monai.networks.schedulers
+
+`Scheduler`
+~~~~~~~~~~~
+.. autoclass:: Scheduler
+  :members:
+
+`NoiseSchedules`
+~~~~~~~~~~~~~~~~
+.. autoclass:: NoiseSchedules
+  :members:
+
+`DDPMScheduler`
+~~~~~~~~~~~~~~~
+.. autoclass:: DDPMScheduler
+  :members:
+
+`DDIMScheduler`
+~~~~~~~~~~~~~~~
+.. autoclass:: DDIMScheduler
+  :members:
+
+`PNDMScheduler`
+~~~~~~~~~~~~~~~
+.. autoclass:: PNDMScheduler
+  :members:
+
+`RFlowScheduler`
+~~~~~~~~~~~~~~~~
+.. autoclass:: RFlowScheduler
+  :members:

monai/apps/deepedit/interaction.py

@@ -72,7 +72,7 @@ def __call__(self, engine: SupervisedTrainer | SupervisedEvaluator, batchdata: d
 
                 with torch.no_grad():
                     if engine.amp:
-                        with torch.cuda.amp.autocast():
+                        with torch.autocast("cuda"):
                             predictions = engine.inferer(inputs, engine.network)
                     else:
                         predictions = engine.inferer(inputs, engine.network)

monai/apps/deepgrow/interaction.py

@@ -67,7 +67,7 @@ def __call__(self, engine: SupervisedTrainer | SupervisedEvaluator, batchdata: d
             engine.network.eval()
             with torch.no_grad():
                 if engine.amp:
-                    with torch.cuda.amp.autocast():
+                    with torch.autocast("cuda"):
                         predictions = engine.inferer(inputs, engine.network)
                 else:
                     predictions = engine.inferer(inputs, engine.network)

monai/apps/detection/networks/retinanet_detector.py

@@ -180,7 +180,7 @@ def forward(self, images: torch.Tensor):
                 nesterov=True,
             )
             torch.save(detector.network.state_dict(), 'model.pt')  # save model
-            detector.network.load_state_dict(torch.load('model.pt'))  # load model
+            detector.network.load_state_dict(torch.load('model.pt', weights_only=True))  # load model
     """
 
     def __init__(

monai/apps/detection/networks/retinanet_network.py

@@ -88,8 +88,8 @@ def __init__(
 
         for layer in self.conv.children():
             if isinstance(layer, conv_type):  # type: ignore
-                torch.nn.init.normal_(layer.weight, std=0.01)
-                torch.nn.init.constant_(layer.bias, 0)
+                torch.nn.init.normal_(layer.weight, std=0.01)  # type: ignore[arg-type]
+                torch.nn.init.constant_(layer.bias, 0)  # type: ignore[arg-type]
 
         self.cls_logits = conv_type(in_channels, num_anchors * num_classes, kernel_size=3, stride=1, padding=1)
         torch.nn.init.normal_(self.cls_logits.weight, std=0.01)
@@ -167,8 +167,8 @@ def __init__(self, in_channels: int, num_anchors: int, spatial_dims: int):
 
         for layer in self.conv.children():
             if isinstance(layer, conv_type):  # type: ignore
-                torch.nn.init.normal_(layer.weight, std=0.01)
-                torch.nn.init.zeros_(layer.bias)
+                torch.nn.init.normal_(layer.weight, std=0.01)  # type: ignore[arg-type]
+                torch.nn.init.zeros_(layer.bias)  # type: ignore[arg-type]
 
     def forward(self, x: list[Tensor]) -> list[Tensor]:
         """
@@ -297,7 +297,7 @@ def __init__(
             )
         self.feature_extractor = feature_extractor
 
-        self.feature_map_channels: int = self.feature_extractor.out_channels
+        self.feature_map_channels: int = self.feature_extractor.out_channels  # type: ignore[assignment]
         self.num_anchors = num_anchors
         self.classification_head = RetinaNetClassificationHead(
             self.feature_map_channels, self.num_anchors, self.num_classes, spatial_dims=self.spatial_dims

monai/apps/detection/utils/box_coder.py

@@ -221,15 +221,15 @@ def decode_single(self, rel_codes: Tensor, reference_boxes: Tensor) -> Tensor:
 
             pred_ctr_xyx_axis = dxyz_axis * whd_axis[:, None] + ctr_xyz_axis[:, None]
             pred_whd_axis = torch.exp(dwhd_axis) * whd_axis[:, None]
-            pred_whd_axis = pred_whd_axis.to(dxyz_axis.dtype)
+            pred_whd_axis = pred_whd_axis.to(dxyz_axis.dtype)  # type: ignore[union-attr]
 
             # When convert float32 to float16, Inf or Nan may occur
             if torch.isnan(pred_whd_axis).any() or torch.isinf(pred_whd_axis).any():
                 raise ValueError("pred_whd_axis is NaN or Inf.")
 
             # Distance from center to box's corner.
             c_to_c_whd_axis = (
-                torch.tensor(0.5, dtype=pred_ctr_xyx_axis.dtype, device=pred_whd_axis.device) * pred_whd_axis
+                torch.tensor(0.5, dtype=pred_ctr_xyx_axis.dtype, device=pred_whd_axis.device) * pred_whd_axis  # type: ignore[arg-type]
             )
 
             pred_boxes.append(pred_ctr_xyx_axis - c_to_c_whd_axis)

monai/apps/generation/maisi/networks/autoencoderkl_maisi.py

@@ -232,6 +232,10 @@ def forward(self, x: torch.Tensor) -> torch.Tensor:
         if self.print_info:
             logger.info(f"Number of splits: {self.num_splits}")
 
+        if self.dim_split <= 1 and self.num_splits <= 1:
+            x = self.conv(x)
+            return x
+
         # compute size of splits
         l = x.size(self.dim_split + 2)
         split_size = l // self.num_splits

monai/apps/mmars/mmars.py

@@ -241,7 +241,7 @@ def load_from_mmar(
         return torch.jit.load(_model_file, map_location=map_location)
 
     # loading with `torch.load`
-    model_dict = torch.load(_model_file, map_location=map_location)
+    model_dict = torch.load(_model_file, map_location=map_location, weights_only=True)
     if weights_only:
         return model_dict.get(model_key, model_dict)  # model_dict[model_key] or model_dict directly
 

monai/apps/reconstruction/networks/blocks/varnetblock.py

@@ -55,7 +55,7 @@ def soft_dc(self, x: Tensor, ref_kspace: Tensor, mask: Tensor) -> Tensor:
         Returns:
             Output of DC block with the same shape as x
         """
-        return torch.where(mask, x - ref_kspace, self.zeros) * self.dc_weight
+        return torch.where(mask, x - ref_kspace, self.zeros) * self.dc_weight  # type: ignore
 
     def forward(self, current_kspace: Tensor, ref_kspace: Tensor, mask: Tensor, sens_maps: Tensor) -> Tensor:
         """

monai/bundle/scripts.py

@@ -760,7 +760,7 @@ def load(
     if load_ts_module is True:
         return load_net_with_metadata(full_path, map_location=torch.device(device), more_extra_files=config_files)
     # loading with `torch.load`
-    model_dict = torch.load(full_path, map_location=torch.device(device))
+    model_dict = torch.load(full_path, map_location=torch.device(device), weights_only=True)
 
     if not isinstance(model_dict, Mapping):
         warnings.warn(f"the state dictionary from {full_path} should be a dictionary but got {type(model_dict)}.")
@@ -1279,9 +1279,8 @@ def verify_net_in_out(
         if input_dtype == torch.float16:
             # fp16 can only be executed in gpu mode
             net.to("cuda")
-            from torch.cuda.amp import autocast
 
-            with autocast():
+            with torch.autocast("cuda"):
                 output = net(test_data.cuda(), **extra_forward_args_)
             net.to(device_)
         else:
@@ -1330,7 +1329,7 @@ def _export(
         # here we use ignite Checkpoint to support nested weights and be compatible with MONAI CheckpointSaver
         Checkpoint.load_objects(to_load={key_in_ckpt: net}, checkpoint=ckpt_file)
     else:
-        ckpt = torch.load(ckpt_file)
+        ckpt = torch.load(ckpt_file, weights_only=True)
         copy_model_state(dst=net, src=ckpt if key_in_ckpt == "" else ckpt[key_in_ckpt])
 
     # Use the given converter to convert a model and save with metadata, config content