fix cuda initialize issue

Signed-off-by: YunLiu <[email protected]>

Author: KumoLiu

generative/networks/layers/vector_quantizer.py

@@ -83,7 +83,6 @@ def __init__(
             range(1, self.spatial_dims + 1)
         )
 
-    @torch.cuda.amp.autocast(enabled=False)
     def quantize(self, inputs: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
         """
         Given an input it projects it to the quantized space and returns additional tensors needed for EMA loss.
@@ -100,28 +99,28 @@ def quantize(self, inputs: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor, to
         encoding_indices_view = list(inputs.shape)
         del encoding_indices_view[1]
 
-        inputs = inputs.float()
+        with torch.cuda.amp.autocast(enabled=False):
+            inputs = inputs.float()
 
-        # Converting to channel last format
-        flat_input = inputs.permute(self.flatten_permutation).contiguous().view(-1, self.embedding_dim)
+            # Converting to channel last format
+            flat_input = inputs.permute(self.flatten_permutation).contiguous().view(-1, self.embedding_dim)
 
-        # Calculate Euclidean distances
-        distances = (
-            (flat_input**2).sum(dim=1, keepdim=True)
-            + (self.embedding.weight.t() ** 2).sum(dim=0, keepdim=True)
-            - 2 * torch.mm(flat_input, self.embedding.weight.t())
-        )
+            # Calculate Euclidean distances
+            distances = (
+                (flat_input**2).sum(dim=1, keepdim=True)
+                + (self.embedding.weight.t() ** 2).sum(dim=0, keepdim=True)
+                - 2 * torch.mm(flat_input, self.embedding.weight.t())
+            )
 
-        # Mapping distances to indexes
-        encoding_indices = torch.max(-distances, dim=1)[1]
-        encodings = torch.nn.functional.one_hot(encoding_indices, self.num_embeddings).float()
+            # Mapping distances to indexes
+            encoding_indices = torch.max(-distances, dim=1)[1]
+            encodings = torch.nn.functional.one_hot(encoding_indices, self.num_embeddings).float()
 
-        # Quantize and reshape
-        encoding_indices = encoding_indices.view(encoding_indices_view)
+            # Quantize and reshape
+            encoding_indices = encoding_indices.view(encoding_indices_view)
 
         return flat_input, encodings, encoding_indices
 
-    @torch.cuda.amp.autocast(enabled=False)
     def embed(self, embedding_indices: torch.Tensor) -> torch.Tensor:
         """
         Given encoding indices of shape [B,D,H,W,1] embeds them in the quantized space
@@ -135,7 +134,8 @@ def embed(self, embedding_indices: torch.Tensor) -> torch.Tensor:
         Returns:
             torch.Tensor: Quantize space representation of encoding_indices in channel first format.
         """
-        return self.embedding(embedding_indices).permute(self.quantization_permutation).contiguous()
+        with torch.cuda.amp.autocast(enabled=False):
+            return self.embedding(embedding_indices).permute(self.quantization_permutation).contiguous()
 
     @torch.jit.unused
     def distributed_synchronization(self, encodings_sum: torch.Tensor, dw: torch.Tensor) -> None: