Merge pull request #6465 from brkirch/fix-training

Fix training for newer PyTorch builds

modules/sd_hijack_clip.py

@@ -247,9 +247,9 @@ class FrozenCLIPEmbedderWithCustomWordsBase(torch.nn.Module):
         # restoring original mean is likely not correct, but it seems to work well to prevent artifacts that happen otherwise
         batch_multipliers = torch.asarray(batch_multipliers).to(devices.device)
         original_mean = z.mean()
-        z *= batch_multipliers.reshape(batch_multipliers.shape + (1,)).expand(z.shape)
+        z = z * batch_multipliers.reshape(batch_multipliers.shape + (1,)).expand(z.shape)
         new_mean = z.mean()
-        z *= original_mean / new_mean
+        z = z * (original_mean / new_mean)
 
         return z