fix issue with mixed precision and gradient clipping

README.md

@@ -830,6 +830,7 @@ Once built, images will be saved to the same directory the command is invoked
 - [ ] extend diffusion head to use diffusion-gan (potentially using lightweight-gan) to speed up inference
 - [ ] bring in cross-scale embedding from iclr paper https://github.com/lucidrains/vit-pytorch/blob/main/vit_pytorch/crossformer.py#L14
 - [ ] figure out if possible to augment with external memory, as described in https://arxiv.org/abs/2204.11824
+- [ ] test out grid attention in cascading ddpm locally, decide whether to keep or remove
 
 ## Citations
 

dalle2_pytorch/train.py

@@ -159,12 +159,13 @@ class DecoderTrainer(nn.Module):
         index = unet_number - 1
         unet = self.decoder.unets[index]
 
-        if exists(self.max_grad_norm):
-            nn.utils.clip_grad_norm_(unet.parameters(), self.max_grad_norm)
-
         optimizer = getattr(self, f'optim{index}')
         scaler = getattr(self, f'scaler{index}')
 
+        if exists(self.max_grad_norm):
+            scaler.unscale_(optimizer)
+            nn.utils.clip_grad_norm_(unet.parameters(), self.max_grad_norm)
+
         scaler.step(optimizer)
         scaler.update()
         optimizer.zero_grad()

dalle2_pytorch/vqgan_vae.py

@@ -285,6 +285,10 @@ class ResnetEncDec(nn.Module):
     def get_encoded_fmap_size(self, image_size):
         return image_size // (2 ** self.layers)
 
+    @property
+    def last_dec_layer(self):
+        return self.decoders[-1].weight
+
     def encode(self, x):
         for enc in self.encoders:
             x = enc(x)
@@ -419,6 +423,10 @@ class ConvNextEncDec(nn.Module):
     def get_encoded_fmap_size(self, image_size):
         return image_size // (2 ** self.layers)
 
+    @property
+    def last_dec_layer(self):
+        return self.decoders[-1].weight
+
     def encode(self, x):
         for enc in self.encoders:
             x = enc(x)
@@ -606,6 +614,10 @@ class ViTEncDec(nn.Module):
     def get_encoded_fmap_size(self, image_size):
         return image_size // self.patch_size
 
+    @property
+    def last_dec_layer(self):
+        return self.decoder[-3][-1].weight
+
     def encode(self, x):
         return self.encoder(x)
 
@@ -843,7 +855,7 @@ class VQGanVAE(nn.Module):
 
         # calculate adaptive weight
 
-        last_dec_layer = self.decoders[-1].weight
+        last_dec_layer = self.enc_dec.last_dec_layer
 
         norm_grad_wrt_gen_loss = grad_layer_wrt_loss(gen_loss, last_dec_layer).norm(p = 2)
         norm_grad_wrt_perceptual_loss = grad_layer_wrt_loss(perceptual_loss, last_dec_layer).norm(p = 2)

setup.py

@@ -10,7 +10,7 @@ setup(
       'dream = dalle2_pytorch.cli:dream'
     ],
   },
-  version = '0.0.88',
+  version = '0.0.90',
   license='MIT',
   description = 'DALL-E 2',
   author = 'Phil Wang',

train_diffusion_prior.py

@@ -17,14 +17,24 @@ os.environ["WANDB_SILENT"] = "true"
 def eval_model(model,device,image_reader,text_reader,start,end,batch_size,loss_type,phase="Validation"):
     model.eval()
     with torch.no_grad():
-        for emb_images,emb_text in zip(image_reader(batch_size=batch_size, start=start, end=end),
+        total_loss = 0.
+        total_samples = 0.
+
+        for emb_images, emb_text in zip(image_reader(batch_size=batch_size, start=start, end=end),
                 text_reader(batch_size=batch_size, start=start, end=end)):
+
             emb_images_tensor = torch.tensor(emb_images[0]).to(device)
             emb_text_tensor = torch.tensor(emb_text[0]).to(device)
+
+            batches = emb_images_tensor.shape[0]
+
             loss = model(text_embed = emb_text_tensor, image_embed = emb_images_tensor)
 
-            # Log to wandb
-            wandb.log({f'{phase} {loss_type}': loss})
+            total_loss += loss.item() * batches
+            total_samples += batches
+
+        avg_loss = (total_loss / total_samples)
+        wandb.log({f'{phase} {loss_type}': avg_loss})
 
 def save_model(save_path,state_dict):
     # Saving State Dict