align the ema model device back after sampling from the cascading ddpm in the decoder

The default config for clip from openai should be ViT-B/32

README.md

@@ -508,7 +508,7 @@ To use a pretrained OpenAI CLIP, simply import `OpenAIClipAdapter` and pass it i
 import torch
 from dalle2_pytorch import DALLE2, DiffusionPriorNetwork, DiffusionPrior, Unet, Decoder, OpenAIClipAdapter
 
-# openai pretrained clip - defaults to ViT/B-32
+# openai pretrained clip - defaults to ViT-B/32
 
 clip = OpenAIClipAdapter()
 

dalle2_pytorch/dalle2_pytorch.py

@@ -831,7 +831,7 @@ class DiffusionPrior(BaseGaussianDiffusion):
         image_channels = 3,
         timesteps = 1000,
         cond_drop_prob = 0.,
-        loss_type = "l1",
+        loss_type = "l2",
         predict_x_start = True,
         beta_schedule = "cosine",
         condition_on_text_encodings = True, # the paper suggests this is needed, but you can turn it off for your CLIP preprocessed text embed -> image embed training
@@ -1492,11 +1492,12 @@ class Unet(nn.Module):
 
         if self.cond_on_image_embeds:
             image_tokens = self.image_to_cond(image_embed)
+            null_image_embed = self.null_image_embed.to(image_tokens.dtype) # for some reason pytorch AMP not working
 
             image_tokens = torch.where(
                 image_keep_mask,
                 image_tokens,
-                self.null_image_embed
+                null_image_embed
             )
 
         # take care of text encodings (optional)
@@ -1520,10 +1521,12 @@ class Unet(nn.Module):
                 text_mask = rearrange(text_mask, 'b n -> b n 1')
                 text_keep_mask = text_mask & text_keep_mask
 
+            null_text_embed = self.null_text_embed.to(text_tokens.dtype) # for some reason pytorch AMP not working
+
             text_tokens = torch.where(
                 text_keep_mask,
                 text_tokens,
-                self.null_text_embed
+                null_text_embed
             )
 
         # main conditioning tokens (c)
@@ -1611,7 +1614,7 @@ class Decoder(BaseGaussianDiffusion):
         timesteps = 1000,
         image_cond_drop_prob = 0.1,
         text_cond_drop_prob = 0.5,
-        loss_type = 'l1',
+        loss_type = 'l2',
         beta_schedule = 'cosine',
         predict_x_start = False,
         predict_x_start_for_latent_diffusion = False,

dalle2_pytorch/train.py

@@ -105,6 +105,10 @@ class EMA(nn.Module):
         self.register_buffer('initted', torch.Tensor([False]))
         self.register_buffer('step', torch.tensor([0.]))
 
+    def restore_ema_model_device(self):
+        device = self.initted.device
+        self.ema_model.to(device)
+
     def update(self):
         self.step += 1
 
@@ -305,6 +309,11 @@ class DecoderTrainer(nn.Module):
 
         if self.use_ema:
             self.decoder.unets = trainable_unets             # restore original training unets
+
+        # cast the ema_model unets back to original device
+        for ema in self.ema_unets:
+            ema.restore_ema_model_device()
+
         return output
 
     def forward(

setup.py

@@ -10,7 +10,7 @@ setup(
       'dream = dalle2_pytorch.cli:dream'
     ],
   },
-  version = '0.2.9',
+  version = '0.2.12',
   license='MIT',
   description = 'DALL-E 2',
   author = 'Phil Wang',