fix everything around normalizing images to -1 to 1 for ddpm training automatically

dalle2_pytorch/dalle2_pytorch.py

@@ -114,10 +114,10 @@ def resize_image_to(image, target_image_size):
 # ddpms expect images to be in the range of -1 to 1
 # but CLIP may otherwise
 
-def normalize_img(img):
+def normalize_neg_one_to_one(img):
     return img * 2 - 1
 
-def unnormalize_img(normed_img):
+def unnormalize_zero_to_one(normed_img):
     return (normed_img + 1) * 0.5
 
 # clip related adapters
@@ -278,7 +278,7 @@ class OpenAIClipAdapter(BaseClipAdapter):
     def embed_image(self, image):
         assert not self.cleared
         image = resize_image_to(image, self.image_size)
-        image = self.clip_normalize(unnormalize_img(image))
+        image = self.clip_normalize(image)
         image_embed = self.clip.encode_image(image)
         return EmbeddedImage(l2norm(image_embed.float()), None)
 
@@ -1821,7 +1821,7 @@ class Decoder(BaseGaussianDiffusion):
             # eq 15 - https://arxiv.org/abs/2102.09672
             min_log = extract(self.posterior_log_variance_clipped, t, x.shape)
             max_log = extract(torch.log(self.betas), t, x.shape)
-            var_interp_frac = unnormalize_img(var_interp_frac_unnormalized)
+            var_interp_frac = unnormalize_zero_to_one(var_interp_frac_unnormalized)
 
             posterior_log_variance = var_interp_frac * max_log + (1 - var_interp_frac) * min_log
             posterior_variance = posterior_log_variance.exp()
@@ -1859,11 +1859,21 @@ class Decoder(BaseGaussianDiffusion):
                 clip_denoised = clip_denoised
             )
 
-        return img
+        unnormalize_img = unnormalize_zero_to_one(img)
+        return unnormalize_img
 
     def p_losses(self, unet, x_start, times, *, image_embed, lowres_cond_img = None, text_encodings = None, text_mask = None, predict_x_start = False, noise = None, learned_variance = False, clip_denoised = False):
         noise = default(noise, lambda: torch.randn_like(x_start))
 
+        # normalize to [-1, 1]
+
+        x_start = normalize_neg_one_to_one(x_start)
+
+        if exists(lowres_cond_img):
+            lowres_cond_img = normalize_neg_one_to_one(lowres_cond_img)
+
+        # get x_t
+
         x_noisy = self.q_sample(x_start = x_start, t = times, noise = noise)
 
         model_output = unet(
@@ -1890,6 +1900,11 @@ class Decoder(BaseGaussianDiffusion):
             # return simple loss if not using learned variance
             return loss
 
+        # most of the code below is transcribed from
+        # https://github.com/hojonathanho/diffusion/blob/master/diffusion_tf/diffusion_utils_2.py
+        # the Improved DDPM paper then further modified it so that the mean is detached (shown a couple lines before), and weighted to be smaller than the l1 or l2 "simple" loss
+        # it is questionable whether this is really needed, looking at some of the figures in the paper, but may as well stay faithful to their implementation
+
         # if learning the variance, also include the extra weight kl loss
 
         true_mean, _, true_log_variance_clipped = self.q_posterior(x_start = x_start, x_t = x_noisy, t = times)

setup.py

@@ -10,7 +10,7 @@ setup(
       'dream = dalle2_pytorch.cli:dream'
     ],
   },
-  version = '0.2.14',
+  version = '0.2.16',
   license='MIT',
   description = 'DALL-E 2',
   author = 'Phil Wang',