make sure classifier free guidance is used only if conditional dropout is present on the DiffusionPrior and Decoder classes. also make sure prior can have a different conditional scale than decoder

README.md

@@ -706,7 +706,7 @@ mock_image_embed = torch.randn(1, 512).cuda()
 images = decoder.sample(mock_image_embed) # (1, 3, 1024, 1024)
 ```
 
-## Training wrapper (wip)
+## Training wrapper
 
 ### Decoder Training
 
@@ -851,6 +851,57 @@ diffusion_prior_trainer.update()  # this will update the optimizer as well as th
 image_embeds = diffusion_prior_trainer.sample(text) # (4, 512) - exponential moving averaged image embeddings
 ```
 
+## Bonus
+
+### Unconditional Training
+
+The repository also contains the means to train unconditional DDPM model, or even cascading DDPMs. You simply have to set `unconditional = True` in the `Decoder`
+
+ex.
+
+```python
+import torch
+from dalle2_pytorch import Unet, Decoder
+
+# unet for the cascading ddpm
+
+unet1 = Unet(
+    dim = 128,
+    dim_mults=(1, 2, 4, 8)
+).cuda()
+
+unet2 = Unet(
+    dim = 32,
+    dim_mults = (1, 2, 4, 8, 16)
+).cuda()
+
+# decoder, which contains the unets
+
+decoder = Decoder(
+    unet = (unet1, unet2),
+    image_sizes = (256, 512),  # first unet up to 256px, then second to 512px
+    timesteps = 1000,
+    unconditional = True
+).cuda()
+
+# mock images (get a lot of this)
+
+images = torch.randn(1, 3, 512, 512).cuda()
+
+# feed images into decoder
+
+for i in (1, 2):
+    loss = decoder(images, unet_number = i)
+    loss.backward()
+
+# do the above for many many many many steps
+# then it will learn to generate images
+
+images = decoder.sample(batch_size = 2) # (2, 3, 512, 512)
+```
+
+## Dataloaders
+
 ### Decoder Dataloaders
 
 In order to make loading data simple and efficient, we include some general dataloaders that can be used to train portions of the network.
@@ -1014,6 +1065,7 @@ Once built, images will be saved to the same directory the command is invoked
 - [ ] allow for unet to be able to condition non-cross attention style as well
 - [ ] for all model classes with hyperparameters that changes the network architecture, make it requirement that they must expose a config property, and write a simple function that asserts that it restores the object correctly
 - [ ] for both diffusion prior and decoder, all exponential moving averaged models needs to be saved and restored as well (as well as the step number)
+- [ ] read the paper, figure it out, and build it https://github.com/lucidrains/DALLE2-pytorch/issues/89
 
 ## Citations
 
@@ -1102,4 +1154,13 @@ Once built, images will be saved to the same directory the command is invoked
 }
 ```
 
+```bibtex
+@article{ho2021cascaded,
+    title   = {Cascaded Diffusion Models for High Fidelity Image Generation},
+    author  = {Ho, Jonathan and Saharia, Chitwan and Chan, William and Fleet, David J and Norouzi, Mohammad and Salimans, Tim},
+    journal = {arXiv preprint arXiv:2106.15282},
+    year    = {2021}
+}
+```
+
 *Creating noise from data is easy; creating data from noise is generative modeling.* - <a href="https://arxiv.org/abs/2011.13456">Yang Song's paper</a>

dalle2_pytorch/dalle2_pytorch.py

@@ -794,7 +794,7 @@ class DiffusionPriorNetwork(nn.Module):
         text_embed,
         text_encodings = None,
         mask = None,
-        cond_drop_prob = 0.2
+        cond_drop_prob = 0.
     ):
         batch, dim, device, dtype = *image_embed.shape, image_embed.device, image_embed.dtype
 
@@ -901,6 +901,7 @@ class DiffusionPrior(BaseGaussianDiffusion):
         self.channels = default(image_channels, lambda: clip.image_channels)
 
         self.cond_drop_prob = cond_drop_prob
+        self.can_classifier_guidance = cond_drop_prob > 0.
         self.condition_on_text_encodings = condition_on_text_encodings
 
         # in paper, they do not predict the noise, but predict x0 directly for image embedding, claiming empirically better results. I'll just offer both.
@@ -914,8 +915,10 @@ class DiffusionPrior(BaseGaussianDiffusion):
         self.training_clamp_l2norm = training_clamp_l2norm
         self.init_image_embed_l2norm = init_image_embed_l2norm
 
-    def p_mean_variance(self, x, t, text_cond, clip_denoised: bool):
+    def p_mean_variance(self, x, t, text_cond, clip_denoised = False, cond_scale = 1.):
-        pred = self.net(x, t, **text_cond)
+        assert not (cond_scale != 1. and not self.can_classifier_guidance), 'the model was not trained with conditional dropout, and thus one cannot use classifier free guidance (cond_scale anything other than 1)'
+
+        pred = self.net.forward_with_cond_scale(x, t, cond_scale = cond_scale, **text_cond)
 
         if self.predict_x_start:
             x_recon = pred
@@ -934,16 +937,16 @@ class DiffusionPrior(BaseGaussianDiffusion):
         return model_mean, posterior_variance, posterior_log_variance
 
     @torch.inference_mode()
-    def p_sample(self, x, t, text_cond = None, clip_denoised = True, repeat_noise = False):
+    def p_sample(self, x, t, text_cond = None, clip_denoised = True, repeat_noise = False, cond_scale = 1.):
         b, *_, device = *x.shape, x.device
-        model_mean, _, model_log_variance = self.p_mean_variance(x = x, t = t, text_cond = text_cond, clip_denoised = clip_denoised)
+        model_mean, _, model_log_variance = self.p_mean_variance(x = x, t = t, text_cond = text_cond, clip_denoised = clip_denoised, cond_scale = cond_scale)
         noise = noise_like(x.shape, device, repeat_noise)
         # no noise when t == 0
         nonzero_mask = (1 - (t == 0).float()).reshape(b, *((1,) * (len(x.shape) - 1)))
         return model_mean + nonzero_mask * (0.5 * model_log_variance).exp() * noise
 
     @torch.inference_mode()
-    def p_sample_loop(self, shape, text_cond):
+    def p_sample_loop(self, shape, text_cond, cond_scale = 1.):
         device = self.betas.device
 
         b = shape[0]
@@ -954,7 +957,7 @@ class DiffusionPrior(BaseGaussianDiffusion):
 
         for i in tqdm(reversed(range(0, self.num_timesteps)), desc='sampling loop time step', total=self.num_timesteps):
             times = torch.full((b,), i, device = device, dtype = torch.long)
-            image_embed = self.p_sample(image_embed, times, text_cond = text_cond)
+            image_embed = self.p_sample(image_embed, times, text_cond = text_cond, cond_scale = cond_scale)
 
         return image_embed
 
@@ -980,19 +983,19 @@ class DiffusionPrior(BaseGaussianDiffusion):
 
     @torch.inference_mode()
     @eval_decorator
-    def sample_batch_size(self, batch_size, text_cond):
+    def sample_batch_size(self, batch_size, text_cond, cond_scale = 1.):
         device = self.betas.device
         shape = (batch_size, self.image_embed_dim)
 
         img = torch.randn(shape, device = device)
 
         for i in tqdm(reversed(range(0, self.num_timesteps)), desc = 'sampling loop time step', total = self.num_timesteps):
-            img = self.p_sample(img, torch.full((batch_size,), i, device = device, dtype = torch.long), text_cond = text_cond)
+            img = self.p_sample(img, torch.full((batch_size,), i, device = device, dtype = torch.long), text_cond = text_cond, cond_scale = cond_scale)
         return img
 
     @torch.inference_mode()
     @eval_decorator
-    def sample(self, text, num_samples_per_batch = 2):
+    def sample(self, text, num_samples_per_batch = 2, cond_scale = 1.):
         # in the paper, what they did was
         # sample 2 image embeddings, choose the top 1 similarity, as judged by CLIP
         text = repeat(text, 'b ... -> (b r) ...', r = num_samples_per_batch)
@@ -1007,7 +1010,7 @@ class DiffusionPrior(BaseGaussianDiffusion):
         if self.condition_on_text_encodings:
             text_cond = {**text_cond, 'text_encodings': text_encodings, 'mask': text_mask}
 
-        image_embeds = self.p_sample_loop((batch_size, image_embed_dim), text_cond = text_cond)
+        image_embeds = self.p_sample_loop((batch_size, image_embed_dim), text_cond = text_cond, cond_scale = cond_scale)
 
         # retrieve original unscaled image embed
 
@@ -1305,7 +1308,7 @@ class Unet(nn.Module):
         self,
         dim,
         *,
-        image_embed_dim,
+        image_embed_dim = None,
         text_embed_dim = None,
         cond_dim = None,
         num_image_tokens = 4,
@@ -1377,7 +1380,7 @@ class Unet(nn.Module):
         self.image_to_cond = nn.Sequential(
             nn.Linear(image_embed_dim, cond_dim * num_image_tokens),
             Rearrange('b (n d) -> b n d', n = num_image_tokens)
-        ) if image_embed_dim != cond_dim else nn.Identity()
+        ) if cond_on_image_embeds and image_embed_dim != cond_dim else nn.Identity()
 
         self.norm_cond = nn.LayerNorm(cond_dim)
         self.norm_mid_cond = nn.LayerNorm(cond_dim)
@@ -1387,7 +1390,8 @@ class Unet(nn.Module):
         self.text_to_cond = None
 
         if cond_on_text_encodings:
-            self.text_to_cond = nn.LazyLinear(cond_dim) if not exists(text_embed_dim) else nn.Linear(text_embed_dim, cond_dim)
+            assert exists(text_embed_dim), 'text_embed_dim must be given to the unet if cond_on_text_encodings is True'
+            self.text_to_cond = nn.Linear(text_embed_dim, cond_dim)
 
         # finer control over whether to condition on image embeddings and text encodings
         # so one can have the latter unets in the cascading DDPMs only focus on super-resoluting
@@ -1701,7 +1705,7 @@ class Decoder(BaseGaussianDiffusion):
         self.unconditional = unconditional
         assert not (condition_on_text_encodings and unconditional), 'unconditional decoder image generation cannot be set to True if conditioning on text is present'
 
-        assert exists(clip) ^ exists(image_size), 'either CLIP is supplied, or you must give the image_size and channels (usually 3 for RGB)'
+        assert self.unconditional or (exists(clip) ^ exists(image_size)), 'either CLIP is supplied, or you must give the image_size and channels (usually 3 for RGB)'
 
         self.clip = None
         if exists(clip):
@@ -1792,6 +1796,7 @@ class Decoder(BaseGaussianDiffusion):
 
         self.image_cond_drop_prob = image_cond_drop_prob
         self.text_cond_drop_prob = text_cond_drop_prob
+        self.can_classifier_guidance = image_cond_drop_prob > 0. or text_cond_drop_prob > 0.
 
         # whether to clip when sampling
 
@@ -1818,6 +1823,8 @@ class Decoder(BaseGaussianDiffusion):
         unet.cpu()
 
     def p_mean_variance(self, unet, x, t, image_embed, text_encodings = None, text_mask = None, lowres_cond_img = None, clip_denoised = True, predict_x_start = False, learned_variance = False, cond_scale = 1., model_output = None):
+        assert not (cond_scale != 1. and not self.can_classifier_guidance), 'the decoder was not trained with conditional dropout, and thus one cannot use classifier free guidance (cond_scale anything other than 1)'
+
         pred = default(model_output, lambda: unet.forward_with_cond_scale(x, t, image_embed = image_embed, text_encodings = text_encodings, text_mask = text_mask, cond_scale = cond_scale, lowres_cond_img = lowres_cond_img))
 
         if learned_variance:
@@ -2036,12 +2043,12 @@ class Decoder(BaseGaussianDiffusion):
 
         times = torch.randint(0, self.num_timesteps, (b,), device = device, dtype = torch.long)
 
-        if not exists(image_embed):
+        if not exists(image_embed) and not self.unconditional:
             assert exists(self.clip), 'if you want to derive CLIP image embeddings automatically, you must supply `clip` to the decoder on init'
             image_embed, _ = self.clip.embed_image(image)
 
         text_encodings = text_mask = None
-        if exists(text) and not exists(text_encodings):
+        if exists(text) and not exists(text_encodings) and not self.unconditional:
             assert exists(self.clip), 'if you are passing in raw text, you need to supply `clip` to the decoder'
             _, text_encodings, text_mask = self.clip.embed_text(text)
 
@@ -2093,6 +2100,7 @@ class DALLE2(nn.Module):
         self,
         text,
         cond_scale = 1.,
+        prior_cond_scale = 1.,
         return_pil_images = False
     ):
         device = next(self.parameters()).device
@@ -2102,7 +2110,7 @@ class DALLE2(nn.Module):
             text = [text] if not isinstance(text, (list, tuple)) else text
             text = tokenizer.tokenize(text).to(device)
 
-        image_embed = self.prior.sample(text, num_samples_per_batch = self.prior_num_samples)
+        image_embed = self.prior.sample(text, num_samples_per_batch = self.prior_num_samples, cond_scale = prior_cond_scale)
 
         text_cond = text if self.decoder_need_text_cond else None
         images = self.decoder.sample(image_embed, text = text_cond, cond_scale = cond_scale)

dalle2_pytorch/train.py

@@ -335,11 +335,6 @@ class DecoderTrainer(nn.Module):
         self.num_unets = len(self.decoder.unets)
 
         self.use_ema = use_ema
-
-        if use_ema:
-            has_lazy_linear = any([type(module) == nn.LazyLinear for module in decoder.modules()])
-            assert not has_lazy_linear, 'you must set the text_embed_dim on your u-nets if you plan on doing automatic exponential moving average'
-
         self.ema_unets = nn.ModuleList([])
 
         self.amp = amp

setup.py

@@ -10,7 +10,7 @@ setup(
       'dream = dalle2_pytorch.cli:dream'
     ],
   },
-  version = '0.2.34',
+  version = '0.2.38',
   license='MIT',
   description = 'DALL-E 2',
   author = 'Phil Wang',