make sure time embedding dimension is kept at 4 x dimension (thanks @borisdayma)

denoising_diffusion_pytorch/denoising_diffusion_pytorch.py

@@ -204,12 +204,12 @@ class Unet(nn.Module):
         in_out = list(zip(dims[:-1], dims[1:]))
 
         if with_time_emb:
-            time_dim = dim
+            time_dim = dim * 4
             self.time_mlp = nn.Sequential(
                 SinusoidalPosEmb(dim),
-                nn.Linear(dim, dim * 4),
+                nn.Linear(dim, time_dim),
                 nn.GELU(),
-                nn.Linear(dim * 4, dim)
+                nn.Linear(time_dim, time_dim)
             )
         else:
             time_dim = None
@@ -293,8 +293,8 @@ def cosine_beta_schedule(timesteps, s = 0.008):
     as proposed in https://openreview.net/forum?id=-NEXDKk8gZ
     """
     steps = timesteps + 1
-    x = torch.linspace(0, steps, steps)
+    x = torch.linspace(0, timesteps, steps)
-    alphas_cumprod = torch.cos(((x / steps) + s) / (1 + s) * torch.pi * 0.5) ** 2
+    alphas_cumprod = torch.cos(((x / timesteps) + s) / (1 + s) * torch.pi * 0.5) ** 2
     alphas_cumprod = alphas_cumprod / alphas_cumprod[0]
     betas = 1 - (alphas_cumprod[1:] / alphas_cumprod[:-1])
     return torch.clip(betas, 0, 0.999)

setup.py

@@ -3,7 +3,7 @@ from setuptools import setup, find_packages
 setup(
   name = 'denoising-diffusion-pytorch',
   packages = find_packages(),
-  version = '0.9.1',
+  version = '0.10.0',
   license='MIT',
   description = 'Denoising Diffusion Probabilistic Models - Pytorch',
   author = 'Phil Wang',