ddpm

app.py

@@ -188,11 +188,45 @@ def sample_ddim_context(n_sample, context, n=20):
     intermediate = np.stack(intermediate)
     return samples, intermediate
 
+# helper function; removes the predicted noise (but adds some noise back in to avoid collapse)
+def denoise_add_noise(x, t, pred_noise, z=None):
+    if z is None:
+        z = torch.randn_like(x)
+    noise = b_t.sqrt()[t] * z
+    mean = (x - pred_noise * ((1 - a_t[t]) / (1 - ab_t[t]).sqrt())) / a_t[t].sqrt()
+    return mean + noise
+
+# sample using standard algorithm
+@torch.no_grad()
+def sample_ddpm(n_sample, save_rate=20):
+    # x_T ~ N(0, 1), sample initial noise
+    samples = torch.randn(n_sample, 3, height, height).to(device)
+
+    # array to keep track of generated steps for plotting
+    intermediate = []
+    for i in range(timesteps, 0, -1):
+        print(f'sampling timestep {i:3d}', end='\r')
+
+        # reshape time tensor
+        t = torch.tensor([i / timesteps])[:, None, None, None].to(device)
+
+        # sample some random noise to inject back in. For i = 1, don't add back in noise
+        z = torch.randn_like(samples) if i > 1 else 0
+
+        eps = nn_model(samples, t)    # predict noise e_(x_t,t)
+        samples = denoise_add_noise(samples, i, eps, z)
+        if i % save_rate ==0 or i==timesteps or i<8:
+            intermediate.append(samples.detach().cpu().numpy())
+
+    intermediate = np.stack(intermediate)
+    return samples, intermediate
+
 def greet(input):
 	steps = int(input)
 	#samples, intermediate = sample_ddim(32, n=steps)
-	ctx = F.one_hot(torch.randint(0, 5, (32,)), 5).to(device=device).float()
-	samples, intermediate = sample_ddim_context(32, ctx, steps)
+	#ctx = F.one_hot(torch.randint(0, 5, (32,)), 5).to(device=device).float()
+	#samples, intermediate = sample_ddim_context(32, ctx, steps)
+	samples, intermediate = sample_ddpm(32, )
 	#response = transform2(transform(np.moveaxis(samples.detach().cpu().numpy(),1,3)[-1]))
 	#response2 = transform2(transform(np.moveaxis(samples.detach().cpu().numpy(),1,3)[1]))
 	#response = im.fromarray(intermediate[24][0][1]).convert("RGB")