save clean-up

models/vision/ddim/modeling_ddim.py

@@ -40,39 +40,34 @@ class DDIM(DiffusionPipeline):
         inference_step_times = range(0, num_trained_timesteps, num_trained_timesteps // num_inference_steps)
 
         self.unet.to(torch_device)
-        x = self.noise_scheduler.sample_noise((batch_size, self.unet.in_channels, self.unet.resolution, self.unet.resolution), device=torch_device, generator=generator)
-
-        b = self.noise_scheduler.betas.to(torch_device)
-
-        seq = inference_step_times
-        seq_next = [-1] + list(seq[:-1])
-#        for t in tqdm.tqdm(reversed(range(num_inference_steps)), total=num_inference_steps):
-#            train_step = inference_step_times[t]
-        for i, j in zip(reversed(seq), reversed(seq_next)):
-
-            n = batch_size
-            x0_preds = []
-            xs = [x]
-
-#            i = train_step
-#            j = inference_step_times[t-1] if t > 0 else -1
-            if True:
-                print(i)
-                t = (torch.ones(n) * i).to(x.device)
-                next_t = (torch.ones(n) * j).to(x.device)
-                at = compute_alpha(b, t.long())
-                at_next = compute_alpha(b, next_t.long())
-                xt = xs[-1].to('cuda')
-                with torch.no_grad():
-                    et = self.unet(xt, t)
-                x0_t = (xt - et * (1 - at).sqrt()) / at.sqrt()
-                x0_preds.append(x0_t.to('cpu'))
-                # eta
-                c1 = (
-                    eta * ((1 - at / at_next) * (1 - at_next) / (1 - at)).sqrt()
-                )
-                c2 = ((1 - at_next) - c1 ** 2).sqrt()
-                xt_next = at_next.sqrt() * x0_t + c1 * torch.randn_like(x) + c2 * et
-                xs.append(xt_next.to('cpu'))
-
-        return xt_next
+        image = self.noise_scheduler.sample_noise((batch_size, self.unet.in_channels, self.unet.resolution, self.unet.resolution), device=torch_device, generator=generator)
+
+        for t in tqdm.tqdm(reversed(range(num_inference_steps)), total=num_inference_steps):
+            # get actual t and t-1
+            train_step = inference_step_times[t]
+            prev_train_step = inference_step_times[t - 1] if t > 0 else - 1
+
+            # compute alphas
+            alpha_prod_t = self.noise_scheduler.get_alpha_prod(train_step)
+            alpha_prod_t_prev = self.noise_scheduler.get_alpha_prod(prev_train_step)
+            alpha_prod_t_rsqrt = 1 / alpha_prod_t.sqrt()
+            alpha_prod_t_prev_rsqrt = 1 / alpha_prod_t_prev.sqrt()
+            beta_prod_t_sqrt = (1 - alpha_prod_t).sqrt()
+            beta_prod_t_prev_sqrt = (1 - alpha_prod_t_prev).sqrt()
+
+            # compute relevant coefficients
+            coeff_1 = (alpha_prod_t_prev - alpha_prod_t).sqrt() * alpha_prod_t_prev_rsqrt * beta_prod_t_prev_sqrt / beta_prod_t_sqrt * eta
+            coeff_2 = ((1 - alpha_prod_t_prev) - coeff_1 ** 2).sqrt()
+
+            with torch.no_grad():
+                noise_residual = self.unet(image, train_step)
+
+            print(train_step)
+
+            pred_mean = (image - noise_residual * beta_prod_t_sqrt) * alpha_prod_t_rsqrt
+            xt_next = alpha_prod_t_prev.sqrt() * pred_mean + coeff_1 * torch.randn_like(image) + coeff_2 * noise_residual
+#            xt_next = 1 / alpha_prod_t_rsqrt * pred_mean + coeff_1 * torch.randn_like(image) + coeff_2 * noise_residual
+            # eta
+            image = xt_next
+
+        return image