Create app.py

app.py

@@ -0,0 +1,82 @@
+from diffusers import DDPMPipeline
+image_pipe = DDPMPipeline.from_pretrained("google/ddpm-celebahq-256")
+image_pipe.to("cuda")
+images = image_pipe().images
+image_pipe
+from diffusers import UNet2DModel
+
+repo_id = "google/ddpm-church-256"
+model = UNet2DModel.from_pretrained(repo_id)
+model
+model.config
+model_random = UNet2DModel(**model.config)
+model_random.save_pretrained("my_model")
+model_random = UNet2DModel.from_pretrained("my_model")
+import torch
+
+torch.manual_seed(0)
+
+noisy_sample = torch.randn(
+    1, model.config.in_channels, model.config.sample_size, model.config.sample_size
+)
+noisy_sample.shape
+with torch.no_grad():
+    noisy_residual = model(sample=noisy_sample, timestep=2).sample
+noisy_residual.shape
+from diffusers import DDPMScheduler
+
+scheduler = DDPMScheduler.from_config(repo_id)
+scheduler.config
+scheduler.save_config("my_scheduler")
+new_scheduler = DDPMScheduler.from_config("my_scheduler")
+less_noisy_sample = scheduler.step(
+    model_output=noisy_residual, timestep=2, sample=noisy_sample
+).prev_sample
+less_noisy_sample.shape
+import PIL.Image
+import numpy as np
+
+def display_sample(sample, i):
+    image_processed = sample.cpu().permute(0, 2, 3, 1)
+    image_processed = (image_processed + 1.0) * 127.5
+    image_processed = image_processed.numpy().astype(np.uint8)
+
+    image_pil = PIL.Image.fromarray(image_processed[0])
+    display(f"Image at step {i}")
+    display(image_pil)
+model.to("cuda")
+noisy_sample = noisy_sample.to("cuda")
+import tqdm
+
+sample = noisy_sample
+
+for i, t in enumerate(tqdm.tqdm(scheduler.timesteps)):
+  # 1. predict noise residual
+  with torch.no_grad():
+      residual = model(sample, t).sample
+
+  # 2. compute less noisy image and set x_t -> x_t-1
+  sample = scheduler.step(residual, t, sample).prev_sample
+
+  # 3. optionally look at image
+  if (i + 1) % 50 == 0:
+      display_sample(sample, i + 1)
+from diffusers import DDIMScheduler
+
+scheduler = DDIMScheduler.from_config(repo_id)
+scheduler.set_timesteps(num_inference_steps=50)
+import tqdm
+
+sample = noisy_sample
+
+for i, t in enumerate(tqdm.tqdm(scheduler.timesteps)):
+  # 1. predict noise residual
+  with torch.no_grad():
+      residual = model(sample, t).sample
+
+  # 2. compute previous image and set x_t -> x_t-1
+  sample = scheduler.step(residual, t, sample).prev_sample
+
+  # 3. optionally look at image
+  if (i + 1) % 10 == 0:
+      display_sample(sample, i + 1)