add pipeline and give image a 2nd try

run.py

@@ -5,6 +5,8 @@ import PIL.Image
 import numpy as np
 import tqdm
 
+# 1. Unroll the full loop
+# ==================================================================
 # load all models
 unet = UNetUnconditionalModel.from_pretrained("./", subfolder="unet")
 vqvae = VQModel.from_pretrained("./", subfolder="vqvae")
@@ -44,8 +46,31 @@ with torch.no_grad():
 
 # process image
 image_processed = image.cpu().permute(0, 2, 3, 1)
-image_processed = (image_processed + 1.0) * 127.5
+image_processed = image_processed * 255.
 image_processed = image_processed.numpy().astype(np.uint8)
 image_pil = PIL.Image.fromarray(image_processed[0])
 
-image_pil.save("generated_image.png")
+image_pil.save("generated_image_unrolled.png")
+
+
+# 2. Use pipeline
+# ==================================================================
+from diffusers import LatentDiffusionUncondPipeline
+import torch
+import PIL.Image
+import numpy as np
+import tqdm
+
+pipeline = LatentDiffusionUncondPipeline.from_pretrained("./")
+
+# generatae image by calling the pipeline
+generator = torch.manual_seed(0)
+image = pipeline(generator=generator, num_inference_steps=50)["sample"]
+
+# process image
+image_processed = image.cpu().permute(0, 2, 3, 1)
+image_processed = image_processed * 255.
+image_processed = image_processed.numpy().astype(np.uint8)
+image_pil = PIL.Image.fromarray(image_processed[0])
+
+image_pil.save("generated_image_pipeline.png")