Merge branch 'main' of https://huggingface.co/tianweiy/DMD2 into main

README.md

@@ -22,9 +22,9 @@ Tianwei Yin [tianweiy@mit.edu](mailto:tianweiy@mit.edu)
 
 ## Huggingface Demo
 
-Our 4-step (much higher quality, 2X slower) Text-to-Image demo is hosted at [DMD2-4step](https://913f7051c61c070e4e.gradio.live)
+Our 4-step (much higher quality, 2X slower) Text-to-Image demo is hosted at [DMD2-4step](https://6cf215173601f32482.gradio.live)
 
-Our 1-step Text-to-Image demo is hosted at [DMD2-1step](https://154dfe6ee5c63946cc.gradio.live)
+Our 1-step Text-to-Image demo is hosted at [DMD2-1step](https://cc2622c0c132346c64.gradio.live)
 
 ## Usage
 
@@ -46,7 +46,9 @@ unet.load_state_dict(torch.load(hf_hub_download(repo_name, ckpt_name), map_locat
 pipe = DiffusionPipeline.from_pretrained(base_model_id, unet=unet, torch_dtype=torch.float16, variant="fp16").to("cuda")
 pipe.scheduler = LCMScheduler.from_config(pipe.scheduler.config)
 prompt="a photo of a cat"
-image=pipe(prompt=prompt, num_inference_steps=4, guidance_scale=0).images[0]
+
+# LCMScheduler's default timesteps are different from the one we used for training 
+image=pipe(prompt=prompt, num_inference_steps=4, guidance_scale=0, timesteps=[999, 749, 499, 249]).images[0]
 ```
 
 #### 1-step generation