Update README.md

README.md

@@ -15,26 +15,52 @@ widget:
 base_model: stabilityai/stable-diffusion-xl-base-1.0
 license: cc-by-nc-nd-4.0
 ---
-# flash-sdxl
+# ⚡ FlashDiffusion: FlashSD ⚡
 
-<Gallery />
 
+Flash Diffusion is a diffusion distillation method proposed in [ADD ARXIV]() *by Clément Chadebec, Onur Tasar and Benjamin Aubin.*
+This model is a **26.4M** LoRA distilled version of SD1.5 model that is able to generate images in **2-4 steps**. The main purpose of this model is to reproduce the main results of the paper.
 
-## Trigger words
 
-You should use `Diffusion` to trigger the image generation.
+<p align="center">
+   <img style="width:700px;" src="images/hf_grid.png">
+</p>
 
-You should use `Distillation` to trigger the image generation.
+# How to use?
 
-You should use `text-to-image` to trigger the image generation.
+The model can be used using the `StableDiffusionPipeline` from `diffusers` library directly. It can allow reducing the number of required sampling steps to **2-4 steps**.
 
-You should use `fast image generation` to trigger the image generation.
+```python
+from diffusers import DiffusionPipeline, LCMScheduler
 
-You should use `sdxl` to trigger the image generation.
+adapter_id = "jasperai/flash-sd"
 
+pipe = DiffusionPipeline.from_pretrained(
+  "stabilityai/stable-diffusion-xl-base-1.0",
+  use_safetensors=True,
+)
 
-## Download model
+pipe.scheduler = LCMScheduler.from_pretrained(
+  "stabilityai/stable-diffusion-xl-base-1.0",
+  subfolder="scheduler",
+  timestep_spacing="trailing",
+)
+pipe.to("cuda")
 
-Weights for this model are available in Safetensors format.
+# Fuse and load LoRA weights
+pipe.load_lora_weights(adapter_id)
+pipe.fuse_lora()
 
-[Download](/jasperai/flash-sdxl/tree/main) them in the Files & versions tab.
+prompt = "A raccoon reading a book in a lush forest."
+
+image = pipe(prompt, num_inference_steps=4, guidance_scale=0).images[0]
+```
+<p align="center">
+   <img style="width:400px;" src="images/raccoon.png">
+</p>
+
+# Training Details
+
+
+## License
+This model is released under the the Creative Commons BY-NC license.