Update README

README.md

@@ -29,38 +29,31 @@ For details on the development and training of our model, please refer to our bl
 
 ### Using the model with 🧨 Diffusers
 
-Install diffusers >= 0.26.0 and some dependencies:
+Install diffusers >= 0.27.0 and the relevant dependencies. For now, you need to install from the `main` diffusers branch in GitHub until a new release is published in PyPi.
 
 ```
+pip install git+https://github.com/huggingface/diffusers.git
 pip install transformers accelerate safetensors
 ```
 
-To run our model, you will need to use our custom pipeline from this gist: https://gist.github.com/aykamko/402e948a8fdbbc9613f9978802d90194
-
 **Notes:**
-- Only the Euler, Heun, and DPM++ 2M Karras schedulers have been tested
-- We recommend using `guidance_scale=7.0` for the Euler/Heun, and `guidance_scale=5.0` for DPM++ 2M Karras
+- The pipeline uses the `EDMEulerScheduler` scheduler. It's an [EDM formulation](https://arxiv.org/abs/2206.00364) of the Euler scheduler.
+- `guidance_scale=5.0` is a good default for this scheduler.
 
 Then, run the following snippet:
 
 ```python
-# copy/paste pipeline code here from gist: https://gist.github.com/aykamko/402e948a8fdbbc9613f9978802d90194
+from diffusers import DiffusionPipeline
+import torch
 
-pipe = PlaygroundV2dot5Pipeline.from_pretrained(
+pipe = DiffusionPipeline.from_pretrained(
     "playgroundai/playground-v2.5-1024px-aesthetic",
     torch_dtype=torch.float16,
-    use_safetensors=True,
-    add_watermarker=False,
     variant="fp16",
-)
-pipe.to("cuda")
-
-# # Optional: use DPM++ 2M Karras scheduler for improved quality on small details
-# from diffusers import DPMSolverMultistepScheduler
-# pipe.scheduler = DPMSolverMultistepScheduler(**common_config, use_karras_sigmas=True)
+).to("cuda")
 
-prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k"
-image  = pipe(prompt=prompt, guidance_scale=7.0).images[0]
+prompt = "a chihuahua riding on the back of a golden retriever, in front of the Eiffel tower"
+image = pipe(prompt=prompt, num_inference_steps=25, guidance_scale=5).images[0]
 ```
 
 ### Using the model with Automatic1111/ComfyUI