​
Try out this model on VESSL Hub.

A simple web app for stable diffusion inference is deployed in this example. Some SD model checkpoints are mounted with a VESSL Model so you can try some generation instantly.

Stable diffusion is a deep learning, text-to-image model that uses a diffusion technique, which generates an image from noise by iterating gradual denoising steps. Unlike other text-to-image models, stable diffusion performs a diffusion process in the latent space with smaller dimensions and reconstructs the result to the image in a real dimension. Also, a cross-attention mechanism is added for multi-modal tasks such as text-to-image and layout-to-image tasks.

​
Running the model

You can run the model with our quick command.

vessl run crewate -f sd-webui.yaml

Here’s a rundown of the sd-webui.yaml file.

name: stable-diffusion-webui
description: A template Run for stable diffusion webui app
resources:
  cluster: google-oregon
  preset: v1.l4-1.mem-42
image: quay.io/vessl-ai/hub:torch2.1.0-cuda12.2-202312070053
import:
  /code/:
    git:
      url: https://github.com/vessl-ai/hub-model
      ref: main
  /models/protogen-infinity/: hf://huggingface.co/darkstorm2150/Protogen_Infinity_Official_Release
  /models/sd-v1-5/: hf://huggingface.co/VESSL/stable-diffusion-v1-5-checkpoint
  /models/sd-v2-1/: hf://huggingface.co/VESSL/stable-diffusion-v2-1-checkpoint
run:
  - command: |-
      pip install -r requirements.txt
      python -u launch.py --no-download-sd-model --ckpt-dir /models --no-half --no-gradio-queue --listen 
    workdir: /code/stable-diffusion-webui
interactive:
  max_runtime: 24h
  jupyter:
    idle_timeout: 120m
ports:
  - name: gradio
    type: http
    port: 7860

Was this page helpful?

SSD-1B PlaygroundWhisper V3 Playground