README.md

metadata

license: openrail++
language:
  - en
tags:
  - stable-diffusion
  - text-to-image
  - diffusers
thumbnail: >-
  https://huggingface.co/nitrosocke/redshift-diffusion-768/resolve/main/images/redshift-diffusion-768-thumbnail.jpg
inference: false

Future Diffusion

This is the fine-tuned Stable Diffusion 2.0 model trained on high quality 3D images with a 768x768 pixel resolution. Use the tokens
redshift style
in your prompts for the effect. Trained on Stability.ai's Stable Diffusion 2.0 with 768x768 resolution.

If you enjoy my work and want to test new models before release, please consider supporting me

• The Weights (or .ckpt file) are not released yet as there is no working conversion script yet. You can try out the model here: Diffusion Space Demo
• Disclaimer: The SD 2.0 model is just a few days old at this point and we still need to figure out how it works exactly. Please view this as an early prototype and experiment with the model.

Characters rendered with the model: Cars and Animals rendered with the model: Landscapes rendered with the model:

Prompt and settings for the Characters:

redshift style portrait black female cyberpunk hacker tattoos colorful short hair wearing a crop top redshift style Negative Prompt: mutated body double head bad anatomy long face long neck long body text watermark signature
Steps: 20, Sampler: Euler a, CFG scale: 7, Size: 768x1024

Prompt and settings for the Landscapes:

redshift style beautiful fjord at sunrise Negative Prompt: fog blurry soft
Steps: 20, Sampler: Euler a, CFG scale: 7, Size: 1536x768

This model was trained using the diffusers based dreambooth training by ShivamShrirao using prior-preservation loss and the train-text-encoder flag in 7.500 steps.

License

This model is open access and available to all, with a CreativeML Open RAIL++-M License further specifying rights and usage. Please read the full license here