archer diffusion

Introduction

Archer Diffusion is a fine-tuned Stable Diffusion model designed for generating images in the style of the TV show "Archer." The model is trained on screenshots from the show and can produce images using prompts with the token "archer style."

Architecture

The model utilizes the Stable Diffusion framework and is compatible with the Diffusers library. It supports various optimizations and exports, including ONNX, MPS, and FLAX/JAX formats.

Training

Archer Diffusion was trained using the Diffusers-based Dreambooth training method, incorporating prior-preservation loss over 4,000 steps. It utilizes the train-text-encoder feature to enhance performance in generating Archer-style images.

Guide: Running Locally

1. Install Dependencies

   pip install diffusers transformers scipy torch
   

2. Load and Run the Model

   from diffusers import StableDiffusionPipeline
   import torch
   
   model_id = "nitrosocke/archer-diffusion"
   pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16)
   pipe = pipe.to("cuda")
   
   prompt = "a magical princess with golden hair, archer style"
   image = pipe(prompt).images[0]
   
   image.save("./magical_princess.png")
   

3. Cloud GPU Suggestions

   • Consider using cloud-based GPUs such as AWS EC2, Google Cloud Platform, or Azure for faster processing and efficiency.

License

This model is available under the CreativeML OpenRAIL-M license. Key points include:

1. Prohibition on using the model for illegal or harmful outputs.
2. No rights are claimed on generated outputs; users are accountable for their use.
3. Redistribution and commercial use are allowed, provided the same usage restrictions are retained and the license is shared with end users.

For full license details, refer to the CreativeML OpenRAIL-M license.