enna sketch style

Introduction

ENNA Sketch Style is a text-to-image model designed to generate character illustrations in a sketch style. It is part of the Hugging Face model repository and employs stable diffusion techniques to produce artistic outputs based on textual prompts.

Architecture

The model is based on the black-forest-labs/FLUX.1-dev architecture, utilizing the SD-LORA template. It incorporates elements of stable diffusion and LoRA (Low-Rank Adaptation) to adjust and fine-tune its illustration outputs. The model is optimized for generating sketch-style illustrations, characterized by a distinct linework style.

Training

The model was trained on a dataset of approximately 12 images. The training process involved adapting the model to capture the nuances of sketch illustration, including a unique linework style influenced by the creator's personal artistic style. Training was conducted on the Replicate platform, and commercial use requires explicit permission from the creator.

Guide: Running Locally

To run the ENNA Sketch Style model locally, follow these steps:

1. Install Dependencies: Ensure you have Python and the necessary libraries installed. Use tools like pip to install packages such as torch, transformers, and diffusers.

2. Download Model Weights: Access the model's weights available in Safetensors format from the Hugging Face Files & versions tab.

3. Set Up Environment:

   • Clone the repository or download the model files.
   • Load the model using a framework like PyTorch or TensorFlow that supports Safetensors.

4. Run Inference: Use the prompt "sketch illustration style" to generate images based on your input text.

For enhanced performance, especially for large-scale image generation tasks, consider using cloud-based GPUs from providers like AWS, Google Cloud, or Azure.

License

The model operates under the non-commercial-flux-dev-with-exceptions license. It is free to use for non-commercial purposes, but commercial use requires contacting the creator for permission. Detailed license information can be found here.