Introduction

The "VHS Box - LoRA (Flux.1 D)" model is a specialized text-to-image generator trained on images of horror and sci-fi VHS boxes from the 80s and 90s. Developed by Purz, this model utilizes the LoRA technique to create vintage-style VHS cover art.

Architecture

The model is based on the "black-forest-labs/FLUX.1-dev" architecture, integrating with the diffusers library for generating images. It supports LoRA weights for enhanced image synthesis capabilities.

Training

The model was trained using images of VHS covers from the 80s and 90s, focusing on horror and sci-fi genres. It leverages LoRA (Low-Rank Adaptation) to fine-tune the base model effectively for this niche task.

Guide: Running Locally

To run the model locally, follow these steps:

  1. Install the diffusers library:

    pip install diffusers
    
  2. Set up the environment:

    from diffusers import AutoPipelineForText2Image
    import torch
    
    device = "cuda" if torch.cuda.is_available() else "cpu"
    
  3. Load the pretrained model and LoRA weights:

    pipeline = AutoPipelineForText2Image.from_pretrained('black-forest-labs/FLUX.1-dev', torch_dtype=torch.bfloat16).to(device)
    pipeline.load_lora_weights('Purz/vhs-box', weight_name='purz-vhs_box.safetensors')
    
  4. Generate an image using a prompt:

    image = pipeline('vhs_box a VHS cover of the movie Fuckin Fart Goblin').images[0]
    

For optimal performance, it is recommended to use a cloud GPU service like AWS, Google Cloud, or Azure.

License

The model is distributed under the bespoke-lora-trained-license. Details can be found here.

More Related APIs in Text To Image