Flux Super Blend Lo R A

Introduction

Flux-Super-Blend-LoRA is a text-to-image model designed to create imaginative and complex scenes based on textual prompts. It combines various image blending techniques and is trained to generate diverse and artistic outputs, often featuring fantastical or surreal elements.

Architecture

The model is based on the black-forest-labs/FLUX.1-dev architecture. It employs LoRA (Low-Rank Adaptation) for efficient fine-tuning, allowing it to manage complex image generation tasks with reduced computational requirements. Key parameters include a learning rate scheduler set to constant, and the optimizer used is AdamW.

Training

The model was trained with a dataset of 30 images using the strangerzonehf/Super-Blend-Image-Collection. The training configuration includes a network dimension of 64 and a network alpha of 32, with a total of 20 epochs. Images are best processed at dimensions of 768x1024 or 1024x1024.

Guide: Running Locally

To run this model locally, follow these steps:

1. Setup Environment: Ensure you have Python and PyTorch installed. Use a virtual environment to manage dependencies.
2. Install Required Libraries:

   pip install torch
   pip install diffusers
   

3. Load the Model:

   import torch
   from pipelines import DiffusionPipeline
   
   base_model = "black-forest-labs/FLUX.1-dev"
   pipe = DiffusionPipeline.from_pretrained(base_model, torch_dtype=torch.bfloat16)
   
   lora_repo = "strangerzonehf/Flux-Super-Blend-LoRA"
   trigger_word = "Super Blend"  
   pipe.load_lora_weights(lora_repo)
   
   device = torch.device("cuda")
   pipe.to(device)
   

4. Generate Images: Use creative prompts with the trigger word "Super Blend" to start image generation.

For optimal performance and faster processing, it is recommended to use cloud GPUs from providers like AWS, Google Cloud, or Azure.

License

The model is licensed under the CreativeML Open RAIL-M license, allowing for both commercial and non-commercial use while adhering to ethical AI use guidelines.