Logo Design Flux Lo R A

Introduction

The Logo-Design-Flux-LoRA is a text-to-image model designed for creating logo designs using textual prompts. It employs a diffusion model leveraging LoRA (Low-Rank Adaptation) techniques to fine-tune the image generation process.

Architecture

This model uses a base model from black-forest-labs/FLUX.1-dev. It applies LoRA for efficient parameter tuning and is optimized with AdamW. The training process includes specific configurations such as a learning rate scheduler, noise offset, and multiresolution noise settings.

Training

• Parameters:
  • LR Scheduler: Constant
  • Optimizer: AdamW
  • Network Dimensions: 64
  • Epochs: 10, with checkpoints saved every epoch
• Data: The model is trained on 14 high-resolution images with a labeling focus on natural language and English.
• Best Output Dimensions: 1024 x 1024 pixels

Guide: Running Locally

To run the model locally, follow these steps:

1. Installation: Ensure you have PyTorch and the necessary dependencies installed.
2. Setup:

   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 = "prithivMLmods/Logo-Design-Flux-LoRA"
   pipe.load_lora_weights(lora_repo)
   
   device = torch.device("cuda")
   pipe.to(device)
   

3. Trigger Word: Use "Logo Design" to initiate the image generation.
4. Download Weights: Access the model weights in Safetensors format via the Files & Versions tab.

For enhanced performance, it is recommended to use cloud GPUs from providers like AWS, Google Cloud, or Azure.

License

The model is licensed under creativeml-openrail-m, which allows for creative and research use while adhering to specified conditions.