Test-1 Model

Introduction

The Test-1 model is a text-to-image generation model utilizing the FLUX.1-dev framework. It employs various cutting-edge techniques, such as diffusers, lora, and replicate, to transform text prompts into images. This model is designed for non-commercial use under the flux-1-dev-non-commercial-license.

Architecture

The model is based on the black-forest-labs/FLUX.1-dev architecture and is optimized for generating images from text prompts. It supports the integration of LoRA (Low-Rank Adaptation) weights for enhanced performance and customization.

Training

The Test-1 model was trained using the Replicate platform, employing the flux-dev-lora-trainer. The training process emphasized the use of trigger words, specifically "DGTOK", to initiate image generation effectively.

Guide: Running Locally

To run the Test-1 model locally, follow these steps:

1. Install the necessary libraries:

   pip install diffusers torch
   

2. Use the code snippet below to load and generate images using the model:

   from diffusers import AutoPipelineForText2Image
   import torch
   
   pipeline = AutoPipelineForText2Image.from_pretrained(
       'black-forest-labs/FLUX.1-dev',
       torch_dtype=torch.float16
   ).to('cuda')
   pipeline.load_lora_weights('rehctiw25/test-1', weight_name='lora.safetensors')
   image = pipeline('your prompt').images[0]
   

3. Ensure you have access to a CUDA-capable GPU to maximize performance. Consider using cloud-based GPU services like AWS or Google Cloud for efficient processing.

For detailed instructions on using and customizing LoRA weights, refer to the diffusers documentation.

License

The Test-1 model is released under the flux-1-dev-non-commercial-license. For full license details, visit the license page.