Overview

Description:

FLUX.1 is a collection of generative image AI models creating high quality, realistic images:

• FLUX.1-dev generates images from simple text prompts.
• FLUX.1-Canny-dev combines the text prompt with an image input processed to canny edges to guide the output image structure.
• FLUX.1-Depth-dev combines the text prompt with an image input processed to depth map leveraging LiheYoung/Depth-anything-large-hf model to guide the output image structure.

This model is ready for non-commercial use. Contact sales@blackforestlabs.ai for commercial terms.

Third-Party Community Consideration:

This model is not owned or developed by NVIDIA. This model has been developed and built to a third-party’s requirements for this application and use case; see link to:

• black-forest-labs/FLUX.1-dev Model Card
• black-forest-labs/FLUX.1-Canny-dev Model Card
• black-forest-labs/FLUX.1-Depth-dev Model Card
• LiheYoung/Depth-anything-large-hf Model Card

Terms of use

GOVERNING TERMS: The trial service is governed by the NVIDIA API Trial Terms of Service. Contact sales@blackforestlabs.ai for commercial terms to use the Flux.1-dev model. ADDITIONAL INFORMATION: Apache 2.0, NVIDIA Community Model License Agreement and Llama 2 Community Model License Agreement.

Deployment Geography:

Global

Use Case:

Creators and professionals can use this model to generate high-quality images from text prompts, simplifying visual communication.

Release Date:

August 1, 2024

References

• Flux on Black Forest Labs
• Flux blog post

Model Architecture:

Architecture Type: Transformer and Convolutional Neural Network (CNN)
Network Architecture: Diffusion Transformer
LiheYoung/Depth-anything-large-hf leverages the DPT architecture with a DINOv2 backbone.

Input:

Input Type: Text, Image (optional)
Input Parameters: Text: 1D. Image: 2D
Input Format: Text: String. Image: Red, Green, Blue (RGB)
Other Properties Related to Input: Steps, Classifier-Free Guidance Scale, Output Image Aspect Ratio, and Seed per the API Reference Page

Output:

Output Type: Image
Output Parameters: 2D
Output Format: Red, Green, Blue (RGB)
Other Properties Related to Output: 1024x1024, 768x1344, 1344x768, 1344x768, 1344x768, 1344x768, 1216x832

Software Integration:

Runtime Engines:

• TensorRT

Supported Hardware Platforms:

• NVIDIA Blackwell
  
• NVIDIA Hopper
  
• NVIDIA Lovelace
  

Supported Operating Systems: Linux, Windows Subsystem for Linux

Model Version(s):

• FLUX.1-dev
• FLUX.1-Canny-dev
• FLUX.1-Depth-dev
• LiheYoung/Depth-anything-large-hf

Training, Testing, and Evaluation Datasets:

Training Dataset:

• Data Collection Method by Dataset: Undisclosed
• Labeling Method by Dataset: Undisclosed

Properties (Quantity, Dataset Descriptions, Sensor(s)): Undisclosed

Testing Dataset:

• Data Collection Method by Dataset: Undisclosed
• Labeling Method by Dataset: Undisclosed

Properties (Quantity, Dataset Descriptions, Sensor(s)): Undisclosed

Evaluation Dataset:

• Data Collection Method by Dataset: Undisclosed
• Labeling Method by Dataset: Undisclosed

Properties (Quantity, Dataset Descriptions, Sensor(s)): Undisclosed

Inference:

Engine: TensorRT
Test Hardware: H100

Ethical Considerations:

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

Please report security vulnerabilities or NVIDIA AI Concerns here.