README.md

metadata

title: FLUX.2 Text to Image
emoji: 🖼
colorFrom: purple
colorTo: red
sdk: gradio
sdk_version: 6.0.2
app_file: app.py
pinned: true
license: apache-2.0
short_description: Generate production-grade AI images using FLUX.2 [dev]

FLUX.2 [dev] Text-to-Image

Generate high-quality images using FLUX.2 [dev], a 32B parameter rectified flow model by Black Forest Labs.

Features

• Text-to-Image Generation: Create images from text prompts
• Image Editing: Upload images for editing and manipulation
• Image Combining: Combine multiple images based on text instructions
• Prompt Upsampling: Automatically enhance prompts using a VLM (optional)
• ZeroGPU Support: Optimized for ZeroGPU inference
• Advanced Controls: Fine-tune generation with seed, guidance scale, inference steps, and dimensions

Setup

This Space requires:

• ZeroGPU: Enable ZeroGPU in your Space settings
• HF_TOKEN: Set your Hugging Face token as a Space secret for gated model access
  • Go to Settings → Secrets → Add HF_TOKEN with your Hugging Face token

Model

• Model: black-forest-labs/FLUX.2-dev
• Blog: FLUX.2 Announcement

Usage

1. Enter a text prompt describing the image you want to generate
2. (Optional) Upload one or more images for editing/combining
3. Adjust advanced settings as needed:
   • Prompt Upsampling: Enable to automatically enhance your prompt
   • Seed: Control randomness (use randomize for variety)
   • Dimensions: Width and height (must be multiples of 8)
   • Inference Steps: More steps = higher quality but slower (default: 30)
   • Guidance Scale: How closely to follow the prompt (default: 4.0)
4. Click "Run" to generate your image

Notes

• The model uses pre-compiled blocks for ZeroGPU optimization
• Prompt upsampling requires a valid HF_TOKEN secret
• Image dimensions are automatically adjusted when uploading images
• Supports image editing and combining multiple images

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference