belentani7 FLUX - Image Generation by Black Forest Labs
Open-weight image generation and editing models. FLUX.1 and FLUX.2 families.
When to Use
- User wants to generate images from text prompts
- User wants image editing (inpainting, outpainting)
- User needs high-quality AI art generation
- User wants open-source alternative to Midjourney/DALL-E
Models
| Model | License | VRAM | Best For |
|---|---|---|---|
| FLUX.1 [schnell] | Apache 2.0 | 8GB+ | Fast text-to-image |
| FLUX.1 [dev] | Non-Commercial | 24GB+ | Maximum quality |
| FLUX.2 [klein] 4B | Apache 2.0 | 8GB+ | Real-time generation |
| FLUX.2 [klein] 9B | Non-Commercial | 16GB+ | High quality |
| FLUX.2 [dev] | Non-Commercial | 80GB+ | Pro quality |
Installation
git clone https://github.com/black-forest-labs/flux.git
cd flux
pip install -r requirements.txt
Quick Start
FLUX.1 schnell (Apache 2.0)
from diffusers import FluxPipeline
import torch
pipe = FluxPipeline.from_pretrained(
"black-forest-labs/FLUX.1-schnell",
torch_dtype=torch.bfloat16
)
pipe.to("cuda")
image = pipe("A beautiful sunset over mountains").images[0]
image.save("output.png")
FLUX.2 klein 4B (Fastest, Apache 2.0)
from diffusers import Flux2Pipeline
pipe = Flux2Pipeline.from_pretrained(
"black-forest-labs/FLUX.2-klein-4B",
torch_dtype=torch.bfloat16
)
image = pipe("A cyberpunk city at night").images[0]
Key Parameters
| Parameter | Default | Description |
|---|---|---|
guidance_scale |
3.5-5.0 | Prompt adherence |
num_inference_steps |
28-50 | Quality vs speed |
width / height |
1024 | Resolution |
max_sequence_length |
512 | Prompt length |
Editing Modes
- Inpainting: Replace parts of an image
- Outpainting: Extend image boundaries
- Multi-reference: Use multiple reference images
- Image-to-Image: Transform existing images
UI Options
easydiffusion: 1-click GUI for Stable Diffusioncomfyui: Node-based workflow editorstable-diffusion-webui: Full-featured web UI