---
license: apache-2.0
datasets:
- CaptionEmporium/coyo-hd-11m-llavanext
- CortexLM/midjourney-v6
language:
- en
base_model:
- black-forest-labs/FLUX.1-dev
pipeline_tag: image-to-image
library_name: diffusers
---

<img src="assets/banner-dark.png?raw=true" alt="Banner Picture 1" style="width:1024px;"/>
<a href="https://discord.gg/FHY2guThfy">
<img src="https://github.com/XLabs-AI/x-flux/blob/main/assets/readme/light/join-our-discord-rev1.png?raw=true" style="width:1024px;"/>
</a>
<img src="assets/ip_adapter_0.jpg?raw=true" alt="example_0" style="width:1024px;"/>
<img src="assets/mona_workflow.jpg?raw=true" alt="Mona Anime Workflow 1" style="width:1024px;"/>

This repository provides a IP-Adapter checkpoint for
[FLUX.1-dev model](https://huggingface.co/black-forest-labs/FLUX.1-dev) by Black Forest Labs

[See our github](https://github.com/XLabs-AI/x-flux-comfyui) for comfy ui workflows.

# Models
The IP adapter is trained on a resolution of 512x512 for 150k steps and 1024x1024 for 350k steps while maintaining the aspect ratio.
We release **v2 version** - which can be used directly in ComfyUI!   

Please, see our [ComfyUI custom nodes installation guide](https://github.com/XLabs-AI/x-flux-comfyui)

# Examples

See examples of our models results below.  
Also, some generation results with input images are provided in "Files and versions"

# Inference

To try our models, you have 3 options:
1. Use main.py from our [official repo](https://github.com/XLabs-AI/x-flux)
2. Use our custom nodes for ComfyUI and test it with provided workflows (check out folder /workflows)
3. Diffusers 🧨

## Instruction for ComfyUI 
1. Go to ComfyUI/custom_nodes
2. Clone [x-flux-comfyui](https://github.com/XLabs-AI/x-flux-comfyui.git), path should be ComfyUI/custom_nodes/x-flux-comfyui/*, where * is all the files in this repo
3. Go to ComfyUI/custom_nodes/x-flux-comfyui/ and run python setup.py
4. Update x-flux-comfy with `git pull` or reinstall it.
5. Download Clip-L `model.safetensors` from [OpenAI VIT CLIP large](https://huggingface.co/openai/clip-vit-large-patch14), and put it to `ComfyUI/models/clip_vision/*`.
6. Download our IPAdapter from [huggingface](https://huggingface.co/XLabs-AI/flux-ip-adapter/tree/main), and put it to `ComfyUI/models/xlabs/ipadapters/*`.
7. Use `Flux Load IPAdapter` and `Apply Flux IPAdapter` nodes, choose right CLIP model and enjoy your genereations.
8. You can find example workflow in folder workflows in this repo.

## Diffusers 🧨
1. Install Diffusers 🧨 `pip install -U diffusers`
2. Run the example
```python
import torch
from diffusers import FluxPipeline
from diffusers.utils import load_image

pipe: FluxPipeline = FluxPipeline.from_pretrained(
    "black-forest-labs/FLUX.1-dev",
    torch_dtype=torch.bfloat16,
).to("cuda")

image = load_image("monalisa.jpg").resize((1024, 1024))

pipe.load_ip_adapter(
    "XLabs-AI/flux-ip-adapter-v2",
    weight_name="ip_adapter.safetensors",
    image_encoder_pretrained_model_name_or_path="openai/clip-vit-large-patch14"
)

def LinearStrengthModel(start, finish, size):
    return [
        (start + (finish - start) * (i / (size - 1))) for i in range(size)
    ]

ip_strengths = LinearStrengthModel(0.4, 1.0, 19)
pipe.set_ip_adapter_scale(ip_strengths)

image = pipe(
    width=1024,
    height=1024,
    prompt='wearing red sunglasses, golden chain and a green cap',
    negative_prompt="",
    true_cfg_scale=1.0,
    generator=torch.Generator().manual_seed(0),
    ip_adapter_image=image,
).images[0]

image.save('result.jpg')
```



If you get bad results, try to set to play with ip strength
### Limitations
The IP Adapter is currently in beta.
We do not guarantee that you will get a good result right away, it may take more attempts to get a result. 
<img src="assets/ip_adapter_2.jpg?raw=true" alt="example_2" style="width:1024px;"/>
<img src="assets/ip_adapter_3.jpg?raw=true" alt="example_3" style="width:1024px;"/>
<img src="assets/ip_adapter_1.jpg?raw=true" alt="example_1" style="width:1024px;"/>
<img src="assets/ip_adapter_4.jpg?raw=true" alt="example_4" style="width:1024px;"/>
<img src="assets/ip_adapter_5.jpg?raw=true" alt="example_5" style="width:1024px;"/>
<img src="assets/ip_adapter_6.jpg?raw=true" alt="example_6" style="width:1024px;"/>


## License
Our weights fall under the [FLUX.1 [dev]](https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md) Non-Commercial License<br/>