|
--- |
|
license: other |
|
license_name: flux-1-dev-non-commercial-license |
|
license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/resolve/main/LICENSE.md |
|
base_model: |
|
- black-forest-labs/FLUX.1-dev |
|
pipeline_tag: text-to-image |
|
library_name: diffusers |
|
tags: |
|
- flux |
|
- text-to-image |
|
--- |
|
|
|
 |
|
|
|
# Flux.1 Lite |
|
|
|
We are thrilled to announce the alpha release of Flux.1 Lite, an 8B parameter transformer model distilled from the FLUX.1-dev model. |
|
|
|
Our goal? To distill FLUX.1-dev into a lighter model, reducing the parameters to just 24 GB, so it can run smoothly on most consumer-grade GPU cards, making high-quality AI models accessible to everyone. |
|
|
|
 |
|
|
|
## Motivation |
|
|
|
As stated by other members of the community like [Ostris](https://ostris.com/2024/09/07/skipping-flux-1-dev-blocks/), it seems that blocks of the Flux1.dev transformer have a different contribution to the final image generation. To explore this, we analyzed the Mean Squared Error (MSE) between the input and output of each block, revealing significant variability. |
|
|
|
Our findings? Not all blocks are created equal. By strategically skipping less impactful blocks, we've managed to achieve incredible efficiency gains without compromising on quality. The results are striking: skipping just one of the early MMDIT blocks can significantly impact model performance, whereas other blocks have a much smaller effect. |
|
|
|
 |
|
 |
|
 |
|
|
|
Furthermore, as displayed in the following image, only when you skip one of the first MMDIT blocks, the performance of the model severely impacts the model's performance. |
|
 |
|
 |
|
|
|
## Text-to-Image Usage |
|
|
|
Flux.1 Lite is ready to unleash your creativity! For the best results, we recommend using a `guidance_scale` of 3.5 and setting `n_steps` between 22 and 30. |
|
|
|
```python |
|
import torch |
|
from diffusers import FluxPipeline |
|
|
|
base_model_id = "Freepik/flux.1-lite-8B-alpha" |
|
torch_dtype = torch.bfloat16 |
|
device = "cuda" |
|
|
|
# Load the pipe |
|
model_id = "Freepik/flux.1-lite-8B-alpha" |
|
pipe = FluxPipeline.from_pretrained( |
|
model_id, torch_dtype=torch_dtype |
|
).to(device) |
|
|
|
# Inference |
|
prompt = "A close-up image of a green alien with fluorescent skin in the middle of a dark purple forest" |
|
|
|
guidance_scale = 3.5 # Keep guidance_scale at 3.5 |
|
n_steps = 28 |
|
seed = 11 |
|
|
|
with torch.inference_mode(): |
|
image = pipe( |
|
prompt=prompt, |
|
generator=torch.Generator(device="cpu").manual_seed(seed), |
|
num_inference_steps=n_steps, |
|
guidance_scale=guidance_scale, |
|
height=1024,s |
|
width=1024, |
|
).images[0] |
|
image.save("output.png") |
|
``` |
|
|
|
## ComfyUI |
|
We've also crafted a ComfyUI workflow to make using Flux.1 Lite even more seamless! Find it in `comfy/flux.1-lite_workflow.json`. |
|
 |
|
|
|
## Checkpoints |
|
* `flux.1-lite-8B-alpha.safetensors`: Transformer checkpoint, in Flux original format. |
|
* `transformers/`: Contains distilled 8B transformer model, in diffusers format. |
|
|
|
## 🤗 Hugging Face space: |
|
Flux.1 Lite demo host on [🤗 flux.1-lite](https://huggingface.co/spaces/Freepik/flux.1-lite) |
|
|
|
## 🔥 News 🔥 |
|
* Oct.18, 2024. Alpha 8B checkpoint and comparison demo 🤗 (i.e. [Flux.1 Lite](https://huggingface.co/spaces/Freepik/flux.1-lite)) is publicly available on [HuggingFace Repo](https://huggingface.co/Freepik/flux.1-lite-8B-alpha). |
|
|
|
## Citation |
|
If you find our work helpful, please cite it! |
|
|
|
```bibtex |
|
@article{flux1-lite, |
|
title={Flux.1 Lite: Distilling Flux1.dev for Efficient Text-to-Image Generation}, |
|
author={Daniel Verdú, Javier Martín}, |
|
email={dverdu@freepik.com, javier.martin@freepik.com}, |
|
year={2024}, |
|
} |
|
``` |
|
|
|
|
|
|
|
|
|
|
