Update README.md
Browse files
README.md
CHANGED
|
@@ -1,90 +1,15 @@
|
|
| 1 |
-
---
|
| 2 |
-
language:
|
| 3 |
-
- en
|
| 4 |
-
license: other
|
| 5 |
-
license_name: flux-1-dev-non-commercial-license
|
| 6 |
-
license_link: LICENSE.md
|
| 7 |
-
extra_gated_prompt: By clicking "Agree", you agree to the [FluxDev Non-Commercial License Agreement](https://huggingface.co/black-forest-labs/FLUX.1-Fill-dev/blob/main/LICENSE.md)
|
| 8 |
-
and acknowledge the [Acceptable Use Policy](https://huggingface.co/black-forest-labs/FLUX.1-Fill-dev/blob/main/POLICY.md).
|
| 9 |
-
tags:
|
| 10 |
-
- image-generation
|
| 11 |
-
- flux
|
| 12 |
-
- diffusion-single-file
|
| 13 |
-
---
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-

|
| 17 |
-
|
| 18 |
-
`FLUX.1 Fill [dev]` is a 12 billion parameter rectified flow transformer capable of filling areas in existing images based on a text description.
|
| 19 |
-
For more information, please read our [blog post](https://blackforestlabs.ai/flux-1-tools/).
|
| 20 |
-
|
| 21 |
-
# Key Features
|
| 22 |
-
1. Cutting-edge output quality, second only to our state-of-the-art model `FLUX.1 Fill [pro]`.
|
| 23 |
-
2. Blends impressive prompt following with completing the structure of your source image.
|
| 24 |
-
3. Trained using guidance distillation, making `FLUX.1 Fill [dev]` more efficient.
|
| 25 |
-
4. Open weights to drive new scientific research, and empower artists to develop innovative workflows.
|
| 26 |
-
5. Generated outputs can be used for personal, scientific, and commercial purposes as described in the [`FLUX.1 [dev]` Non-Commercial License](https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md).
|
| 27 |
-
|
| 28 |
-
# Usage
|
| 29 |
-
We provide a reference implementation of `FLUX.1 Fill [dev]`, as well as sampling code, in a dedicated [github repository](https://github.com/black-forest-labs/flux).
|
| 30 |
-
Developers and creatives looking to build on top of `FLUX.1 Fill [dev]` are encouraged to use this as a starting point.
|
| 31 |
-
|
| 32 |
-
## API Endpoints
|
| 33 |
-
The FLUX.1 models are also available in our API [bfl.ml](https://docs.bfl.ml/)
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-

|
| 37 |
-
|
| 38 |
-
## Diffusers
|
| 39 |
-
|
| 40 |
-
To use `FLUX.1 Fill [dev]` with the 🧨 diffusers python library, first install or upgrade diffusers
|
| 41 |
-
|
| 42 |
-
```shell
|
| 43 |
-
pip install -U diffusers
|
| 44 |
-
```
|
| 45 |
-
|
| 46 |
-
Then you can use `FluxFillPipeline` to run the model
|
| 47 |
-
|
| 48 |
-
```python
|
| 49 |
-
import torch
|
| 50 |
-
from diffusers import FluxFillPipeline
|
| 51 |
-
from diffusers.utils import load_image
|
| 52 |
-
|
| 53 |
-
image = load_image("https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/cup.png")
|
| 54 |
-
mask = load_image("https://huggingface.co/datasets/diffusers/diffusers-images-docs/resolve/main/cup_mask.png")
|
| 55 |
-
|
| 56 |
-
pipe = FluxFillPipeline.from_pretrained("black-forest-labs/FLUX.1-Fill-dev", torch_dtype=torch.bfloat16).to("cuda")
|
| 57 |
-
image = pipe(
|
| 58 |
-
prompt="a white paper cup",
|
| 59 |
-
image=image,
|
| 60 |
-
mask_image=mask,
|
| 61 |
-
height=1632,
|
| 62 |
-
width=1232,
|
| 63 |
-
guidance_scale=30,
|
| 64 |
-
num_inference_steps=50,
|
| 65 |
-
max_sequence_length=512,
|
| 66 |
-
generator=torch.Generator("cpu").manual_seed(0)
|
| 67 |
-
).images[0]
|
| 68 |
-
image.save(f"flux-fill-dev.png")
|
| 69 |
-
```
|
| 70 |
-
|
| 71 |
-
To learn more check out the [diffusers](https://huggingface.co/docs/diffusers/main/en/api/pipelines/flux) documentation
|
| 72 |
-
|
| 73 |
-
---
|
| 74 |
-
|
| 75 |
-
# Limitations
|
| 76 |
-
- This model is not intended or able to provide factual information.
|
| 77 |
-
- As a statistical model this checkpoint might amplify existing societal biases.
|
| 78 |
-
- The model may fail to generate output that matches the prompts.
|
| 79 |
-
- Prompt following is heavily influenced by the prompting-style.
|
| 80 |
-
- There may be slight-color shifts in areas that are not filled in
|
| 81 |
-
- Filling in complex textures may produce lines at the edges of the filled-area.
|
| 82 |
-
|
| 83 |
-
# Out-of-Scope Use
|
| 84 |
-
The model and its derivatives may not be used
|
| 85 |
-
|
| 86 |
-
- In any way that violates any applicable national, federal, state, local or international law or regulation.
|
| 87 |
-
- For the purpose of exploiting, harming or attempting to exploit or harm minors in any way; including but not limited to the solicitation, creation, acquisition, or dissemination of child exploitative content.
|
| 88 |
-
- To generate or disseminate verifiably false information and/or content with the purpose of harming others.
|
| 89 |
-
- To generate or disseminate personal identifiable information that can be used to harm an individual.
|
| 90 |
-
- To harass, abuse, threat
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
license: other
|
| 5 |
+
license_name: flux-1-dev-non-commercial-license
|
| 6 |
+
license_link: LICENSE.md
|
| 7 |
+
extra_gated_prompt: By clicking "Agree", you agree to the [FluxDev Non-Commercial License Agreement](https://huggingface.co/black-forest-labs/FLUX.1-Fill-dev/blob/main/LICENSE.md)
|
| 8 |
+
and acknowledge the [Acceptable Use Policy](https://huggingface.co/black-forest-labs/FLUX.1-Fill-dev/blob/main/POLICY.md).
|
| 9 |
+
tags:
|
| 10 |
+
- image-generation
|
| 11 |
+
- flux
|
| 12 |
+
- diffusion-single-file
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
Merge [OneReward](https://huggingface.co/bytedance-research/OneReward) into Flux.1-Dev-Fill.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|