zsxkib / flux-dev-inpainting

🎨 Fill in masked parts of images with FLUX.1-dev 🖌️

Cold

Public
347.3K runs
A100 (80GB)
GitHub
License

Run with an API

Playground API Examples README Versions

Input

image

*file

Preview

Input image for inpainting

mask

*file

Preview

Mask image

prompt

*string

Shift + Return to add a new line

Text prompt for inpainting

strength

number

(minimum: 0, maximum: 1)

Strength of inpainting. Higher values allow for more deviation from the original image.

Default: 0.85

num_inference_steps

integer

(minimum: 1, maximum: 50)

Number of denoising steps. More steps usually lead to a higher quality image at the expense of slower inference.

Default: 30

guidance_scale

number

(minimum: 1, maximum: 20)

Guidance scale as defined in Classifier-Free Diffusion Guidance. Higher guidance scale encourages images that are closely linked to the text prompt, usually at the expense of lower image quality.

Default: 7

height

integer

(minimum: 128, maximum: 2048)

Height of the output image. Will be rounded to the nearest multiple of 8.

Default: 1024

width

integer

(minimum: 128, maximum: 2048)

Width of the output image. Will be rounded to the nearest multiple of 8.

Default: 1024

num_outputs

integer

(minimum: 1, maximum: 8)

Number of images to generate per prompt.

Default: 1

seed

integer

Random seed. Leave blank to randomize the seed

output_format

string

Format of the output image

Default: "webp"

output_quality

integer

(minimum: 0, maximum: 100)

Quality of the output image, from 0 to 100. 100 is best quality, 0 is lowest quality.

Default: 80

Run this model in Node.js with one line of code:

npx create-replicate --model=zsxkib/flux-dev-inpainting

or set up a project from scratch

Install Replicate’s Node.js client library:

npm install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import and set up the client:

import Replicate from "replicate";

const replicate = new Replicate({
  auth: process.env.REPLICATE_API_TOKEN,
});

Run zsxkib/flux-dev-inpainting using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "zsxkib/flux-dev-inpainting:ca8350ff748d56b3ebbd5a12bd3436c2214262a4ff8619de9890ecc41751a008",
  {
    input: {
      mask: "https://replicate.delivery/pbxt/HtGQBqO9MtVbPm0G0K43nsvvjBB0E0PaWOhuNRrRBBT4ttbf/mask.png",
      image: "https://replicate.delivery/pbxt/HtGQBfA5TrqFYZBf0UL18NTqHrzt8UiSIsAkUuMHtjvFDO6p/overture-creations-5sI6fQgYIuo.png",
      width: 1024,
      height: 1024,
      prompt: "small cute cat sat on a park bench",
      strength: 1,
      num_outputs: 1,
      output_format: "webp",
      guidance_scale: 7,
      output_quality: 90,
      num_inference_steps: 30
    }
  }
);
console.log(output);

To learn more, take a look at the guide on getting started with Node.js.

Install Replicate’s Python client library:

pip install replicate

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Import the client:

import replicate

Run zsxkib/flux-dev-inpainting using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "zsxkib/flux-dev-inpainting:ca8350ff748d56b3ebbd5a12bd3436c2214262a4ff8619de9890ecc41751a008",
    input={
        "mask": "https://replicate.delivery/pbxt/HtGQBqO9MtVbPm0G0K43nsvvjBB0E0PaWOhuNRrRBBT4ttbf/mask.png",
        "image": "https://replicate.delivery/pbxt/HtGQBfA5TrqFYZBf0UL18NTqHrzt8UiSIsAkUuMHtjvFDO6p/overture-creations-5sI6fQgYIuo.png",
        "width": 1024,
        "height": 1024,
        "prompt": "small cute cat sat on a park bench",
        "strength": 1,
        "num_outputs": 1,
        "output_format": "webp",
        "guidance_scale": 7,
        "output_quality": 90,
        "num_inference_steps": 30
    }
)

# The zsxkib/flux-dev-inpainting model can stream output as it's running.
# The predict method returns an iterator, and you can iterate over that output.
for item in output:
    # https://replicate.com/zsxkib/flux-dev-inpainting/api#output-schema
    print(item)

To learn more, take a look at the guide on getting started with Python.

Set the REPLICATE_API_TOKEN environment variable:

export REPLICATE_API_TOKEN=<paste-your-token-here>

Find your API token in your account settings.

Run zsxkib/flux-dev-inpainting using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

curl -s -X POST \
  -H "Authorization: Bearer $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -H "Prefer: wait" \
  -d $'{
    "version": "ca8350ff748d56b3ebbd5a12bd3436c2214262a4ff8619de9890ecc41751a008",
    "input": {
      "mask": "https://replicate.delivery/pbxt/HtGQBqO9MtVbPm0G0K43nsvvjBB0E0PaWOhuNRrRBBT4ttbf/mask.png",
      "image": "https://replicate.delivery/pbxt/HtGQBfA5TrqFYZBf0UL18NTqHrzt8UiSIsAkUuMHtjvFDO6p/overture-creations-5sI6fQgYIuo.png",
      "width": 1024,
      "height": 1024,
      "prompt": "small cute cat sat on a park bench",
      "strength": 1,
      "num_outputs": 1,
      "output_format": "webp",
      "guidance_scale": 7,
      "output_quality": 90,
      "num_inference_steps": 30
    }
  }' \
  https://api.replicate.com/v1/predictions

To learn more, take a look at Replicate’s HTTP API reference docs.

Output

{
  "completed_at": "2024-08-16T13:11:50.808526Z",
  "created_at": "2024-08-16T13:11:34.001000Z",
  "data_removed": false,
  "error": null,
  "id": "kxt9x4wqe5rj20chbd88bnsd9m",
  "input": {
    "mask": "https://replicate.delivery/pbxt/HtGQBqO9MtVbPm0G0K43nsvvjBB0E0PaWOhuNRrRBBT4ttbf/mask.png",
    "image": "https://replicate.delivery/pbxt/HtGQBfA5TrqFYZBf0UL18NTqHrzt8UiSIsAkUuMHtjvFDO6p/overture-creations-5sI6fQgYIuo.png",
    "prompt": "small cute cat sat on a park bench",
    "strength": 1,
    "output_format": "webp",
    "output_quality": 90,
    "num_inference_steps": 30
  },
  "logs": "Using seed: 632880580\n  0%|          | 0/34 [00:00<?, ?it/s]\n  3%|▎         | 1/34 [00:00<00:18,  1.79it/s]\n  6%|▌         | 2/34 [00:01<00:16,  1.99it/s]\n  9%|▉         | 3/34 [00:01<00:14,  2.07it/s]\n 12%|█▏        | 4/34 [00:01<00:14,  2.11it/s]\n 15%|█▍        | 5/34 [00:02<00:13,  2.13it/s]\n 18%|█▊        | 6/34 [00:02<00:13,  2.14it/s]\n 21%|██        | 7/34 [00:03<00:12,  2.15it/s]\n 24%|██▎       | 8/34 [00:03<00:12,  2.15it/s]\n 26%|██▋       | 9/34 [00:04<00:11,  2.16it/s]\n 29%|██▉       | 10/34 [00:04<00:11,  2.16it/s]\n 32%|███▏      | 11/34 [00:05<00:10,  2.16it/s]\n 35%|███▌      | 12/34 [00:05<00:10,  2.16it/s]\n 38%|███▊      | 13/34 [00:06<00:09,  2.16it/s]\n 41%|████      | 14/34 [00:06<00:09,  2.16it/s]\n 44%|████▍     | 15/34 [00:07<00:08,  2.16it/s]\n 47%|████▋     | 16/34 [00:07<00:08,  2.16it/s]\n 50%|█████     | 17/34 [00:07<00:07,  2.16it/s]\n 53%|█████▎    | 18/34 [00:08<00:07,  2.16it/s]\n 56%|█████▌    | 19/34 [00:08<00:06,  2.16it/s]\n 59%|█████▉    | 20/34 [00:09<00:06,  2.16it/s]\n 62%|██████▏   | 21/34 [00:09<00:06,  2.16it/s]\n 65%|██████▍   | 22/34 [00:10<00:05,  2.16it/s]\n 68%|██████▊   | 23/34 [00:10<00:05,  2.16it/s]\n 71%|███████   | 24/34 [00:11<00:04,  2.16it/s]\n 74%|███████▎  | 25/34 [00:11<00:04,  2.16it/s]\n 76%|███████▋  | 26/34 [00:12<00:03,  2.16it/s]\n 79%|███████▉  | 27/34 [00:12<00:03,  2.16it/s]\n 82%|████████▏ | 28/34 [00:13<00:02,  2.16it/s]\n 85%|████████▌ | 29/34 [00:13<00:02,  2.16it/s]\n 88%|████████▊ | 30/34 [00:13<00:01,  2.16it/s]\n 91%|█████████ | 31/34 [00:14<00:01,  2.16it/s]\n 94%|█████████▍| 32/34 [00:14<00:00,  2.16it/s]\n 97%|█████████▋| 33/34 [00:15<00:00,  2.16it/s]\n100%|██████████| 34/34 [00:15<00:00,  2.45it/s]\n100%|██████████| 34/34 [00:15<00:00,  2.17it/s]\nINFERENCE DONE\n[~] Saving to /tmp/output.webp...\n[~] Output format: WEBP\n[~] Output quality: 90",
  "metrics": {
    "predict_time": 16.799082725,
    "total_time": 16.807526
  },
  "output": "https://replicate.delivery/yhqm/M3kYRwlZQ0YiG1PJaeIsDEnsO1JJozDQVbrVqMNjqETLEopJA/output.webp",
  "started_at": "2024-08-16T13:11:34.009444Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/kxt9x4wqe5rj20chbd88bnsd9m",
    "cancel": "https://api.replicate.com/v1/predictions/kxt9x4wqe5rj20chbd88bnsd9m/cancel"
  },
  "version": "11cca3274341de7aef06f04e4dab3d651ea8ac04eff003f23603d4fdf5b56ff0"
}

Generated in

16.8 seconds

Tweak it ShareReport

Using seed: 632880580
  0%|          | 0/34 [00:00<?, ?it/s]
  3%|▎         | 1/34 [00:00<00:18,  1.79it/s]
  6%|▌         | 2/34 [00:01<00:16,  1.99it/s]
  9%|▉         | 3/34 [00:01<00:14,  2.07it/s]
 12%|█▏        | 4/34 [00:01<00:14,  2.11it/s]
 15%|█▍        | 5/34 [00:02<00:13,  2.13it/s]
 18%|█▊        | 6/34 [00:02<00:13,  2.14it/s]
 21%|██        | 7/34 [00:03<00:12,  2.15it/s]
 24%|██▎       | 8/34 [00:03<00:12,  2.15it/s]
 26%|██▋       | 9/34 [00:04<00:11,  2.16it/s]
 29%|██▉       | 10/34 [00:04<00:11,  2.16it/s]
 32%|███▏      | 11/34 [00:05<00:10,  2.16it/s]
 35%|███▌      | 12/34 [00:05<00:10,  2.16it/s]
 38%|███▊      | 13/34 [00:06<00:09,  2.16it/s]
 41%|████      | 14/34 [00:06<00:09,  2.16it/s]
 44%|████▍     | 15/34 [00:07<00:08,  2.16it/s]
 47%|████▋     | 16/34 [00:07<00:08,  2.16it/s]
 50%|█████     | 17/34 [00:07<00:07,  2.16it/s]
 53%|█████▎    | 18/34 [00:08<00:07,  2.16it/s]
 56%|█████▌    | 19/34 [00:08<00:06,  2.16it/s]
 59%|█████▉    | 20/34 [00:09<00:06,  2.16it/s]
 62%|██████▏   | 21/34 [00:09<00:06,  2.16it/s]
 65%|██████▍   | 22/34 [00:10<00:05,  2.16it/s]
 68%|██████▊   | 23/34 [00:10<00:05,  2.16it/s]
 71%|███████   | 24/34 [00:11<00:04,  2.16it/s]
 74%|███████▎  | 25/34 [00:11<00:04,  2.16it/s]
 76%|███████▋  | 26/34 [00:12<00:03,  2.16it/s]
 79%|███████▉  | 27/34 [00:12<00:03,  2.16it/s]
 82%|████████▏ | 28/34 [00:13<00:02,  2.16it/s]
 85%|████████▌ | 29/34 [00:13<00:02,  2.16it/s]
 88%|████████▊ | 30/34 [00:13<00:01,  2.16it/s]
 91%|█████████ | 31/34 [00:14<00:01,  2.16it/s]
 94%|█████████▍| 32/34 [00:14<00:00,  2.16it/s]
 97%|█████████▋| 33/34 [00:15<00:00,  2.16it/s]
100%|██████████| 34/34 [00:15<00:00,  2.45it/s]
100%|██████████| 34/34 [00:15<00:00,  2.17it/s]
INFERENCE DONE
[~] Saving to /tmp/output.webp...
[~] Output format: WEBP
[~] Output quality: 90

This example was created by a different version, zsxkib/flux-dev-inpainting:11cca327.

Examples

View more examples

Run time and cost

This model costs approximately $0.040 to run on Replicate, or 25 runs per $1, but this varies depending on your inputs. It is also open source and you can run it on your own computer with Docker.

This model runs on Nvidia A100 (80GB) GPU hardware. Predictions typically complete within 29 seconds. The predict time for this model varies significantly based on the inputs.

Readme

FLUX.1 DEV Inpainting Model

About

This is a version of the Flux DEV inpainting model by @skalskip92. We’ve changed it to keep the original image’s shape. The model expects the mask to be the same size as the input image, but you can change this with some settings.

Big thanks to @Gothos13 for helping create this clever inpainting method.

Tips for Use

For better results, try using more steps (between 20-30) when creating the image.
Play around with the inpainting ‘strength’. Numbers between 0.85-1.0 often work well, but you might need different strengths for different prompts.
Keep in mind that Flux wasn’t specially trained for inpainting. There are no specific inpainting parts, which makes this a smart way to use the model’s skills.
The model can still handle text in images.

Note

This inpainting method can make great images, but you might need to try a few times to get what you want. Don’t give up if your first try doesn’t look perfect!

For more info and updates, check out the original tweet thread.

Support

If you like my work, please follow me! @zsakib_