catacolabs / sdxl-ad-inpaint

Product advertising image generator using SDXL

Cold

Public
379K runs
L40S
GitHub
License

Run with an API

Playground API Examples README Versions

Input

image

*file

Remove background from this image

product_fill

string

What percentage of the image width to fill with product

Default: "Original"

prompt

*string

Shift + Return to add a new line

a modern sofa in a contemporary living room, stylish decora modern sofa in a contemporary living room, stylish decor

Describe the new setting for your product

negative_prompt

string

Shift + Return to add a new line

Describe what you do not want in your setting

Default: "low quality, out of frame, illustration, 3d, sepia, painting, cartoons, sketch, watermark, text, Logo, advertisement"

img_size

string

Possible SDXL image sizes

Default: "1024, 1024"

scheduler

string

scheduler

Default: "K_EULER"

num_inference_steps

integer

Inference Steps

Default: 40

guidance_scale

number

Guidance Scale

Default: 7.5

condition_scale

number

(minimum: 0.3, maximum: 0.9)

controlnet conditioning scale for generalization

Default: 0.9

num_refine_steps

integer

(minimum: 0, maximum: 40)

Number of steps to refine

Default: 10

apply_img

boolean

Applies the original product image to the final result

Default: true

seed

integer

Empty or 0 for a random image

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

npx create-replicate --model=catacolabs/sdxl-ad-inpaint

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 catacolabs/sdxl-ad-inpaint using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "catacolabs/sdxl-ad-inpaint:9c0cb4c579c54432431d96c70924afcca18983de872e8a221777fb1416253359",
  {
    input: {
      seed: 24603,
      image: "https://replicate.delivery/pbxt/JX7yjB7jAgeCC1tUmWUUZlWg3IDWW9vLqIjwqWOlj9p6zyyn/sofa.png",
      prompt: "a modern sofa in a contemporary living room, stylish decor",
      img_size: "1024, 1024",
      apply_img: false,
      scheduler: "K_EULER",
      product_fill: "80",
      guidance_scale: 7.5,
      condition_scale: 0.8,
      negative_prompt: "",
      num_refine_steps: 20,
      num_inference_steps: 40
    }
  }
);

// To access the file URL:
console.log(output.url()); //=> "http://example.com"

// To write the file to disk:
fs.writeFile("my-image.png", 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 catacolabs/sdxl-ad-inpaint using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "catacolabs/sdxl-ad-inpaint:9c0cb4c579c54432431d96c70924afcca18983de872e8a221777fb1416253359",
    input={
        "seed": 24603,
        "image": "https://replicate.delivery/pbxt/JX7yjB7jAgeCC1tUmWUUZlWg3IDWW9vLqIjwqWOlj9p6zyyn/sofa.png",
        "prompt": "a modern sofa in a contemporary living room, stylish decor",
        "img_size": "1024, 1024",
        "apply_img": False,
        "scheduler": "K_EULER",
        "product_fill": "80",
        "guidance_scale": 7.5,
        "condition_scale": 0.8,
        "negative_prompt": "",
        "num_refine_steps": 20,
        "num_inference_steps": 40
    }
)
print(output)

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 catacolabs/sdxl-ad-inpaint 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": "9c0cb4c579c54432431d96c70924afcca18983de872e8a221777fb1416253359",
    "input": {
      "seed": 24603,
      "image": "https://replicate.delivery/pbxt/JX7yjB7jAgeCC1tUmWUUZlWg3IDWW9vLqIjwqWOlj9p6zyyn/sofa.png",
      "prompt": "a modern sofa in a contemporary living room, stylish decor",
      "img_size": "1024, 1024",
      "apply_img": false,
      "scheduler": "K_EULER",
      "product_fill": "80",
      "guidance_scale": 7.5,
      "condition_scale": 0.8,
      "negative_prompt": "",
      "num_refine_steps": 20,
      "num_inference_steps": 40
    }
  }' \
  https://api.replicate.com/v1/predictions

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

Output

{
  "completed_at": "2023-09-15T18:43:09.006396Z",
  "created_at": "2023-09-15T18:42:51.245376Z",
  "data_removed": false,
  "error": null,
  "id": "vhroqttby6kocim755245syvxi",
  "input": {
    "seed": 24603,
    "image": "https://replicate.delivery/pbxt/JX7yjB7jAgeCC1tUmWUUZlWg3IDWW9vLqIjwqWOlj9p6zyyn/sofa.png",
    "prompt": "a modern sofa in a contemporary living room, stylish decor",
    "img_size": "1024, 1024",
    "apply_img": false,
    "scheduler": "K_EULER",
    "product_fill": "80",
    "guidance_scale": 7.5,
    "condition_scale": 0.8,
    "negative_prompt": "",
    "num_refine_steps": 20,
    "num_inference_steps": 40
  },
  "logs": "Using seed: 24603\nProduct img W:666, H:375\nScale factor: 0.8\nFinal img W: 1024, H:1024\n  0%|          | 0/40 [00:00<?, ?it/s]\n  2%|▎         | 1/40 [00:00<00:11,  3.38it/s]\n  5%|▌         | 2/40 [00:00<00:11,  3.38it/s]\n  8%|▊         | 3/40 [00:00<00:10,  3.38it/s]\n 10%|█         | 4/40 [00:01<00:10,  3.37it/s]\n 12%|█▎        | 5/40 [00:01<00:10,  3.37it/s]\n 15%|█▌        | 6/40 [00:01<00:10,  3.37it/s]\n 18%|█▊        | 7/40 [00:02<00:09,  3.37it/s]\n 20%|██        | 8/40 [00:02<00:09,  3.37it/s]\n 22%|██▎       | 9/40 [00:02<00:09,  3.37it/s]\n 25%|██▌       | 10/40 [00:02<00:08,  3.36it/s]\n 28%|██▊       | 11/40 [00:03<00:08,  3.36it/s]\n 30%|███       | 12/40 [00:03<00:08,  3.36it/s]\n 32%|███▎      | 13/40 [00:03<00:08,  3.36it/s]\n 35%|███▌      | 14/40 [00:04<00:07,  3.36it/s]\n 38%|███▊      | 15/40 [00:04<00:07,  3.32it/s]\n 40%|████      | 16/40 [00:04<00:07,  3.32it/s]\n 42%|████▎     | 17/40 [00:05<00:06,  3.33it/s]\n 45%|████▌     | 18/40 [00:05<00:06,  3.33it/s]\n 48%|████▊     | 19/40 [00:05<00:06,  3.34it/s]\n 50%|█████     | 20/40 [00:05<00:05,  3.34it/s]\n 52%|█████▎    | 21/40 [00:06<00:05,  3.34it/s]\n 55%|█████▌    | 22/40 [00:06<00:05,  3.35it/s]\n 57%|█████▊    | 23/40 [00:06<00:05,  3.34it/s]\n 60%|██████    | 24/40 [00:07<00:04,  3.34it/s]\n 62%|██████▎   | 25/40 [00:07<00:04,  3.34it/s]\n 65%|██████▌   | 26/40 [00:07<00:04,  3.34it/s]\n 68%|██████▊   | 27/40 [00:08<00:03,  3.34it/s]\n 70%|███████   | 28/40 [00:08<00:03,  3.34it/s]\n 72%|███████▎  | 29/40 [00:08<00:03,  3.34it/s]\n 75%|███████▌  | 30/40 [00:08<00:02,  3.34it/s]\n 78%|███████▊  | 31/40 [00:09<00:02,  3.35it/s]\n 80%|████████  | 32/40 [00:09<00:02,  3.34it/s]\n 82%|████████▎ | 33/40 [00:09<00:02,  3.34it/s]\n 85%|████████▌ | 34/40 [00:10<00:01,  3.34it/s]\n 88%|████████▊ | 35/40 [00:10<00:01,  3.34it/s]\n 90%|█████████ | 36/40 [00:10<00:01,  3.34it/s]\n 92%|█████████▎| 37/40 [00:11<00:00,  3.34it/s]\n 95%|█████████▌| 38/40 [00:11<00:00,  3.34it/s]\n 98%|█████████▊| 39/40 [00:11<00:00,  3.34it/s]\n100%|██████████| 40/40 [00:11<00:00,  3.34it/s]\n100%|██████████| 40/40 [00:11<00:00,  3.35it/s]\n  0%|          | 0/6 [00:00<?, ?it/s]\n 17%|█▋        | 1/6 [00:00<00:01,  4.31it/s]\n 33%|███▎      | 2/6 [00:00<00:00,  4.30it/s]\n 50%|█████     | 3/6 [00:00<00:00,  4.30it/s]\n 67%|██████▋   | 4/6 [00:00<00:00,  4.28it/s]\n 83%|████████▎ | 5/6 [00:01<00:00,  4.29it/s]\n100%|██████████| 6/6 [00:01<00:00,  4.28it/s]\n100%|██████████| 6/6 [00:01<00:00,  4.29it/s]",
  "metrics": {
    "predict_time": 17.757526,
    "total_time": 17.76102
  },
  "output": "https://pbxt.replicate.delivery/ORbuWtoy0y6NI9f4DrJ2fxs92LgviBaOlzOVdYTr3pT8eKJjA/7-out.png",
  "started_at": "2023-09-15T18:42:51.248870Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/vhroqttby6kocim755245syvxi",
    "cancel": "https://api.replicate.com/v1/predictions/vhroqttby6kocim755245syvxi/cancel"
  },
  "version": "9c0cb4c579c54432431d96c70924afcca18983de872e8a221777fb1416253359"
}

Generated in

17.8 seconds

Report

Using seed: 24603
Product img W:666, H:375
Scale factor: 0.8
Final img W: 1024, H:1024
  0%|          | 0/40 [00:00<?, ?it/s]
  2%|▎         | 1/40 [00:00<00:11,  3.38it/s]
  5%|▌         | 2/40 [00:00<00:11,  3.38it/s]
  8%|▊         | 3/40 [00:00<00:10,  3.38it/s]
 10%|█         | 4/40 [00:01<00:10,  3.37it/s]
 12%|█▎        | 5/40 [00:01<00:10,  3.37it/s]
 15%|█▌        | 6/40 [00:01<00:10,  3.37it/s]
 18%|█▊        | 7/40 [00:02<00:09,  3.37it/s]
 20%|██        | 8/40 [00:02<00:09,  3.37it/s]
 22%|██▎       | 9/40 [00:02<00:09,  3.37it/s]
 25%|██▌       | 10/40 [00:02<00:08,  3.36it/s]
 28%|██▊       | 11/40 [00:03<00:08,  3.36it/s]
 30%|███       | 12/40 [00:03<00:08,  3.36it/s]
 32%|███▎      | 13/40 [00:03<00:08,  3.36it/s]
 35%|███▌      | 14/40 [00:04<00:07,  3.36it/s]
 38%|███▊      | 15/40 [00:04<00:07,  3.32it/s]
 40%|████      | 16/40 [00:04<00:07,  3.32it/s]
 42%|████▎     | 17/40 [00:05<00:06,  3.33it/s]
 45%|████▌     | 18/40 [00:05<00:06,  3.33it/s]
 48%|████▊     | 19/40 [00:05<00:06,  3.34it/s]
 50%|█████     | 20/40 [00:05<00:05,  3.34it/s]
 52%|█████▎    | 21/40 [00:06<00:05,  3.34it/s]
 55%|█████▌    | 22/40 [00:06<00:05,  3.35it/s]
 57%|█████▊    | 23/40 [00:06<00:05,  3.34it/s]
 60%|██████    | 24/40 [00:07<00:04,  3.34it/s]
 62%|██████▎   | 25/40 [00:07<00:04,  3.34it/s]
 65%|██████▌   | 26/40 [00:07<00:04,  3.34it/s]
 68%|██████▊   | 27/40 [00:08<00:03,  3.34it/s]
 70%|███████   | 28/40 [00:08<00:03,  3.34it/s]
 72%|███████▎  | 29/40 [00:08<00:03,  3.34it/s]
 75%|███████▌  | 30/40 [00:08<00:02,  3.34it/s]
 78%|███████▊  | 31/40 [00:09<00:02,  3.35it/s]
 80%|████████  | 32/40 [00:09<00:02,  3.34it/s]
 82%|████████▎ | 33/40 [00:09<00:02,  3.34it/s]
 85%|████████▌ | 34/40 [00:10<00:01,  3.34it/s]
 88%|████████▊ | 35/40 [00:10<00:01,  3.34it/s]
 90%|█████████ | 36/40 [00:10<00:01,  3.34it/s]
 92%|█████████▎| 37/40 [00:11<00:00,  3.34it/s]
 95%|█████████▌| 38/40 [00:11<00:00,  3.34it/s]
 98%|█████████▊| 39/40 [00:11<00:00,  3.34it/s]
100%|██████████| 40/40 [00:11<00:00,  3.34it/s]
100%|██████████| 40/40 [00:11<00:00,  3.35it/s]
  0%|          | 0/6 [00:00<?, ?it/s]
 17%|█▋        | 1/6 [00:00<00:01,  4.31it/s]
 33%|███▎      | 2/6 [00:00<00:00,  4.30it/s]
 50%|█████     | 3/6 [00:00<00:00,  4.30it/s]
 67%|██████▋   | 4/6 [00:00<00:00,  4.28it/s]
 83%|████████▎ | 5/6 [00:01<00:00,  4.29it/s]
100%|██████████| 6/6 [00:01<00:00,  4.28it/s]
100%|██████████| 6/6 [00:01<00:00,  4.29it/s]

Examples

View more examples

Run time and cost

This model costs approximately $0.032 to run on Replicate, or 31 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 L40S GPU hardware. Predictions typically complete within 33 seconds. The predict time for this model varies significantly based on the inputs.

Readme

About

This is a custom model, and is my attempt at tackling the challenge of: creating an Ad Image generator

How It Works

Based on SDXL and ControlNet to create a background setting for your product image

If your input image does not have the product images centered, turn off “apply_img”

Support Me

Follow my work on Twitter! @lucataco93