ali-vilab / anydoor

Anydoor: zero-shot object-level image customization

Cold

Public
2K runs
L40S
GitHub
Paper
License

Run with an API

Playground API Examples README Versions

Input

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

npx create-replicate --model=ali-vilab/anydoor

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 ali-vilab/anydoor using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "ali-vilab/anydoor:542c963129c4661ab53a875b1b9a84b2102ca784cf872ef2752a468721c0eb2a",
  {
    input: {
      steps: 30,
      bg_mask_path: "https://replicate.delivery/pbxt/KAr5sZjqxgOrXGCOo2PAkvrywmnD8s6gtGYEKAzxV1n8UsKN/burger-mask.png",
      bg_image_path: "https://replicate.delivery/pbxt/KAr5t2fIFVQ4eNy4JwN6LAZpHyTYzXHaO78TinDeHzYWK9RS/burger.png",
      guidance_scale: 4.5,
      control_strength: 1,
      enable_shape_control: false,
      reference_image_mask: "https://replicate.delivery/pbxt/KAr5swy9Bdkv4cM2IKAvLoxCMWtLLmMLFSK1SpwTLDxzdYgW/sloth-mask.png",
      reference_image_path: "https://replicate.delivery/pbxt/KAr5sJGlXZyZPR2SfOduHMvnUgS1CIc81ynRVCdHV24JdTo0/sloth.png"
    }
  }
);
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 ali-vilab/anydoor using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "ali-vilab/anydoor:542c963129c4661ab53a875b1b9a84b2102ca784cf872ef2752a468721c0eb2a",
    input={
        "steps": 30,
        "bg_mask_path": "https://replicate.delivery/pbxt/KAr5sZjqxgOrXGCOo2PAkvrywmnD8s6gtGYEKAzxV1n8UsKN/burger-mask.png",
        "bg_image_path": "https://replicate.delivery/pbxt/KAr5t2fIFVQ4eNy4JwN6LAZpHyTYzXHaO78TinDeHzYWK9RS/burger.png",
        "guidance_scale": 4.5,
        "control_strength": 1,
        "enable_shape_control": False,
        "reference_image_mask": "https://replicate.delivery/pbxt/KAr5swy9Bdkv4cM2IKAvLoxCMWtLLmMLFSK1SpwTLDxzdYgW/sloth-mask.png",
        "reference_image_path": "https://replicate.delivery/pbxt/KAr5sJGlXZyZPR2SfOduHMvnUgS1CIc81ynRVCdHV24JdTo0/sloth.png"
    }
)
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 ali-vilab/anydoor 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": "542c963129c4661ab53a875b1b9a84b2102ca784cf872ef2752a468721c0eb2a",
    "input": {
      "steps": 30,
      "bg_mask_path": "https://replicate.delivery/pbxt/KAr5sZjqxgOrXGCOo2PAkvrywmnD8s6gtGYEKAzxV1n8UsKN/burger-mask.png",
      "bg_image_path": "https://replicate.delivery/pbxt/KAr5t2fIFVQ4eNy4JwN6LAZpHyTYzXHaO78TinDeHzYWK9RS/burger.png",
      "guidance_scale": 4.5,
      "control_strength": 1,
      "enable_shape_control": false,
      "reference_image_mask": "https://replicate.delivery/pbxt/KAr5swy9Bdkv4cM2IKAvLoxCMWtLLmMLFSK1SpwTLDxzdYgW/sloth-mask.png",
      "reference_image_path": "https://replicate.delivery/pbxt/KAr5sJGlXZyZPR2SfOduHMvnUgS1CIc81ynRVCdHV24JdTo0/sloth.png"
    }
  }' \
  https://api.replicate.com/v1/predictions

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

You can run this model locally using Cog. First, install Cog:

brew install cog

If you don’t have Homebrew, there are other installation options available.

Run this to download the model and run it in your local environment:

cog predict r8.im/ali-vilab/anydoor@sha256:542c963129c4661ab53a875b1b9a84b2102ca784cf872ef2752a468721c0eb2a \
  -i 'steps=30' \
  -i 'bg_mask_path="https://replicate.delivery/pbxt/KAr5sZjqxgOrXGCOo2PAkvrywmnD8s6gtGYEKAzxV1n8UsKN/burger-mask.png"' \
  -i 'bg_image_path="https://replicate.delivery/pbxt/KAr5t2fIFVQ4eNy4JwN6LAZpHyTYzXHaO78TinDeHzYWK9RS/burger.png"' \
  -i 'guidance_scale=4.5' \
  -i 'control_strength=1' \
  -i 'enable_shape_control=false' \
  -i 'reference_image_mask="https://replicate.delivery/pbxt/KAr5swy9Bdkv4cM2IKAvLoxCMWtLLmMLFSK1SpwTLDxzdYgW/sloth-mask.png"' \
  -i 'reference_image_path="https://replicate.delivery/pbxt/KAr5sJGlXZyZPR2SfOduHMvnUgS1CIc81ynRVCdHV24JdTo0/sloth.png"'

To learn more, take a look at the Cog documentation.

Run this to download the model and run it in your local environment:

docker run -d -p 5000:5000 --gpus=all r8.im/ali-vilab/anydoor@sha256:542c963129c4661ab53a875b1b9a84b2102ca784cf872ef2752a468721c0eb2a
curl -s -X POST \
  -H "Content-Type: application/json" \
  -d $'{
    "input": {
      "steps": 30,
      "bg_mask_path": "https://replicate.delivery/pbxt/KAr5sZjqxgOrXGCOo2PAkvrywmnD8s6gtGYEKAzxV1n8UsKN/burger-mask.png",
      "bg_image_path": "https://replicate.delivery/pbxt/KAr5t2fIFVQ4eNy4JwN6LAZpHyTYzXHaO78TinDeHzYWK9RS/burger.png",
      "guidance_scale": 4.5,
      "control_strength": 1,
      "enable_shape_control": false,
      "reference_image_mask": "https://replicate.delivery/pbxt/KAr5swy9Bdkv4cM2IKAvLoxCMWtLLmMLFSK1SpwTLDxzdYgW/sloth-mask.png",
      "reference_image_path": "https://replicate.delivery/pbxt/KAr5sJGlXZyZPR2SfOduHMvnUgS1CIc81ynRVCdHV24JdTo0/sloth.png"
    }
  }' \
  http://localhost:5000/predictions

To learn more, take a look at the Cog documentation.

Output

{
  "completed_at": "2024-01-05T16:14:05.078761Z",
  "created_at": "2024-01-05T16:13:57.255429Z",
  "data_removed": false,
  "error": null,
  "id": "vtxdxidb4u2u64w7ffou7tdvgy",
  "input": {
    "steps": 30,
    "bg_mask_path": "https://replicate.delivery/pbxt/KAr5sZjqxgOrXGCOo2PAkvrywmnD8s6gtGYEKAzxV1n8UsKN/burger-mask.png",
    "bg_image_path": "https://replicate.delivery/pbxt/KAr5t2fIFVQ4eNy4JwN6LAZpHyTYzXHaO78TinDeHzYWK9RS/burger.png",
    "guidance_scale": 4.5,
    "control_strength": 1,
    "enable_shape_control": false,
    "reference_image_mask": "https://replicate.delivery/pbxt/KAr5swy9Bdkv4cM2IKAvLoxCMWtLLmMLFSK1SpwTLDxzdYgW/sloth-mask.png",
    "reference_image_path": "https://replicate.delivery/pbxt/KAr5sJGlXZyZPR2SfOduHMvnUgS1CIc81ynRVCdHV24JdTo0/sloth.png"
  },
  "logs": "Using seed: 2627894035\nData shape for DDIM sampling is (1, 4, 64, 64), eta 0.0\nRunning DDIM Sampling with 31 timesteps\nDDIM Sampler:   0%|          | 0/31 [00:00<?, ?it/s]\nDDIM Sampler:   3%|▎         | 1/31 [00:00<00:04,  6.18it/s]\nDDIM Sampler:   6%|▋         | 2/31 [00:00<00:04,  6.25it/s]\nDDIM Sampler:  10%|▉         | 3/31 [00:00<00:04,  6.28it/s]\nDDIM Sampler:  13%|█▎        | 4/31 [00:00<00:04,  6.29it/s]\nDDIM Sampler:  16%|█▌        | 5/31 [00:00<00:04,  6.29it/s]\nDDIM Sampler:  19%|█▉        | 6/31 [00:00<00:03,  6.28it/s]\nDDIM Sampler:  23%|██▎       | 7/31 [00:01<00:03,  6.28it/s]\nDDIM Sampler:  26%|██▌       | 8/31 [00:01<00:03,  6.28it/s]\nDDIM Sampler:  29%|██▉       | 9/31 [00:01<00:03,  6.28it/s]\nDDIM Sampler:  32%|███▏      | 10/31 [00:01<00:03,  6.27it/s]\nDDIM Sampler:  35%|███▌      | 11/31 [00:01<00:03,  6.27it/s]\nDDIM Sampler:  39%|███▊      | 12/31 [00:01<00:03,  6.28it/s]\nDDIM Sampler:  42%|████▏     | 13/31 [00:02<00:02,  6.28it/s]\nDDIM Sampler:  45%|████▌     | 14/31 [00:02<00:02,  6.29it/s]\nDDIM Sampler:  48%|████▊     | 15/31 [00:02<00:02,  6.31it/s]\nDDIM Sampler:  52%|█████▏    | 16/31 [00:02<00:02,  6.32it/s]\nDDIM Sampler:  55%|█████▍    | 17/31 [00:02<00:02,  6.33it/s]\nDDIM Sampler:  58%|█████▊    | 18/31 [00:02<00:02,  6.33it/s]\nDDIM Sampler:  61%|██████▏   | 19/31 [00:03<00:01,  6.33it/s]\nDDIM Sampler:  65%|██████▍   | 20/31 [00:03<00:01,  6.33it/s]\nDDIM Sampler:  68%|██████▊   | 21/31 [00:03<00:01,  6.33it/s]\nDDIM Sampler:  71%|███████   | 22/31 [00:03<00:01,  6.33it/s]\nDDIM Sampler:  74%|███████▍  | 23/31 [00:03<00:01,  6.33it/s]\nDDIM Sampler:  77%|███████▋  | 24/31 [00:03<00:01,  6.33it/s]\nDDIM Sampler:  81%|████████  | 25/31 [00:03<00:00,  6.33it/s]\nDDIM Sampler:  84%|████████▍ | 26/31 [00:04<00:00,  6.33it/s]\nDDIM Sampler:  87%|████████▋ | 27/31 [00:04<00:00,  6.33it/s]\nDDIM Sampler:  90%|█████████ | 28/31 [00:04<00:00,  6.33it/s]\nDDIM Sampler:  94%|█████████▎| 29/31 [00:04<00:00,  6.33it/s]\nDDIM Sampler:  97%|█████████▋| 30/31 [00:04<00:00,  6.32it/s]\nDDIM Sampler: 100%|██████████| 31/31 [00:04<00:00,  6.32it/s]\nDDIM Sampler: 100%|██████████| 31/31 [00:04<00:00,  6.31it/s]",
  "metrics": {
    "predict_time": 7.787525,
    "total_time": 7.823332
  },
  "output": "https://replicate.delivery/pbxt/1yuh0r2VbN6mK50eijySUVLRGrFqtvpBb6ICrVSdcBam5uEJA/output.png",
  "started_at": "2024-01-05T16:13:57.291236Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/vtxdxidb4u2u64w7ffou7tdvgy",
    "cancel": "https://api.replicate.com/v1/predictions/vtxdxidb4u2u64w7ffou7tdvgy/cancel"
  },
  "version": "542c963129c4661ab53a875b1b9a84b2102ca784cf872ef2752a468721c0eb2a"
}

Generated in

7.8 seconds

Tweak itReport

Using seed: 2627894035
Data shape for DDIM sampling is (1, 4, 64, 64), eta 0.0
Running DDIM Sampling with 31 timesteps
DDIM Sampler:   0%|          | 0/31 [00:00<?, ?it/s]
DDIM Sampler:   3%|▎         | 1/31 [00:00<00:04,  6.18it/s]
DDIM Sampler:   6%|▋         | 2/31 [00:00<00:04,  6.25it/s]
DDIM Sampler:  10%|▉         | 3/31 [00:00<00:04,  6.28it/s]
DDIM Sampler:  13%|█▎        | 4/31 [00:00<00:04,  6.29it/s]
DDIM Sampler:  16%|█▌        | 5/31 [00:00<00:04,  6.29it/s]
DDIM Sampler:  19%|█▉        | 6/31 [00:00<00:03,  6.28it/s]
DDIM Sampler:  23%|██▎       | 7/31 [00:01<00:03,  6.28it/s]
DDIM Sampler:  26%|██▌       | 8/31 [00:01<00:03,  6.28it/s]
DDIM Sampler:  29%|██▉       | 9/31 [00:01<00:03,  6.28it/s]
DDIM Sampler:  32%|███▏      | 10/31 [00:01<00:03,  6.27it/s]
DDIM Sampler:  35%|███▌      | 11/31 [00:01<00:03,  6.27it/s]
DDIM Sampler:  39%|███▊      | 12/31 [00:01<00:03,  6.28it/s]
DDIM Sampler:  42%|████▏     | 13/31 [00:02<00:02,  6.28it/s]
DDIM Sampler:  45%|████▌     | 14/31 [00:02<00:02,  6.29it/s]
DDIM Sampler:  48%|████▊     | 15/31 [00:02<00:02,  6.31it/s]
DDIM Sampler:  52%|█████▏    | 16/31 [00:02<00:02,  6.32it/s]
DDIM Sampler:  55%|█████▍    | 17/31 [00:02<00:02,  6.33it/s]
DDIM Sampler:  58%|█████▊    | 18/31 [00:02<00:02,  6.33it/s]
DDIM Sampler:  61%|██████▏   | 19/31 [00:03<00:01,  6.33it/s]
DDIM Sampler:  65%|██████▍   | 20/31 [00:03<00:01,  6.33it/s]
DDIM Sampler:  68%|██████▊   | 21/31 [00:03<00:01,  6.33it/s]
DDIM Sampler:  71%|███████   | 22/31 [00:03<00:01,  6.33it/s]
DDIM Sampler:  74%|███████▍  | 23/31 [00:03<00:01,  6.33it/s]
DDIM Sampler:  77%|███████▋  | 24/31 [00:03<00:01,  6.33it/s]
DDIM Sampler:  81%|████████  | 25/31 [00:03<00:00,  6.33it/s]
DDIM Sampler:  84%|████████▍ | 26/31 [00:04<00:00,  6.33it/s]
DDIM Sampler:  87%|████████▋ | 27/31 [00:04<00:00,  6.33it/s]
DDIM Sampler:  90%|█████████ | 28/31 [00:04<00:00,  6.33it/s]
DDIM Sampler:  94%|█████████▎| 29/31 [00:04<00:00,  6.33it/s]
DDIM Sampler:  97%|█████████▋| 30/31 [00:04<00:00,  6.32it/s]
DDIM Sampler: 100%|██████████| 31/31 [00:04<00:00,  6.32it/s]
DDIM Sampler: 100%|██████████| 31/31 [00:04<00:00,  6.31it/s]

Examples

View more examples

Run time and cost

This model costs approximately $0.0067 to run on Replicate, or 149 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 7 seconds.

Readme

Cog Implementation of ali-vilab/AnyDoor

AnyDoor: Zero-shot Object-level Image Customization

banner

Note that AnyDoor does not contain any specific design/tuning for tryon, we think it would be helpful to add skeleton infos or warped garment, and tune on tryon data to make it better :) clothes

@article{chen2023anydoor,
  title={Anydoor: Zero-shot object-level image customization},
  author={Chen, Xi and Huang, Lianghua and Liu, Yu and Shen, Yujun and Zhao, Deli and Zhao, Hengshuang},
  journal={arXiv preprint arXiv:2307.09481},
  year={2023}
}