open-mmlab / pia

Personalized Image Animator

Cold

Public
103.1K runs
L40S
GitHub
Paper
License

Run with an API

Playground API Examples README Versions

Input

prompt

*string

Shift + Return to add a new line

Input prompt.

image

*file

Input image

negative_prompt

string

Shift + Return to add a new line

wrong white balance, dark, sketches,worst quality,low quality, deformed, distorted, disfigured, bad eyes, wrong lips, weird mouth, bad teeth, mutated hands and fingers, bad anatomy,wrong anatomy, amputation, extra limb, missing limb, floating,limbs, disconnected limbs, mutation, ugly, disgusting, bad_pictures, negative_hand-negwrong white balance, dark, sketches,worst quality,low quality, deformed, distorted, disfigured, bad eyes, wrong lips, weird mouth, bad teeth, mutated hands and fingers, bad anatomy,wrong anatomy, amputation, extra limb, missing limb, floating,limbs, disconnected limbs, mutation, ugly, disgusting, bad_pictures, negative_hand-neg

Things do not show in the output.

Default: "wrong white balance, dark, sketches,worst quality,low quality, deformed, distorted, disfigured, bad eyes, wrong lips, weird mouth, bad teeth, mutated hands and fingers, bad anatomy,wrong anatomy, amputation, extra limb, missing limb, floating,limbs, disconnected limbs, mutation, ugly, disgusting, bad_pictures, negative_hand-neg"

style

string

Choose a style

Default: "3d_cartoon"

max_size

integer

Max size (The long edge of the input image will be resized to this value, larger value means slower inference speed)

Default: 512

motion_scale

integer

(minimum: 1, maximum: 3)

Larger value means larger motion but less identity consistency.

Default: 1

sampling_steps

integer

(minimum: 10, maximum: 100)

Number of denoising steps

Default: 25

animation_length

integer

(minimum: 8, maximum: 24)

Length of the output

Default: 16

guidance_scale

number

(minimum: 1, maximum: 20)

Scale for classifier-free guidance

Default: 7.5

ip_adapter_scale

number

(minimum: 0, maximum: 1)

Scale for classifier-free guidance

Default: 0

seed

integer

Random seed. Leave blank to randomize the seed

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

npx create-replicate --model=open-mmlab/pia

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 open-mmlab/pia using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "open-mmlab/pia:7a673cec329e8f9c51a96791030103467e04aad905cc0319ad6501eb2d54e91b",
  {
    input: {
      image: "https://replicate.delivery/pbxt/KA4hkvNyViiiEafjSAKBKQO5yv3Bt6gtwyFXarsMnP8jlDYD/harry.png",
      style: "3d_cartoon",
      prompt: "1boy smiling",
      max_size: 512,
      motion_scale: 1,
      guidance_scale: 7.5,
      sampling_steps: 25,
      negative_prompt: "wrong white balance, dark, sketches,worst quality,low quality, deformed, distorted, disfigured, bad eyes, wrong lips, weird mouth, bad teeth, mutated hands and fingers, bad anatomy,wrong anatomy, amputation, extra limb, missing limb, floating,limbs, disconnected limbs, mutation, ugly, disgusting, bad_pictures, negative_hand-neg",
      animation_length: 16,
      ip_adapter_scale: 0
    }
  }
);
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 open-mmlab/pia using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "open-mmlab/pia:7a673cec329e8f9c51a96791030103467e04aad905cc0319ad6501eb2d54e91b",
    input={
        "image": "https://replicate.delivery/pbxt/KA4hkvNyViiiEafjSAKBKQO5yv3Bt6gtwyFXarsMnP8jlDYD/harry.png",
        "style": "3d_cartoon",
        "prompt": "1boy smiling",
        "max_size": 512,
        "motion_scale": 1,
        "guidance_scale": 7.5,
        "sampling_steps": 25,
        "negative_prompt": "wrong white balance, dark, sketches,worst quality,low quality, deformed, distorted, disfigured, bad eyes, wrong lips, weird mouth, bad teeth, mutated hands and fingers, bad anatomy,wrong anatomy, amputation, extra limb, missing limb, floating,limbs, disconnected limbs, mutation, ugly, disgusting, bad_pictures, negative_hand-neg",
        "animation_length": 16,
        "ip_adapter_scale": 0
    }
)
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 open-mmlab/pia 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": "7a673cec329e8f9c51a96791030103467e04aad905cc0319ad6501eb2d54e91b",
    "input": {
      "image": "https://replicate.delivery/pbxt/KA4hkvNyViiiEafjSAKBKQO5yv3Bt6gtwyFXarsMnP8jlDYD/harry.png",
      "style": "3d_cartoon",
      "prompt": "1boy smiling",
      "max_size": 512,
      "motion_scale": 1,
      "guidance_scale": 7.5,
      "sampling_steps": 25,
      "negative_prompt": "wrong white balance, dark, sketches,worst quality,low quality, deformed, distorted, disfigured, bad eyes, wrong lips, weird mouth, bad teeth, mutated hands and fingers, bad anatomy,wrong anatomy, amputation, extra limb, missing limb, floating,limbs, disconnected limbs, mutation, ugly, disgusting, bad_pictures, negative_hand-neg",
      "animation_length": 16,
      "ip_adapter_scale": 0
    }
  }' \
  https://api.replicate.com/v1/predictions

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

Output

{
  "completed_at": "2024-01-03T11:29:58.244116Z",
  "created_at": "2024-01-03T11:29:26.956553Z",
  "data_removed": false,
  "error": null,
  "id": "rysmfv3b6r32rsppghdv2sfuq4",
  "input": {
    "image": "https://replicate.delivery/pbxt/KA4hkvNyViiiEafjSAKBKQO5yv3Bt6gtwyFXarsMnP8jlDYD/harry.png",
    "style": "3d_cartoon",
    "prompt": "1boy smiling",
    "max_size": 512,
    "motion_scale": 1,
    "guidance_scale": 7.5,
    "sampling_steps": 25,
    "negative_prompt": "wrong white balance, dark, sketches,worst quality,low quality, deformed, distorted, disfigured, bad eyes, wrong lips, weird mouth, bad teeth, mutated hands and fingers, bad anatomy,wrong anatomy, amputation, extra limb, missing limb, floating,limbs, disconnected limbs, mutation, ugly, disgusting, bad_pictures, negative_hand-neg",
    "animation_length": 16,
    "ip_adapter_scale": 0
  },
  "logs": "Using seed: 17054993394507681024\n  0%|          | 0/25 [00:00<?, ?it/s]\n  4%|▍         | 1/25 [00:00<00:17,  1.37it/s]\n  8%|▊         | 2/25 [00:01<00:16,  1.37it/s]\n 12%|█▏        | 3/25 [00:02<00:16,  1.37it/s]\n 16%|█▌        | 4/25 [00:02<00:15,  1.37it/s]\n 20%|██        | 5/25 [00:03<00:14,  1.37it/s]\n 24%|██▍       | 6/25 [00:04<00:13,  1.37it/s]\n 28%|██▊       | 7/25 [00:05<00:13,  1.37it/s]\n 32%|███▏      | 8/25 [00:05<00:12,  1.37it/s]\n 36%|███▌      | 9/25 [00:06<00:11,  1.37it/s]\n 40%|████      | 10/25 [00:07<00:10,  1.37it/s]\n 44%|████▍     | 11/25 [00:08<00:10,  1.37it/s]\n 48%|████▊     | 12/25 [00:08<00:09,  1.37it/s]\n 52%|█████▏    | 13/25 [00:09<00:08,  1.37it/s]\n 56%|█████▌    | 14/25 [00:10<00:08,  1.37it/s]\n 60%|██████    | 15/25 [00:10<00:07,  1.37it/s]\n 64%|██████▍   | 16/25 [00:11<00:06,  1.37it/s]\n 68%|██████▊   | 17/25 [00:12<00:05,  1.37it/s]\n 72%|███████▏  | 18/25 [00:13<00:05,  1.37it/s]\n 76%|███████▌  | 19/25 [00:13<00:04,  1.37it/s]\n 80%|████████  | 20/25 [00:14<00:03,  1.37it/s]\n 84%|████████▍ | 21/25 [00:15<00:02,  1.37it/s]\n 88%|████████▊ | 22/25 [00:16<00:02,  1.37it/s]\n 92%|█████████▏| 23/25 [00:16<00:01,  1.37it/s]\n 96%|█████████▌| 24/25 [00:17<00:00,  1.37it/s]\n100%|██████████| 25/25 [00:18<00:00,  1.37it/s]\n100%|██████████| 25/25 [00:18<00:00,  1.37it/s]\n  0%|          | 0/16 [00:00<?, ?it/s]\n 31%|███▏      | 5/16 [00:00<00:00, 42.26it/s]\n 62%|██████▎   | 10/16 [00:00<00:00, 22.92it/s]\n 81%|████████▏ | 13/16 [00:00<00:00, 20.70it/s]\n100%|██████████| 16/16 [00:00<00:00, 19.49it/s]\n100%|██████████| 16/16 [00:00<00:00, 21.28it/s]",
  "metrics": {
    "predict_time": 20.751483,
    "total_time": 31.287563
  },
  "output": "https://replicate.delivery/pbxt/CuwMAPeDNk3MAaPp54aI5fhFGGgxfeIjcemJt44u9ogrm7FRC/out.mp4",
  "started_at": "2024-01-03T11:29:37.492633Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/rysmfv3b6r32rsppghdv2sfuq4",
    "cancel": "https://api.replicate.com/v1/predictions/rysmfv3b6r32rsppghdv2sfuq4/cancel"
  },
  "version": "7a673cec329e8f9c51a96791030103467e04aad905cc0319ad6501eb2d54e91b"
}

Generated in

20.8 seconds

Tweak itReport

Using seed: 17054993394507681024
  0%|          | 0/25 [00:00<?, ?it/s]
  4%|▍         | 1/25 [00:00<00:17,  1.37it/s]
  8%|▊         | 2/25 [00:01<00:16,  1.37it/s]
 12%|█▏        | 3/25 [00:02<00:16,  1.37it/s]
 16%|█▌        | 4/25 [00:02<00:15,  1.37it/s]
 20%|██        | 5/25 [00:03<00:14,  1.37it/s]
 24%|██▍       | 6/25 [00:04<00:13,  1.37it/s]
 28%|██▊       | 7/25 [00:05<00:13,  1.37it/s]
 32%|███▏      | 8/25 [00:05<00:12,  1.37it/s]
 36%|███▌      | 9/25 [00:06<00:11,  1.37it/s]
 40%|████      | 10/25 [00:07<00:10,  1.37it/s]
 44%|████▍     | 11/25 [00:08<00:10,  1.37it/s]
 48%|████▊     | 12/25 [00:08<00:09,  1.37it/s]
 52%|█████▏    | 13/25 [00:09<00:08,  1.37it/s]
 56%|█████▌    | 14/25 [00:10<00:08,  1.37it/s]
 60%|██████    | 15/25 [00:10<00:07,  1.37it/s]
 64%|██████▍   | 16/25 [00:11<00:06,  1.37it/s]
 68%|██████▊   | 17/25 [00:12<00:05,  1.37it/s]
 72%|███████▏  | 18/25 [00:13<00:05,  1.37it/s]
 76%|███████▌  | 19/25 [00:13<00:04,  1.37it/s]
 80%|████████  | 20/25 [00:14<00:03,  1.37it/s]
 84%|████████▍ | 21/25 [00:15<00:02,  1.37it/s]
 88%|████████▊ | 22/25 [00:16<00:02,  1.37it/s]
 92%|█████████▏| 23/25 [00:16<00:01,  1.37it/s]
 96%|█████████▌| 24/25 [00:17<00:00,  1.37it/s]
100%|██████████| 25/25 [00:18<00:00,  1.37it/s]
100%|██████████| 25/25 [00:18<00:00,  1.37it/s]
  0%|          | 0/16 [00:00<?, ?it/s]
 31%|███▏      | 5/16 [00:00<00:00, 42.26it/s]
 62%|██████▎   | 10/16 [00:00<00:00, 22.92it/s]
 81%|████████▏ | 13/16 [00:00<00:00, 20.70it/s]
100%|██████████| 16/16 [00:00<00:00, 19.49it/s]
100%|██████████| 16/16 [00:00<00:00, 21.28it/s]

Examples

View more examples

Run time and cost

This model costs approximately $0.070 to run on Replicate, or 14 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 72 seconds. The predict time for this model varies significantly based on the inputs.

Readme

PIA：Personalized Image Animator

PIA is a personalized image animation method which can generate videos with high motion controllability and strong text and image alignment.

AnimateBench

We have open-sourced AnimateBench on HuggingFace which includes images, prompts and configs to evaluate PIA and other image animation methods.

Contact Us

Yiming Zhang: zhangyiming@pjlab.org.cn

Zhening Xing: xingzhening@pjlab.org.cn

Yanhong Zeng: zengyanhong@pjlab.org.cn

Acknowledgements

The code is built upon AnimateDiff, Tune-a-Video and PySceneDetect