lucataco / ip-adapter-faceid

(Research only) IP-Adapter-FaceID can generate various style images conditioned on a face with only text prompts

Cold

Public
30.1K runs
L40S
GitHub
Paper

Run with an API

Playground API Examples README Versions

Input

face_image

*file

Input face image

prompt

string

Shift + Return to add a new line

photo of a woman in red dress in a gardenphoto of a woman in red dress in a garden

Input prompt

Default: "photo of a woman in red dress in a garden"

negative_prompt

string

Shift + Return to add a new line

monochrome, lowres, bad anatomy, worst quality, low quality, blurry, multiple peoplemonochrome, lowres, bad anatomy, worst quality, low quality, blurry, multiple people

Input Negative Prompt

Default: "monochrome, lowres, bad anatomy, worst quality, low quality, blurry"

width

integer

Width of output image

Default: 1024

height

integer

Height of output image

Default: 1024

num_outputs

integer

(minimum: 1, maximum: 4)

Number of images to output

Default: 1

num_inference_steps

integer

(minimum: 1, maximum: 200)

Number of denoising steps

Default: 30

seed

integer

Random seed. Leave blank to randomize the seed

agree_to_research_only

boolean

You must agree to use this model only for research. It is not for commercial use.

Default: false

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

npx create-replicate --model=lucataco/ip-adapter-faceid

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 lucataco/ip-adapter-faceid using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "lucataco/ip-adapter-faceid:fb81ef963e74776af72e6f380949013533d46dd5c6228a9e586c57db6303d7cd",
  {
    input: {
      seed: 2212213399,
      width: 1024,
      height: 1024,
      prompt: "photo of a woman in red dress in a garden",
      face_image: "https://replicate.delivery/pbxt/K5DSwf3aUzIpS4srbRhNQkESybPovfXwEfjIuDMj3Dz86tDV/demo.png",
      num_outputs: 1,
      negative_prompt: "monochrome, lowres, bad anatomy, worst quality, low quality, blurry, multiple people",
      num_inference_steps: 30,
      agree_to_research_only: true
    }
  }
);

// To access the file URL:
console.log(output[0].url()); //=> "http://example.com"

// To write the file to disk:
fs.writeFile("my-image.png", output[0]);

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 lucataco/ip-adapter-faceid using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "lucataco/ip-adapter-faceid:fb81ef963e74776af72e6f380949013533d46dd5c6228a9e586c57db6303d7cd",
    input={
        "seed": 2212213399,
        "width": 1024,
        "height": 1024,
        "prompt": "photo of a woman in red dress in a garden",
        "face_image": "https://replicate.delivery/pbxt/K5DSwf3aUzIpS4srbRhNQkESybPovfXwEfjIuDMj3Dz86tDV/demo.png",
        "num_outputs": 1,
        "negative_prompt": "monochrome, lowres, bad anatomy, worst quality, low quality, blurry, multiple people",
        "num_inference_steps": 30,
        "agree_to_research_only": True
    }
)
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 lucataco/ip-adapter-faceid 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": "fb81ef963e74776af72e6f380949013533d46dd5c6228a9e586c57db6303d7cd",
    "input": {
      "seed": 2212213399,
      "width": 1024,
      "height": 1024,
      "prompt": "photo of a woman in red dress in a garden",
      "face_image": "https://replicate.delivery/pbxt/K5DSwf3aUzIpS4srbRhNQkESybPovfXwEfjIuDMj3Dz86tDV/demo.png",
      "num_outputs": 1,
      "negative_prompt": "monochrome, lowres, bad anatomy, worst quality, low quality, blurry, multiple people",
      "num_inference_steps": 30,
      "agree_to_research_only": true
    }
  }' \
  https://api.replicate.com/v1/predictions

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

Output

{
  "completed_at": "2023-12-20T19:00:27.975310Z",
  "created_at": "2023-12-20T19:00:08.612807Z",
  "data_removed": false,
  "error": null,
  "id": "yzcljc3bq3iap6xal2q3cpequy",
  "input": {
    "seed": 2212213399,
    "width": 1024,
    "height": 1024,
    "prompt": "photo of a woman in red dress in a garden",
    "face_image": "https://replicate.delivery/pbxt/K5DSwf3aUzIpS4srbRhNQkESybPovfXwEfjIuDMj3Dz86tDV/demo.png",
    "num_outputs": 1,
    "negative_prompt": "monochrome, lowres, bad anatomy, worst quality, low quality, blurry, multiple people",
    "num_inference_steps": 30,
    "agree_to_research_only": true
  },
  "logs": "Using seed: 2212213399\nset det-size: (640, 640)\nwarning: det_size is already set in detection model, ignore\n  0%|          | 0/30 [00:00<?, ?it/s]\n  3%|▎         | 1/30 [00:00<00:16,  1.73it/s]\n  7%|▋         | 2/30 [00:01<00:16,  1.73it/s]\n 10%|█         | 3/30 [00:01<00:15,  1.73it/s]\n 13%|█▎        | 4/30 [00:02<00:15,  1.73it/s]\n 17%|█▋        | 5/30 [00:02<00:14,  1.73it/s]\n 20%|██        | 6/30 [00:03<00:13,  1.72it/s]\n 23%|██▎       | 7/30 [00:04<00:13,  1.72it/s]\n 27%|██▋       | 8/30 [00:04<00:12,  1.72it/s]\n 30%|███       | 9/30 [00:05<00:12,  1.72it/s]\n 33%|███▎      | 10/30 [00:05<00:11,  1.72it/s]\n 37%|███▋      | 11/30 [00:06<00:11,  1.72it/s]\n 40%|████      | 12/30 [00:06<00:10,  1.72it/s]\n 43%|████▎     | 13/30 [00:07<00:09,  1.72it/s]\n 47%|████▋     | 14/30 [00:08<00:09,  1.72it/s]\n 50%|█████     | 15/30 [00:08<00:08,  1.72it/s]\n 53%|█████▎    | 16/30 [00:09<00:08,  1.72it/s]\n 57%|█████▋    | 17/30 [00:09<00:07,  1.72it/s]\n 60%|██████    | 18/30 [00:10<00:06,  1.72it/s]\n 63%|██████▎   | 19/30 [00:11<00:06,  1.72it/s]\n 67%|██████▋   | 20/30 [00:11<00:05,  1.72it/s]\n 70%|███████   | 21/30 [00:12<00:05,  1.72it/s]\n 73%|███████▎  | 22/30 [00:12<00:04,  1.72it/s]\n 77%|███████▋  | 23/30 [00:13<00:04,  1.72it/s]\n 80%|████████  | 24/30 [00:13<00:03,  1.72it/s]\n 83%|████████▎ | 25/30 [00:14<00:02,  1.72it/s]\n 87%|████████▋ | 26/30 [00:15<00:02,  1.72it/s]\n 90%|█████████ | 27/30 [00:15<00:01,  1.72it/s]\n 93%|█████████▎| 28/30 [00:16<00:01,  1.72it/s]\n 97%|█████████▋| 29/30 [00:16<00:00,  1.72it/s]\n100%|██████████| 30/30 [00:17<00:00,  1.72it/s]\n100%|██████████| 30/30 [00:17<00:00,  1.72it/s]",
  "metrics": {
    "predict_time": 19.324499,
    "total_time": 19.362503
  },
  "output": [
    "https://replicate.delivery/pbxt/HAS5ByGkTaIYCprOe0Z0FPSs5zC3o2zYSHIMlO8CONrlXHCJA/out-0.png"
  ],
  "started_at": "2023-12-20T19:00:08.650811Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/yzcljc3bq3iap6xal2q3cpequy",
    "cancel": "https://api.replicate.com/v1/predictions/yzcljc3bq3iap6xal2q3cpequy/cancel"
  },
  "version": "fb81ef963e74776af72e6f380949013533d46dd5c6228a9e586c57db6303d7cd"
}

Generated in

19.3 seconds

Tweak itReport

Using seed: 2212213399
set det-size: (640, 640)
warning: det_size is already set in detection model, ignore
  0%|          | 0/30 [00:00<?, ?it/s]
  3%|▎         | 1/30 [00:00<00:16,  1.73it/s]
  7%|▋         | 2/30 [00:01<00:16,  1.73it/s]
 10%|█         | 3/30 [00:01<00:15,  1.73it/s]
 13%|█▎        | 4/30 [00:02<00:15,  1.73it/s]
 17%|█▋        | 5/30 [00:02<00:14,  1.73it/s]
 20%|██        | 6/30 [00:03<00:13,  1.72it/s]
 23%|██▎       | 7/30 [00:04<00:13,  1.72it/s]
 27%|██▋       | 8/30 [00:04<00:12,  1.72it/s]
 30%|███       | 9/30 [00:05<00:12,  1.72it/s]
 33%|███▎      | 10/30 [00:05<00:11,  1.72it/s]
 37%|███▋      | 11/30 [00:06<00:11,  1.72it/s]
 40%|████      | 12/30 [00:06<00:10,  1.72it/s]
 43%|████▎     | 13/30 [00:07<00:09,  1.72it/s]
 47%|████▋     | 14/30 [00:08<00:09,  1.72it/s]
 50%|█████     | 15/30 [00:08<00:08,  1.72it/s]
 53%|█████▎    | 16/30 [00:09<00:08,  1.72it/s]
 57%|█████▋    | 17/30 [00:09<00:07,  1.72it/s]
 60%|██████    | 18/30 [00:10<00:06,  1.72it/s]
 63%|██████▎   | 19/30 [00:11<00:06,  1.72it/s]
 67%|██████▋   | 20/30 [00:11<00:05,  1.72it/s]
 70%|███████   | 21/30 [00:12<00:05,  1.72it/s]
 73%|███████▎  | 22/30 [00:12<00:04,  1.72it/s]
 77%|███████▋  | 23/30 [00:13<00:04,  1.72it/s]
 80%|████████  | 24/30 [00:13<00:03,  1.72it/s]
 83%|████████▎ | 25/30 [00:14<00:02,  1.72it/s]
 87%|████████▋ | 26/30 [00:15<00:02,  1.72it/s]
 90%|█████████ | 27/30 [00:15<00:01,  1.72it/s]
 93%|█████████▎| 28/30 [00:16<00:01,  1.72it/s]
 97%|█████████▋| 29/30 [00:16<00:00,  1.72it/s]
100%|██████████| 30/30 [00:17<00:00,  1.72it/s]
100%|██████████| 30/30 [00:17<00:00,  1.72it/s]

Examples

View more examples

Run time and cost

This model costs approximately $0.0075 to run on Replicate, or 133 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 8 seconds.

Readme

Implementation of h94/IP-Adapter-FaceID

Introduction

An experimental version of IP-Adapter-FaceID: we use face ID embedding from a face recognition model instead of CLIP image embedding, additionally, we use LoRA to improve ID consistency. IP-Adapter-FaceID can generate various style images conditioned on a face with only text prompts.

example

Limitations and Bias

The model does not achieve perfect photorealism and ID consistency.
The generalization of the model is limited due to limitations of the training data, base model and face recognition model.
Non-commercial use

This model is released exclusively for research purposes and is not intended for commercial use.