cjwbw/supir – Run with an API on Replicate

cjwbw / supir

Practicing Model Scaling for Photo-Realistic Image Restoration In the Wild. This version uses LLaVA-13b for captioning.

Cold

Public
185.8K runs
A100 (80GB)
GitHub
Paper
License

Run with an API

Playground API Examples README Versions

Input

model_name

string

Choose a model. SUPIR-v0Q is the default training settings with paper. SUPIR-v0F is high generalization and high image quality in most cases. Training with light degradation settings. Stage1 encoder of SUPIR-v0F remains more details when facing light degradations.

Default: "SUPIR-v0Q"

image

*file

Low quality input image.

upscale

integer

Upsampling ratio of given inputs.

Default: 1

min_size

number

Minimum resolution of output images.

Default: 1024

edm_steps

integer

(minimum: 1, maximum: 500)

Number of steps for EDM Sampling Schedule.

Default: 50

use_llava

boolean

Use LLaVA model to get captions.

Default: true

a_prompt

string

Shift + Return to add a new line

Cinematic, High Contrast, highly detailed, taken using a Canon EOS R camera, hyper detailed photo - realistic maximum detail, 32k, Color Grading, ultra HD, extreme meticulous detailing, skin pore detailing, hyper sharpness, perfect without deformations.Cinematic, High Contrast, highly detailed, taken using a Canon EOS R camera, hyper detailed photo - realistic maximum detail, 32k, Color Grading, ultra HD, extreme meticulous detailing, skin pore detailing, hyper sharpness, perfect without deformations.

Additive positive prompt for the inputs.

Default: "Cinematic, High Contrast, highly detailed, taken using a Canon EOS R camera, hyper detailed photo - realistic maximum detail, 32k, Color Grading, ultra HD, extreme meticulous detailing, skin pore detailing, hyper sharpness, perfect without deformations."

n_prompt

string

Shift + Return to add a new line

painting, oil painting, illustration, drawing, art, sketch, oil painting, cartoon, CG Style, 3D render, unreal engine, blurring, dirty, messy, worst quality, low quality, frames, watermark, signature, jpeg artifacts, deformed, lowres, over-smoothpainting, oil painting, illustration, drawing, art, sketch, oil painting, cartoon, CG Style, 3D render, unreal engine, blurring, dirty, messy, worst quality, low quality, frames, watermark, signature, jpeg artifacts, deformed, lowres, over-smooth

Negative prompt for the inputs.

Default: "painting, oil painting, illustration, drawing, art, sketch, oil painting, cartoon, CG Style, 3D render, unreal engine, blurring, dirty, messy, worst quality, low quality, frames, watermark, signature, jpeg artifacts, deformed, lowres, over-smooth"

color_fix_type

string

Color Fixing Type..

Default: "Wavelet"

s_stage1

integer

Control Strength of Stage1 (negative means invalid).

Default: -1

s_churn

number

Original churn hy-param of EDM.

Default: 5

s_noise

number

Original noise hy-param of EDM.

Default: 1.003

s_cfg

number

(minimum: 1, maximum: 20)

Classifier-free guidance scale for prompts.

Default: 7.5

s_stage2

number

Control Strength of Stage2.

Default: 1

linear_CFG

boolean

Linearly (with sigma) increase CFG from 'spt_linear_CFG' to s_cfg.

Default: false

linear_s_stage2

boolean

Linearly (with sigma) increase s_stage2 from 'spt_linear_s_stage2' to s_stage2.

Default: false

spt_linear_CFG

number

Start point of linearly increasing CFG.

Default: 1

spt_linear_s_stage2

number

Start point of linearly increasing s_stage2.

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=cjwbw/supir

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 cjwbw/supir using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "cjwbw/supir:1302b550b4f7681da87ed0e405016d443fe1fafd64dabce6673401855a5039b5",
  {
    input: {
      image: "https://replicate.delivery/pbxt/KSFJ6sJWyEiPWysywNsAr2enzigJA3NZpG4YMrxrSFEIEShi/05.png",
      s_cfg: 7.5,
      s_churn: 5,
      s_noise: 1.003,
      upscale: 1,
      a_prompt: "Cinematic, High Contrast, highly detailed, taken using a Canon EOS R camera, hyper detailed photo - realistic maximum detail, 32k, Color Grading, ultra HD, extreme meticulous detailing, skin pore detailing, hyper sharpness, perfect without deformations.",
      min_size: 1024,
      n_prompt: "painting, oil painting, illustration, drawing, art, sketch, oil painting, cartoon, CG Style, 3D render, unreal engine, blurring, dirty, messy, worst quality, low quality, frames, watermark, signature, jpeg artifacts, deformed, lowres, over-smooth",
      s_stage1: -1,
      s_stage2: 1,
      edm_steps: 50,
      use_llava: true,
      linear_CFG: false,
      model_name: "SUPIR-v0Q",
      color_fix_type: "Wavelet",
      spt_linear_CFG: 1,
      linear_s_stage2: false,
      spt_linear_s_stage2: 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 cjwbw/supir using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "cjwbw/supir:1302b550b4f7681da87ed0e405016d443fe1fafd64dabce6673401855a5039b5",
    input={
        "image": "https://replicate.delivery/pbxt/KSFJ6sJWyEiPWysywNsAr2enzigJA3NZpG4YMrxrSFEIEShi/05.png",
        "s_cfg": 7.5,
        "s_churn": 5,
        "s_noise": 1.003,
        "upscale": 1,
        "a_prompt": "Cinematic, High Contrast, highly detailed, taken using a Canon EOS R camera, hyper detailed photo - realistic maximum detail, 32k, Color Grading, ultra HD, extreme meticulous detailing, skin pore detailing, hyper sharpness, perfect without deformations.",
        "min_size": 1024,
        "n_prompt": "painting, oil painting, illustration, drawing, art, sketch, oil painting, cartoon, CG Style, 3D render, unreal engine, blurring, dirty, messy, worst quality, low quality, frames, watermark, signature, jpeg artifacts, deformed, lowres, over-smooth",
        "s_stage1": -1,
        "s_stage2": 1,
        "edm_steps": 50,
        "use_llava": True,
        "linear_CFG": False,
        "model_name": "SUPIR-v0Q",
        "color_fix_type": "Wavelet",
        "spt_linear_CFG": 1,
        "linear_s_stage2": False,
        "spt_linear_s_stage2": 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 cjwbw/supir 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": "1302b550b4f7681da87ed0e405016d443fe1fafd64dabce6673401855a5039b5",
    "input": {
      "image": "https://replicate.delivery/pbxt/KSFJ6sJWyEiPWysywNsAr2enzigJA3NZpG4YMrxrSFEIEShi/05.png",
      "s_cfg": 7.5,
      "s_churn": 5,
      "s_noise": 1.003,
      "upscale": 1,
      "a_prompt": "Cinematic, High Contrast, highly detailed, taken using a Canon EOS R camera, hyper detailed photo - realistic maximum detail, 32k, Color Grading, ultra HD, extreme meticulous detailing, skin pore detailing, hyper sharpness, perfect without deformations.",
      "min_size": 1024,
      "n_prompt": "painting, oil painting, illustration, drawing, art, sketch, oil painting, cartoon, CG Style, 3D render, unreal engine, blurring, dirty, messy, worst quality, low quality, frames, watermark, signature, jpeg artifacts, deformed, lowres, over-smooth",
      "s_stage1": -1,
      "s_stage2": 1,
      "edm_steps": 50,
      "use_llava": true,
      "linear_CFG": false,
      "model_name": "SUPIR-v0Q",
      "color_fix_type": "Wavelet",
      "spt_linear_CFG": 1,
      "linear_s_stage2": false,
      "spt_linear_s_stage2": 0
    }
  }' \
  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/cjwbw/supir@sha256:1302b550b4f7681da87ed0e405016d443fe1fafd64dabce6673401855a5039b5 \
  -i 'image="https://replicate.delivery/pbxt/KSFJ6sJWyEiPWysywNsAr2enzigJA3NZpG4YMrxrSFEIEShi/05.png"' \
  -i 's_cfg=7.5' \
  -i 's_churn=5' \
  -i 's_noise=1.003' \
  -i 'upscale=1' \
  -i 'a_prompt="Cinematic, High Contrast, highly detailed, taken using a Canon EOS R camera, hyper detailed photo - realistic maximum detail, 32k, Color Grading, ultra HD, extreme meticulous detailing, skin pore detailing, hyper sharpness, perfect without deformations."' \
  -i 'min_size=1024' \
  -i 'n_prompt="painting, oil painting, illustration, drawing, art, sketch, oil painting, cartoon, CG Style, 3D render, unreal engine, blurring, dirty, messy, worst quality, low quality, frames, watermark, signature, jpeg artifacts, deformed, lowres, over-smooth"' \
  -i 's_stage1=-1' \
  -i 's_stage2=1' \
  -i 'edm_steps=50' \
  -i 'use_llava=true' \
  -i 'linear_CFG=false' \
  -i 'model_name="SUPIR-v0Q"' \
  -i 'color_fix_type="Wavelet"' \
  -i 'spt_linear_CFG=1' \
  -i 'linear_s_stage2=false' \
  -i 'spt_linear_s_stage2=0'

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/cjwbw/supir@sha256:1302b550b4f7681da87ed0e405016d443fe1fafd64dabce6673401855a5039b5
curl -s -X POST \
  -H "Content-Type: application/json" \
  -d $'{
    "input": {
      "image": "https://replicate.delivery/pbxt/KSFJ6sJWyEiPWysywNsAr2enzigJA3NZpG4YMrxrSFEIEShi/05.png",
      "s_cfg": 7.5,
      "s_churn": 5,
      "s_noise": 1.003,
      "upscale": 1,
      "a_prompt": "Cinematic, High Contrast, highly detailed, taken using a Canon EOS R camera, hyper detailed photo - realistic maximum detail, 32k, Color Grading, ultra HD, extreme meticulous detailing, skin pore detailing, hyper sharpness, perfect without deformations.",
      "min_size": 1024,
      "n_prompt": "painting, oil painting, illustration, drawing, art, sketch, oil painting, cartoon, CG Style, 3D render, unreal engine, blurring, dirty, messy, worst quality, low quality, frames, watermark, signature, jpeg artifacts, deformed, lowres, over-smooth",
      "s_stage1": -1,
      "s_stage2": 1,
      "edm_steps": 50,
      "use_llava": true,
      "linear_CFG": false,
      "model_name": "SUPIR-v0Q",
      "color_fix_type": "Wavelet",
      "spt_linear_CFG": 1,
      "linear_s_stage2": false,
      "spt_linear_s_stage2": 0
    }
  }' \
  http://localhost:5000/predictions

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

Output

We were unable to load these images. Please make sure the URLs are valid.

{
  "input": "https://replicate.delivery/pbxt/KSFJ6sJWyEiPWysywNsAr2enzigJA3NZpG4YMrxrSFEIEShi/05.png",
  "outut": "https://replicate.delivery/pbxt/tsZTAUpHvDorAJcaZDd1BpYsBcijNfwhXfDUCAIvrqTsdnZSA/out.png"
}

{
  "completed_at": "2024-02-23T16:18:52.773773Z",
  "created_at": "2024-02-23T16:04:32.875679Z",
  "data_removed": false,
  "error": null,
  "id": "o4b2mqtbpcxyuo52addzll2g6i",
  "input": {
    "image": "https://replicate.delivery/pbxt/KSFJ6sJWyEiPWysywNsAr2enzigJA3NZpG4YMrxrSFEIEShi/05.png",
    "s_cfg": 7.5,
    "s_churn": 5,
    "s_noise": 1.003,
    "upscale": 1,
    "a_prompt": "Cinematic, High Contrast, highly detailed, taken using a Canon EOS R camera, hyper detailed photo - realistic maximum detail, 32k, Color Grading, ultra HD, extreme meticulous detailing, skin pore detailing, hyper sharpness, perfect without deformations.",
    "min_size": 1024,
    "n_prompt": "painting, oil painting, illustration, drawing, art, sketch, oil painting, cartoon, CG Style, 3D render, unreal engine, blurring, dirty, messy, worst quality, low quality, frames, watermark, signature, jpeg artifacts, deformed, lowres, over-smooth",
    "s_stage1": -1,
    "s_stage2": 1,
    "edm_steps": 50,
    "use_llava": true,
    "linear_CFG": false,
    "model_name": "SUPIR-v0Q",
    "color_fix_type": "Wavelet",
    "spt_linear_CFG": 1,
    "linear_s_stage2": false,
    "spt_linear_s_stage2": 0
  },
  "logs": "Using seed: 12774\nCaptions from LLaVA: ['The image features a beautiful bird perched on a thin, dry branch in a field. The bird is colorful and stands out against the brown background of the field. The branch is located in the middle of the scene, with the bird sitting on it, drawing attention to itself. The field is filled with a variety of flowers, adding a vibrant touch to the scene. The overall style of the image is artistic, with a focus on the bird and its natural surroundings.']\nSeed set to 12774",
  "metrics": {
    "predict_time": 18.567178,
    "total_time": 859.898094
  },
  "output": "https://replicate.delivery/pbxt/tsZTAUpHvDorAJcaZDd1BpYsBcijNfwhXfDUCAIvrqTsdnZSA/out.png",
  "started_at": "2024-02-23T16:18:34.206595Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/o4b2mqtbpcxyuo52addzll2g6i",
    "cancel": "https://api.replicate.com/v1/predictions/o4b2mqtbpcxyuo52addzll2g6i/cancel"
  },
  "version": "1302b550b4f7681da87ed0e405016d443fe1fafd64dabce6673401855a5039b5"
}

Generated in

18.6 seconds

Tweak itReport

Examples

View more examples

Run time and cost

This model costs approximately $0.41 to run on Replicate, or 2 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 5 minutes. The predict time for this model varies significantly based on the inputs.

Readme

Scaling Up to Excellence: Practicing Model Scaling for Photo-Realistic Image Restoration In the Wild

NOTE: Due to the memory requirement for LLaVA-13b and both SUPIR-v0F and SUPIR-v0Q, this model is posted on 80G A100. To run on 40G A40, try models https://replicate.com/cjwbw/supir-v0q or https://replicate.com/cjwbw/supir-v0f that do not include LLaVA-13b.

SUPIR-v0Q: Default training settings with paper. High generalization and high image quality in most cases.
SUPIR-v0F: Training with light degradation settings. Stage1 encoder of SUPIR-v0F remains more details when facing light degradations.

BibTeX

@misc{yu2024scaling,
  title={Scaling Up to Excellence: Practicing Model Scaling for Photo-Realistic Image Restoration In the Wild}, 
  author={Fanghua Yu and Jinjin Gu and Zheyuan Li and Jinfan Hu and Xiangtao Kong and Xintao Wang and Jingwen He and Yu Qiao and Chao Dong},
  year={2024},
  eprint={2401.13627},
  archivePrefix={arXiv},
  primaryClass={cs.CV}
}