juergengunz/ultimate-portrait-upscale | Run with an API on Replicate

juergengunz / ultimate-portrait-upscale

Upscale Portrait Images with ControlNet Tile (Updated 1 year, 6 months ago)

Cold

Public
47.9K runs
A100 (80GB)

Iterate in playground

Run with an API

Playground API Examples README Versions

Input

image

*file

Input image

positive_prompt

string

Shift + Return to add a new line

business portrait of a man, detailed skin, perfect skin, soft lighting, beautiful blue eyes, photorealisticbusiness portrait of a man, detailed skin, perfect skin, soft lighting, beautiful blue eyes, photorealistic

Positive Prompt

Default: "business portrait, detailed skin, perfect skin, soft lighting, beautiful eyes, photorealistic, perfect teeth"

negative_prompt

string

Shift + Return to add a new line

cartoon, cgi, render, painting, illustration, drawingcartoon, cgi, render, painting, illustration, drawing

Negative Prompt

Default: "cartoon, cgi, render, painting, illustration, drawing"

upscaler

string

Upscaler

Default: "4x-UltraSharp"

upscale_by

number

Upscale By

Default: 2

use_controlnet_tile

boolean

Use ControlNet Tile

Default: true

controlnet_strength

number

ControlNet Strength

Default: 1

seed

integer

Sampling seed, leave Empty for Random

steps

integer

Steps

Default: 20

cfg

number

CFG

Default: 8

sampler_name

string

Sampler

Default: "euler"

scheduler

string

Scheduler

Default: "normal"

denoise

number

Denoise

Default: 0.1

mode_type

string

Mode Type

Default: "Linear"

tile_width

integer

Tile Width

Default: 512

tile_height

integer

Tile Height

Default: 512

mask_blur

integer

Mask Blur

Default: 8

tile_padding

integer

Tile Padding

Default: 32

seam_fix_mode

string

Seam Fix Mode

Default: "None"

seam_fix_denoise

number

Seam Fix Denoise

Default: 1

seam_fix_width

integer

Seam Fix Width

Default: 64

seam_fix_mask_blur

integer

Seam Fix Mask Blur

Default: 8

seam_fix_padding

integer

Seam Fix Padding

Default: 16

force_uniform_tiles

boolean

Force Uniform Tiles

Default: true

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

npx create-replicate --model=juergengunz/ultimate-portrait-upscale

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";
import fs from "node:fs";

const replicate = new Replicate({
  auth: process.env.REPLICATE_API_TOKEN,
});

Run juergengunz/ultimate-portrait-upscale using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "juergengunz/ultimate-portrait-upscale:f7fdace4ec7adab7fa02688a160eee8057f070ead7fbb84e0904864fd2324be5",
  {
    input: {
      cfg: 8,
      image: "https://replicate.delivery/pbxt/K1jy8x3lgsayBRp1SwKmFoQrFGoEOvQ56VcwGIHM6giTiyPm/test.png",
      steps: 20,
      denoise: 0.1,
      upscaler: "4x-UltraSharp",
      mask_blur: 8,
      mode_type: "Linear",
      scheduler: "normal",
      tile_width: 512,
      upscale_by: 2,
      tile_height: 512,
      sampler_name: "euler",
      tile_padding: 32,
      seam_fix_mode: "None",
      seam_fix_width: 64,
      negative_prompt: "cartoon, cgi, render, painting, illustration, drawing",
      positive_prompt: "business portrait of a man, detailed skin, perfect skin, soft lighting, beautiful blue eyes, photorealistic",
      seam_fix_denoise: 1,
      seam_fix_padding: 16,
      seam_fix_mask_blur: 8,
      controlnet_strength: 1,
      force_uniform_tiles: true,
      use_controlnet_tile: true
    }
  }
);

// 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 juergengunz/ultimate-portrait-upscale using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "juergengunz/ultimate-portrait-upscale:f7fdace4ec7adab7fa02688a160eee8057f070ead7fbb84e0904864fd2324be5",
    input={
        "cfg": 8,
        "image": "https://replicate.delivery/pbxt/K1jy8x3lgsayBRp1SwKmFoQrFGoEOvQ56VcwGIHM6giTiyPm/test.png",
        "steps": 20,
        "denoise": 0.1,
        "upscaler": "4x-UltraSharp",
        "mask_blur": 8,
        "mode_type": "Linear",
        "scheduler": "normal",
        "tile_width": 512,
        "upscale_by": 2,
        "tile_height": 512,
        "sampler_name": "euler",
        "tile_padding": 32,
        "seam_fix_mode": "None",
        "seam_fix_width": 64,
        "negative_prompt": "cartoon, cgi, render, painting, illustration, drawing",
        "positive_prompt": "business portrait of a man, detailed skin, perfect skin, soft lighting, beautiful blue eyes, photorealistic",
        "seam_fix_denoise": 1,
        "seam_fix_padding": 16,
        "seam_fix_mask_blur": 8,
        "controlnet_strength": 1,
        "force_uniform_tiles": True,
        "use_controlnet_tile": 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 juergengunz/ultimate-portrait-upscale 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": "juergengunz/ultimate-portrait-upscale:f7fdace4ec7adab7fa02688a160eee8057f070ead7fbb84e0904864fd2324be5",
    "input": {
      "cfg": 8,
      "image": "https://replicate.delivery/pbxt/K1jy8x3lgsayBRp1SwKmFoQrFGoEOvQ56VcwGIHM6giTiyPm/test.png",
      "steps": 20,
      "denoise": 0.1,
      "upscaler": "4x-UltraSharp",
      "mask_blur": 8,
      "mode_type": "Linear",
      "scheduler": "normal",
      "tile_width": 512,
      "upscale_by": 2,
      "tile_height": 512,
      "sampler_name": "euler",
      "tile_padding": 32,
      "seam_fix_mode": "None",
      "seam_fix_width": 64,
      "negative_prompt": "cartoon, cgi, render, painting, illustration, drawing",
      "positive_prompt": "business portrait of a man, detailed skin, perfect skin, soft lighting, beautiful blue eyes, photorealistic",
      "seam_fix_denoise": 1,
      "seam_fix_padding": 16,
      "seam_fix_mask_blur": 8,
      "controlnet_strength": 1,
      "force_uniform_tiles": true,
      "use_controlnet_tile": true
    }
  }' \
  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/juergengunz/ultimate-portrait-upscale@sha256:f7fdace4ec7adab7fa02688a160eee8057f070ead7fbb84e0904864fd2324be5 \
  -i 'cfg=8' \
  -i 'image="https://replicate.delivery/pbxt/K1jy8x3lgsayBRp1SwKmFoQrFGoEOvQ56VcwGIHM6giTiyPm/test.png"' \
  -i 'steps=20' \
  -i 'denoise=0.1' \
  -i 'upscaler="4x-UltraSharp"' \
  -i 'mask_blur=8' \
  -i 'mode_type="Linear"' \
  -i 'scheduler="normal"' \
  -i 'tile_width=512' \
  -i 'upscale_by=2' \
  -i 'tile_height=512' \
  -i 'sampler_name="euler"' \
  -i 'tile_padding=32' \
  -i 'seam_fix_mode="None"' \
  -i 'seam_fix_width=64' \
  -i 'negative_prompt="cartoon, cgi, render, painting, illustration, drawing"' \
  -i 'positive_prompt="business portrait of a man, detailed skin, perfect skin, soft lighting, beautiful blue eyes, photorealistic"' \
  -i 'seam_fix_denoise=1' \
  -i 'seam_fix_padding=16' \
  -i 'seam_fix_mask_blur=8' \
  -i 'controlnet_strength=1' \
  -i 'force_uniform_tiles=true' \
  -i 'use_controlnet_tile=true'

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/juergengunz/ultimate-portrait-upscale@sha256:f7fdace4ec7adab7fa02688a160eee8057f070ead7fbb84e0904864fd2324be5
curl -s -X POST \
  -H "Content-Type: application/json" \
  -d $'{
    "input": {
      "cfg": 8,
      "image": "https://replicate.delivery/pbxt/K1jy8x3lgsayBRp1SwKmFoQrFGoEOvQ56VcwGIHM6giTiyPm/test.png",
      "steps": 20,
      "denoise": 0.1,
      "upscaler": "4x-UltraSharp",
      "mask_blur": 8,
      "mode_type": "Linear",
      "scheduler": "normal",
      "tile_width": 512,
      "upscale_by": 2,
      "tile_height": 512,
      "sampler_name": "euler",
      "tile_padding": 32,
      "seam_fix_mode": "None",
      "seam_fix_width": 64,
      "negative_prompt": "cartoon, cgi, render, painting, illustration, drawing",
      "positive_prompt": "business portrait of a man, detailed skin, perfect skin, soft lighting, beautiful blue eyes, photorealistic",
      "seam_fix_denoise": 1,
      "seam_fix_padding": 16,
      "seam_fix_mask_blur": 8,
      "controlnet_strength": 1,
      "force_uniform_tiles": true,
      "use_controlnet_tile": true
    }
  }' \
  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/K1jy8x3lgsayBRp1SwKmFoQrFGoEOvQ56VcwGIHM6giTiyPm/test.png",
  "outut": "https://replicate.delivery/pbxt/EykrnMfm3RwgWiLVz5iZV39V0iBqu2MtkeuDj0GDMXFmIABSA/out-1926293.png"
}

{
  "completed_at": "2023-12-10T23:56:22.876546Z",
  "created_at": "2023-12-10T23:56:13.187729Z",
  "data_removed": false,
  "error": null,
  "id": "d2wp6xdboshn2nanm7hghzq6gm",
  "input": {
    "cfg": 8,
    "image": "https://replicate.delivery/pbxt/K1jy8x3lgsayBRp1SwKmFoQrFGoEOvQ56VcwGIHM6giTiyPm/test.png",
    "steps": 20,
    "denoise": 0.1,
    "upscaler": "4x-UltraSharp",
    "mask_blur": 8,
    "mode_type": "Linear",
    "scheduler": "normal",
    "tile_width": 512,
    "upscale_by": 2,
    "tile_height": 512,
    "sampler_name": "euler",
    "tile_padding": 32,
    "seam_fix_mode": "None",
    "seam_fix_width": 64,
    "negative_prompt": "cartoon, cgi, render, painting, illustration, drawing",
    "positive_prompt": "business portrait of a man, detailed skin, perfect skin, soft lighting, beautiful blue eyes, photorealistic",
    "seam_fix_denoise": 1,
    "seam_fix_padding": 16,
    "seam_fix_mask_blur": 8,
    "controlnet_strength": 1,
    "force_uniform_tiles": true,
    "use_controlnet_tile": true
  },
  "logs": "Using seed: 1926293\nUsing ControlNet tile with Ultimate SD Upscale\ngot prompt\nCanva size: 800x1072\nImage size: 400x533\nScale factor: 3\nUpscaling iteration 1 with scale factor 3\nTile size: 512x512\nTiles amount: 6\nGrid: 3x2\nRedraw enabled: True\nSeams fix mode: NONE\nRequested to load ControlNet\nLoading 1 new model\n  0%|          | 0/20 [00:00<?, ?it/s]\n 10%|█         | 2/20 [00:00<00:00, 18.57it/s]\n 25%|██▌       | 5/20 [00:00<00:00, 19.62it/s]\n 40%|████      | 8/20 [00:00<00:00, 20.05it/s]\n 50%|█████     | 10/20 [00:00<00:00, 19.82it/s]\n 65%|██████▌   | 13/20 [00:00<00:00, 19.71it/s]\n 75%|███████▌  | 15/20 [00:00<00:00, 19.76it/s]\n 90%|█████████ | 18/20 [00:00<00:00, 20.00it/s]\n100%|██████████| 20/20 [00:01<00:00, 19.94it/s]\n100%|██████████| 20/20 [00:01<00:00, 19.83it/s]\nRequested to load ControlNet\nLoading 1 new model\nunload clone 2\n  0%|          | 0/20 [00:00<?, ?it/s]\n 10%|█         | 2/20 [00:00<00:00, 19.90it/s]\n 25%|██▌       | 5/20 [00:00<00:00, 20.09it/s]\n 40%|████      | 8/20 [00:00<00:00, 20.15it/s]\n 55%|█████▌    | 11/20 [00:00<00:00, 20.27it/s]\n 70%|███████   | 14/20 [00:00<00:00, 20.35it/s]\n 85%|████████▌ | 17/20 [00:00<00:00, 19.85it/s]\n100%|██████████| 20/20 [00:00<00:00, 20.09it/s]\n100%|██████████| 20/20 [00:00<00:00, 20.11it/s]\nRequested to load ControlNet\nLoading 1 new model\nunload clone 2\n  0%|          | 0/20 [00:00<?, ?it/s]\n 10%|█         | 2/20 [00:00<00:00, 19.61it/s]\n 25%|██▌       | 5/20 [00:00<00:00, 20.10it/s]\n 40%|████      | 8/20 [00:00<00:00, 20.25it/s]\n 55%|█████▌    | 11/20 [00:00<00:00, 19.64it/s]\n 70%|███████   | 14/20 [00:00<00:00, 19.94it/s]\n 85%|████████▌ | 17/20 [00:00<00:00, 20.07it/s]\n100%|██████████| 20/20 [00:00<00:00, 20.17it/s]\n100%|██████████| 20/20 [00:00<00:00, 20.05it/s]\nRequested to load ControlNet\nLoading 1 new model\nunload clone 2\n  0%|          | 0/20 [00:00<?, ?it/s]\n 10%|█         | 2/20 [00:00<00:01, 17.92it/s]\n 20%|██        | 4/20 [00:00<00:00, 18.92it/s]\n 35%|███▌      | 7/20 [00:00<00:00, 19.76it/s]\n 50%|█████     | 10/20 [00:00<00:00, 19.95it/s]\n 65%|██████▌   | 13/20 [00:00<00:00, 20.15it/s]\n 80%|████████  | 16/20 [00:00<00:00, 20.17it/s]\n 95%|█████████▌| 19/20 [00:00<00:00, 20.24it/s]\n100%|██████████| 20/20 [00:01<00:00, 19.98it/s]\nRequested to load ControlNet\nLoading 1 new model\nunload clone 2\n  0%|          | 0/20 [00:00<?, ?it/s]\n 10%|█         | 2/20 [00:00<00:00, 19.62it/s]\n 25%|██▌       | 5/20 [00:00<00:00, 19.91it/s]\n 40%|████      | 8/20 [00:00<00:00, 20.07it/s]\n 55%|█████▌    | 11/20 [00:00<00:00, 20.25it/s]\n 70%|███████   | 14/20 [00:00<00:00, 20.28it/s]\n 85%|████████▌ | 17/20 [00:00<00:00, 20.21it/s]\n100%|██████████| 20/20 [00:00<00:00, 20.15it/s]\n100%|██████████| 20/20 [00:00<00:00, 20.14it/s]\nRequested to load ControlNet\nLoading 1 new model\nunload clone 2\n  0%|          | 0/20 [00:00<?, ?it/s]\n 10%|█         | 2/20 [00:00<00:00, 19.84it/s]\n 25%|██▌       | 5/20 [00:00<00:00, 20.24it/s]\n 40%|████      | 8/20 [00:00<00:00, 20.12it/s]\n 55%|█████▌    | 11/20 [00:00<00:00, 19.81it/s]\n 70%|███████   | 14/20 [00:00<00:00, 20.08it/s]\n 85%|████████▌ | 17/20 [00:00<00:00, 20.20it/s]\n100%|██████████| 20/20 [00:00<00:00, 20.25it/s]\n100%|██████████| 20/20 [00:00<00:00, 20.15it/s]\nPrompt executed in 8.69 seconds\nnode output:  {'images': [{'filename': 'ComfyUI_00035_.png', 'subfolder': '', 'type': 'output'}]}\noutput",
  "metrics": {
    "predict_time": 9.670054,
    "total_time": 9.688817
  },
  "output": "https://replicate.delivery/pbxt/EykrnMfm3RwgWiLVz5iZV39V0iBqu2MtkeuDj0GDMXFmIABSA/out-1926293.png",
  "started_at": "2023-12-10T23:56:13.206492Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/d2wp6xdboshn2nanm7hghzq6gm",
    "cancel": "https://api.replicate.com/v1/predictions/d2wp6xdboshn2nanm7hghzq6gm/cancel"
  },
  "version": "f7fdace4ec7adab7fa02688a160eee8057f070ead7fbb84e0904864fd2324be5"
}

Generated in

9.7 seconds

Tweak itReport

Using seed: 1926293
Using ControlNet tile with Ultimate SD Upscale
got prompt
Canva size: 800x1072
Image size: 400x533
Scale factor: 3
Upscaling iteration 1 with scale factor 3
Tile size: 512x512
Tiles amount: 6
Grid: 3x2
Redraw enabled: True
Seams fix mode: NONE
Requested to load ControlNet
Loading 1 new model
  0%|          | 0/20 [00:00<?, ?it/s]
 10%|█         | 2/20 [00:00<00:00, 18.57it/s]
 25%|██▌       | 5/20 [00:00<00:00, 19.62it/s]
 40%|████      | 8/20 [00:00<00:00, 20.05it/s]
 50%|█████     | 10/20 [00:00<00:00, 19.82it/s]
 65%|██████▌   | 13/20 [00:00<00:00, 19.71it/s]
 75%|███████▌  | 15/20 [00:00<00:00, 19.76it/s]
 90%|█████████ | 18/20 [00:00<00:00, 20.00it/s]
100%|██████████| 20/20 [00:01<00:00, 19.94it/s]
100%|██████████| 20/20 [00:01<00:00, 19.83it/s]
Requested to load ControlNet
Loading 1 new model
unload clone 2
  0%|          | 0/20 [00:00<?, ?it/s]
 10%|█         | 2/20 [00:00<00:00, 19.90it/s]
 25%|██▌       | 5/20 [00:00<00:00, 20.09it/s]
 40%|████      | 8/20 [00:00<00:00, 20.15it/s]
 55%|█████▌    | 11/20 [00:00<00:00, 20.27it/s]
 70%|███████   | 14/20 [00:00<00:00, 20.35it/s]
 85%|████████▌ | 17/20 [00:00<00:00, 19.85it/s]
100%|██████████| 20/20 [00:00<00:00, 20.09it/s]
100%|██████████| 20/20 [00:00<00:00, 20.11it/s]
Requested to load ControlNet
Loading 1 new model
unload clone 2
  0%|          | 0/20 [00:00<?, ?it/s]
 10%|█         | 2/20 [00:00<00:00, 19.61it/s]
 25%|██▌       | 5/20 [00:00<00:00, 20.10it/s]
 40%|████      | 8/20 [00:00<00:00, 20.25it/s]
 55%|█████▌    | 11/20 [00:00<00:00, 19.64it/s]
 70%|███████   | 14/20 [00:00<00:00, 19.94it/s]
 85%|████████▌ | 17/20 [00:00<00:00, 20.07it/s]
100%|██████████| 20/20 [00:00<00:00, 20.17it/s]
100%|██████████| 20/20 [00:00<00:00, 20.05it/s]
Requested to load ControlNet
Loading 1 new model
unload clone 2
  0%|          | 0/20 [00:00<?, ?it/s]
 10%|█         | 2/20 [00:00<00:01, 17.92it/s]
 20%|██        | 4/20 [00:00<00:00, 18.92it/s]
 35%|███▌      | 7/20 [00:00<00:00, 19.76it/s]
 50%|█████     | 10/20 [00:00<00:00, 19.95it/s]
 65%|██████▌   | 13/20 [00:00<00:00, 20.15it/s]
 80%|████████  | 16/20 [00:00<00:00, 20.17it/s]
 95%|█████████▌| 19/20 [00:00<00:00, 20.24it/s]
100%|██████████| 20/20 [00:01<00:00, 19.98it/s]
Requested to load ControlNet
Loading 1 new model
unload clone 2
  0%|          | 0/20 [00:00<?, ?it/s]
 10%|█         | 2/20 [00:00<00:00, 19.62it/s]
 25%|██▌       | 5/20 [00:00<00:00, 19.91it/s]
 40%|████      | 8/20 [00:00<00:00, 20.07it/s]
 55%|█████▌    | 11/20 [00:00<00:00, 20.25it/s]
 70%|███████   | 14/20 [00:00<00:00, 20.28it/s]
 85%|████████▌ | 17/20 [00:00<00:00, 20.21it/s]
100%|██████████| 20/20 [00:00<00:00, 20.15it/s]
100%|██████████| 20/20 [00:00<00:00, 20.14it/s]
Requested to load ControlNet
Loading 1 new model
unload clone 2
  0%|          | 0/20 [00:00<?, ?it/s]
 10%|█         | 2/20 [00:00<00:00, 19.84it/s]
 25%|██▌       | 5/20 [00:00<00:00, 20.24it/s]
 40%|████      | 8/20 [00:00<00:00, 20.12it/s]
 55%|█████▌    | 11/20 [00:00<00:00, 19.81it/s]
 70%|███████   | 14/20 [00:00<00:00, 20.08it/s]
 85%|████████▌ | 17/20 [00:00<00:00, 20.20it/s]
100%|██████████| 20/20 [00:00<00:00, 20.25it/s]
100%|██████████| 20/20 [00:00<00:00, 20.15it/s]
Prompt executed in 8.69 seconds
node output:  {'images': [{'filename': 'ComfyUI_00035_.png', 'subfolder': '', 'type': 'output'}]}
output

Examples

View more examples

Run time and cost

This model costs approximately $0.14 to run on Replicate, or 7 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 98 seconds. The predict time for this model varies significantly based on the inputs.

Readme

This model doesn't have a readme.