mserro/upscaler-pro | Run with an API on Replicate

mserro / upscaler-pro

AI Photorealistic Image Ultra-Resolution, Restoration and Upscale! (Updated 1 year ago)

Cold

Public
85.2K runs
A100 (80GB)

Iterate in playground

Run with an API

Playground API Examples README Versions

Input

image

*file

Preview

input image

prompt

string

Shift + Return to add a new line

masterpiece, best quality, highres, <lora:more_details:0.5> <lora:SDXLrender_v2.0:1>masterpiece, best quality, highres, <lora:more_details:0.5> <lora:SDXLrender_v2.0:1>

Prompt

Default: "masterpiece, best quality, highres, <lora:more_details:0.5> <lora:SDXLrender_v2.0:1>"

negative_prompt

string

Shift + Return to add a new line

(worst quality, low quality, normal quality:2) JuggernautNegative-neg(worst quality, low quality, normal quality:2) JuggernautNegative-neg

Negative Prompt

Default: "(worst quality, low quality, normal quality:2) JuggernautNegative-neg"

scale_factor

number

Scale factor

Default: 2

dynamic

number

(minimum: 1, maximum: 50)

HDR, try from 3 - 9

Default: 6

creativity

number

(minimum: 0, maximum: 1)

Creativity, try from 0.3 - 0.9

Default: 0.35

resemblance

number

(minimum: 0, maximum: 3)

Resemblance, try from 0.3 - 1.6

Default: 0.6

tiling_width

integer

Fractality, set lower tile width for a high Fractality

Default: 112

tiling_height

integer

Fractality, set lower tile height for a high Fractality

Default: 144

sd_model

string

Stable Diffusion model checkpoint

Default: "juggernaut_reborn.safetensors [338b85bc4f]"

scheduler

string

scheduler

Default: "DPM++ 3M SDE Karras"

num_inference_steps

integer

(minimum: 1, maximum: 100)

Number of denoising steps

Default: 18

seed

integer

Random seed. Leave blank to randomize the seed

Default: 1337

downscaling

boolean

Downscale the image before upscaling. Can improve quality and speed for images with high resolution but lower quality

Default: false

downscaling_resolution

integer

Downscaling resolution

Default: 768

lora_links

string

Shift + Return to add a new line

Link to a lora file you want to use in your upscaling. Multiple links possible, seperated by comma

Default: ""

custom_sd_model

string

Shift + Return to add a new line

Default: ""

sharpen

number

(minimum: 0, maximum: 10)

Sharpen the image after upscaling. The higher the value, the more sharpening is applied. 0 for no sharpening

Default: 0

mask

file

Mask image to mark areas that should be preserved during upscaling

handfix

string

Use clarity to fix hands in the image

Default: "disabled"

output_format

string

Format of the output images

Default: "png"

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

npx create-replicate --model=mserro/upscaler-pro

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 mserro/upscaler-pro using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "mserro/upscaler-pro:94a45ebf7e50ce3477a86b7aa21dcefc31866bea6f867cbc1b69f9b1c8bea91d",
  {
    input: {
      seed: 1337,
      image: "https://replicate.delivery/pbxt/L0jFLiSGvFQV0OrL4EcAKtpTXiG1UfMoRZ3YJGf8JwtM44yx/cat07blowres_p.png",
      prompt: "masterpiece, best quality, highres, <lora:more_details:0.5> <lora:SDXLrender_v2.0:1>",
      dynamic: 6,
      handfix: "disabled",
      sharpen: 0,
      sd_model: "juggernaut_reborn.safetensors [338b85bc4f]",
      scheduler: "DPM++ 3M SDE Karras",
      creativity: 0.35,
      lora_links: "",
      downscaling: false,
      resemblance: 0.6,
      scale_factor: 2,
      tiling_width: 112,
      output_format: "png",
      tiling_height: 144,
      custom_sd_model: "",
      negative_prompt: "(worst quality, low quality, normal quality:2) JuggernautNegative-neg",
      num_inference_steps: 18,
      downscaling_resolution: 768
    }
  }
);

// 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 mserro/upscaler-pro using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "mserro/upscaler-pro:94a45ebf7e50ce3477a86b7aa21dcefc31866bea6f867cbc1b69f9b1c8bea91d",
    input={
        "seed": 1337,
        "image": "https://replicate.delivery/pbxt/L0jFLiSGvFQV0OrL4EcAKtpTXiG1UfMoRZ3YJGf8JwtM44yx/cat07blowres_p.png",
        "prompt": "masterpiece, best quality, highres, <lora:more_details:0.5> <lora:SDXLrender_v2.0:1>",
        "dynamic": 6,
        "handfix": "disabled",
        "sharpen": 0,
        "sd_model": "juggernaut_reborn.safetensors [338b85bc4f]",
        "scheduler": "DPM++ 3M SDE Karras",
        "creativity": 0.35,
        "lora_links": "",
        "downscaling": False,
        "resemblance": 0.6,
        "scale_factor": 2,
        "tiling_width": 112,
        "output_format": "png",
        "tiling_height": 144,
        "custom_sd_model": "",
        "negative_prompt": "(worst quality, low quality, normal quality:2) JuggernautNegative-neg",
        "num_inference_steps": 18,
        "downscaling_resolution": 768
    }
)
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 mserro/upscaler-pro 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": "mserro/upscaler-pro:94a45ebf7e50ce3477a86b7aa21dcefc31866bea6f867cbc1b69f9b1c8bea91d",
    "input": {
      "seed": 1337,
      "image": "https://replicate.delivery/pbxt/L0jFLiSGvFQV0OrL4EcAKtpTXiG1UfMoRZ3YJGf8JwtM44yx/cat07blowres_p.png",
      "prompt": "masterpiece, best quality, highres, <lora:more_details:0.5> <lora:SDXLrender_v2.0:1>",
      "dynamic": 6,
      "handfix": "disabled",
      "sharpen": 0,
      "sd_model": "juggernaut_reborn.safetensors [338b85bc4f]",
      "scheduler": "DPM++ 3M SDE Karras",
      "creativity": 0.35,
      "lora_links": "",
      "downscaling": false,
      "resemblance": 0.6,
      "scale_factor": 2,
      "tiling_width": 112,
      "output_format": "png",
      "tiling_height": 144,
      "custom_sd_model": "",
      "negative_prompt": "(worst quality, low quality, normal quality:2) JuggernautNegative-neg",
      "num_inference_steps": 18,
      "downscaling_resolution": 768
    }
  }' \
  https://api.replicate.com/v1/predictions

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

Output

{
  "completed_at": "2024-05-30T19:36:08.871670Z",
  "created_at": "2024-05-30T19:36:02.400000Z",
  "data_removed": false,
  "error": null,
  "id": "k2mrjbwec1rge0cfsc09104qe8",
  "input": {
    "seed": 1337,
    "image": "https://replicate.delivery/pbxt/L0jFLiSGvFQV0OrL4EcAKtpTXiG1UfMoRZ3YJGf8JwtM44yx/cat07blowres_p.png",
    "prompt": "masterpiece, best quality, highres, <lora:more_details:0.5> <lora:SDXLrender_v2.0:1>",
    "dynamic": 6,
    "handfix": "disabled",
    "sharpen": 0,
    "sd_model": "juggernaut_reborn.safetensors [338b85bc4f]",
    "scheduler": "DPM++ 3M SDE Karras",
    "creativity": 0.35,
    "lora_links": "",
    "downscaling": false,
    "resemblance": 0.6,
    "scale_factor": 2,
    "tiling_width": 112,
    "output_format": "png",
    "tiling_height": 144,
    "custom_sd_model": "",
    "negative_prompt": "(worst quality, low quality, normal quality:2) JuggernautNegative-neg",
    "num_inference_steps": 18,
    "downscaling_resolution": 768
  },
  "logs": "Running prediction\nUpscaling with scale_factor:  2.0\nlibpng warning: iCCP: known incorrect sRGB profile\n[Tiled Diffusion] upscaling image with 4x-UltraSharp...\n[Tiled Diffusion] ignore tiling when there's only 1 tile or nothing to do :)\n2024-05-30 19:36:04,791 - ControlNet - \u001b[0;32mINFO\u001b[0m - unit_separate = False, style_align = False\n2024-05-30 19:36:04,791 - ControlNet - \u001b[0;32mINFO\u001b[0m - Loading model from cache: control_v11f1e_sd15_tile\nlibpng warning: iCCP: known incorrect sRGB profile\n2024-05-30 19:36:04,795 - ControlNet - \u001b[0;32mINFO\u001b[0m - Using preprocessor: tile_resample\n2024-05-30 19:36:04,796 - ControlNet - \u001b[0;32mINFO\u001b[0m - preprocessor resolution = 608\n2024-05-30 19:36:04,881 - ControlNet - \u001b[0;32mINFO\u001b[0m - ControlNet Hooked - Time = 0.09721875190734863\n[Tiled VAE]: the input size is tiny and unnecessary to tile.\n  0%|          | 0/7 [00:00<?, ?it/s]\u001b[A\nTotal progress:   0%|          | 0/7 [00:00<?, ?it/s]\u001b[A\n 14%|█▍        | 1/7 [00:00<00:02,  2.52it/s]\u001b[A\nTotal progress:  29%|██▊       | 2/7 [00:00<00:00,  5.57it/s]\u001b[A\n 29%|██▊       | 2/7 [00:00<00:01,  2.70it/s]\u001b[A\nTotal progress:  43%|████▎     | 3/7 [00:00<00:00,  4.01it/s]\u001b[A\n 43%|████▎     | 3/7 [00:01<00:01,  2.76it/s]\u001b[A\nTotal progress:  57%|█████▋    | 4/7 [00:01<00:00,  3.46it/s]\u001b[A\n 57%|█████▋    | 4/7 [00:01<00:01,  2.80it/s]\u001b[A\nTotal progress:  71%|███████▏  | 5/7 [00:01<00:00,  3.26it/s]\u001b[A\n 71%|███████▏  | 5/7 [00:01<00:00,  2.85it/s]\u001b[A\nTotal progress:  86%|████████▌ | 6/7 [00:01<00:00,  3.11it/s]\u001b[A\n 86%|████████▌ | 6/7 [00:02<00:00,  2.86it/s]\u001b[A\n100%|██████████| 7/7 [00:02<00:00,  2.89it/s]\u001b[A\n100%|██████████| 7/7 [00:02<00:00,  2.83it/s]\nTotal progress: 100%|██████████| 7/7 [00:02<00:00,  3.05it/s]\u001b[A[Tiled VAE]: the input size is tiny and unnecessary to tile.\nMultiDiffusion Sampling:   0%|          | 0/7 [01:00<?, ?it/s]\nTotal progress: 100%|██████████| 7/7 [00:02<00:00,  3.05it/s]\u001b[A\nTotal progress: 100%|██████████| 7/7 [00:02<00:00,  2.49it/s]\nPrediction took 5.9 seconds",
  "metrics": {
    "predict_time": 6.454658,
    "total_time": 6.47167
  },
  "output": [
    "https://replicate.delivery/czjl/Bz3wuzE4PLYYON3qIIddiHm8PRKsgMIVsuCDbwQMfZSUO0cJA/1337-dcdc69ee-1ebb-11ef-9d7e-c2e7039b51d3.png"
  ],
  "started_at": "2024-05-30T19:36:02.417012Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/k2mrjbwec1rge0cfsc09104qe8",
    "cancel": "https://api.replicate.com/v1/predictions/k2mrjbwec1rge0cfsc09104qe8/cancel"
  },
  "version": "94a45ebf7e50ce3477a86b7aa21dcefc31866bea6f867cbc1b69f9b1c8bea91d"
}

Generated in

6.5 seconds

Tweak it ShareReport

Running prediction
Upscaling with scale_factor:  2.0
libpng warning: iCCP: known incorrect sRGB profile
[Tiled Diffusion] upscaling image with 4x-UltraSharp...
[Tiled Diffusion] ignore tiling when there's only 1 tile or nothing to do :)
2024-05-30 19:36:04,791 - ControlNet - INFO - unit_separate = False, style_align = False
2024-05-30 19:36:04,791 - ControlNet - INFO - Loading model from cache: control_v11f1e_sd15_tile
libpng warning: iCCP: known incorrect sRGB profile
2024-05-30 19:36:04,795 - ControlNet - INFO - Using preprocessor: tile_resample
2024-05-30 19:36:04,796 - ControlNet - INFO - preprocessor resolution = 608
2024-05-30 19:36:04,881 - ControlNet - INFO - ControlNet Hooked - Time = 0.09721875190734863
[Tiled VAE]: the input size is tiny and unnecessary to tile.
  0%|          | 0/7 [00:00<?, ?it/s]
Total progress:   0%|          | 0/7 [00:00<?, ?it/s]
 14%|█▍        | 1/7 [00:00<00:02,  2.52it/s]
Total progress:  29%|██▊       | 2/7 [00:00<00:00,  5.57it/s]
 29%|██▊       | 2/7 [00:00<00:01,  2.70it/s]
Total progress:  43%|████▎     | 3/7 [00:00<00:00,  4.01it/s]
 43%|████▎     | 3/7 [00:01<00:01,  2.76it/s]
Total progress:  57%|█████▋    | 4/7 [00:01<00:00,  3.46it/s]
 57%|█████▋    | 4/7 [00:01<00:01,  2.80it/s]
Total progress:  71%|███████▏  | 5/7 [00:01<00:00,  3.26it/s]
 71%|███████▏  | 5/7 [00:01<00:00,  2.85it/s]
Total progress:  86%|████████▌ | 6/7 [00:01<00:00,  3.11it/s]
 86%|████████▌ | 6/7 [00:02<00:00,  2.86it/s]
100%|██████████| 7/7 [00:02<00:00,  2.89it/s]
100%|██████████| 7/7 [00:02<00:00,  2.83it/s]
Total progress: 100%|██████████| 7/7 [00:02<00:00,  3.05it/s][Tiled VAE]: the input size is tiny and unnecessary to tile.
MultiDiffusion Sampling:   0%|          | 0/7 [01:00<?, ?it/s]
Total progress: 100%|██████████| 7/7 [00:02<00:00,  3.05it/s]
Total progress: 100%|██████████| 7/7 [00:02<00:00,  2.49it/s]
Prediction took 5.9 seconds

Examples

View more examples

Run time and cost

This model costs approximately $0.030 to run on Replicate, or 33 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 22 seconds. The predict time for this model varies significantly based on the inputs.

Readme

This model costs approximately $0.0199 to run on Replicate, but this varies depending on your inputs.

This model runs on Nvidia T4 GPU hardware, which costs $0.000225/sec. Predictions typically complete within 70~80 seconds. The predict time for this model varies significantly based on the inputs.