qr2ai/outline – Run with an API on Replicate

qr2ai / outline

From Sketch to Reality: Transforming Outlines into Lifelike Images

Cold

Public
46K runs
A100 (80GB)

Run with an API

Playground API Examples README Versions

Input

image

file

Input image for img2img or inpaint mode

prompt

string

Shift + Return to add a new line

Modern skyscraper, glass facade, urban skyline, clear day.Modern skyscraper, glass facade, urban skyline, clear day.

Input prompt

Default: "An acrylic painting featuring flowing, elegant strokes, set against a swirling backdrop of deep charcoal, slate gray, and ethereal white. Accents of radiant gold and soft amber cascade across the canvas, highlighted by delicate specks of luminescent gold dust."

suffix_prompt

string

Shift + Return to add a new line

Futuristic Concept, Cutting-Edge Technology, Historical Landmark, Timeless Aesthetics, sharp focus, 8k, uhd, file grain, masterpieceFuturistic Concept, Cutting-Edge Technology, Historical Landmark, Timeless Aesthetics, sharp focus, 8k, uhd, file grain, masterpiece

Additional prompt

Default: "Imagine the harmonious blend of graceful forms and cosmic elegance, where each curve and line tells a story amidst the celestial backdrop, captured in a luxurious interplay of dark and light hues."

negative_prompt

string

Shift + Return to add a new line

deformed, animation, anime, cartoon, comic, cropped, out of frame, low res, draft, cgi, low quality render, thumbnaildeformed, animation, anime, cartoon, comic, cropped, out of frame, low res, draft, cgi, low quality render, thumbnail

Input Negative Prompt

Default: "worst quality, low quality, low resolution, blurry, ugly, disfigured, uncrafted, filled ring, packed ring, cross, star, distorted, stagnant, watermark"

sketch_type

string

Choose the type of sketch detector

Default: "HedPidNet"

weight_primary

number

Weight for the primary sketch in the combination (e.g., PidiNet in PidiNetCanny).

Default: 0.7

weight_secondary

number

Weight for the secondary sketch in the combination (e.g., Canny in PidiNetCanny).

Default: 0.6

blur_size

integer

Gaussian blur kernel size (must be an odd number).

Default: 3

kernel_size

integer

Kernel size for morphological operations (must be an odd number).

Default: 3

dilation_iterations

integer

Number of iterations for dilation.

Default: 5

erosion_iterations

integer

Number of iterations for erosion.

Default: 2

use_canny

boolean

Whether to combine with canny detector for better details

Default: false

lora_input

string

Shift + Return to add a new line

Comma-separated list of LoRA models from Hugging Face or local paths. Leave empty to skip LoRA.

Default: ""

lora_scale

string

Shift + Return to add a new line

Comma-separated list of scales for each LoRA model. Must match the number of LoRAs. Leave empty if no LoRA is used.

Default: ""

width

integer

Width of output image

Default: 1920

height

integer

Height of output image

Default: 1088

generate_square

boolean

Whether to generate a square image (assert height == width)

Default: false

num_outputs

integer

(minimum: 1, maximum: 4)

Number of images to output.

Default: 1

num_inference_steps

integer

(minimum: 1, maximum: 500)

Number of denoising steps

Default: 35

guidance_scale

number

(minimum: 1, maximum: 50)

Scale for classifier-free guidance

Default: 12

adapter_conditioning_scale

number

(minimum: 0, maximum: 2)

Scale for adapter module

Default: 0.97

seed

integer

Random seed. Enter 0 to randomize the seed.

Default: 0

sampler

string

The sampling method

Default: "Euler a"

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

npx create-replicate --model=qr2ai/outline

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 qr2ai/outline using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "qr2ai/outline:6f713aeb58eb5034ad353de02d7dd56c9efa79f2214e6b89a790dad8ca67ef49",
  {
    input: {
      seed: 0,
      image: "https://replicate.delivery/pbxt/Lcg9dtkX66pSWfULefLA7UMExDQoonK9lmdA9y1R6kbEF9Q6/outline.png",
      width: 1024,
      height: 1024,
      prompt: "Modern skyscraper, glass facade, urban skyline, clear day.",
      sampler: "Euler a",
      blur_size: 3,
      use_canny: false,
      lora_input: "",
      lora_scale: "",
      kernel_size: 3,
      num_outputs: 1,
      sketch_type: "HedPidNet",
      suffix_prompt: "Futuristic Concept, Cutting-Edge Technology, Historical Landmark, Timeless Aesthetics, sharp focus, 8k, uhd, file grain, masterpiece",
      guidance_scale: 7.5,
      weight_primary: 0.7,
      generate_square: false,
      negative_prompt: "deformed, animation, anime, cartoon, comic, cropped, out of frame, low res, draft, cgi, low quality render, thumbnail",
      weight_secondary: 0.6,
      erosion_iterations: 2,
      dilation_iterations: 1,
      num_inference_steps: 35,
      adapter_conditioning_scale: 0.9
    }
  }
);
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 qr2ai/outline using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "qr2ai/outline:6f713aeb58eb5034ad353de02d7dd56c9efa79f2214e6b89a790dad8ca67ef49",
    input={
        "seed": 0,
        "image": "https://replicate.delivery/pbxt/Lcg9dtkX66pSWfULefLA7UMExDQoonK9lmdA9y1R6kbEF9Q6/outline.png",
        "width": 1024,
        "height": 1024,
        "prompt": "Modern skyscraper, glass facade, urban skyline, clear day.",
        "sampler": "Euler a",
        "blur_size": 3,
        "use_canny": False,
        "lora_input": "",
        "lora_scale": "",
        "kernel_size": 3,
        "num_outputs": 1,
        "sketch_type": "HedPidNet",
        "suffix_prompt": "Futuristic Concept, Cutting-Edge Technology, Historical Landmark, Timeless Aesthetics, sharp focus, 8k, uhd, file grain, masterpiece",
        "guidance_scale": 7.5,
        "weight_primary": 0.7,
        "generate_square": False,
        "negative_prompt": "deformed, animation, anime, cartoon, comic, cropped, out of frame, low res, draft, cgi, low quality render, thumbnail",
        "weight_secondary": 0.6,
        "erosion_iterations": 2,
        "dilation_iterations": 1,
        "num_inference_steps": 35,
        "adapter_conditioning_scale": 0.9
    }
)
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 qr2ai/outline 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": "6f713aeb58eb5034ad353de02d7dd56c9efa79f2214e6b89a790dad8ca67ef49",
    "input": {
      "seed": 0,
      "image": "https://replicate.delivery/pbxt/Lcg9dtkX66pSWfULefLA7UMExDQoonK9lmdA9y1R6kbEF9Q6/outline.png",
      "width": 1024,
      "height": 1024,
      "prompt": "Modern skyscraper, glass facade, urban skyline, clear day.",
      "sampler": "Euler a",
      "blur_size": 3,
      "use_canny": false,
      "lora_input": "",
      "lora_scale": "",
      "kernel_size": 3,
      "num_outputs": 1,
      "sketch_type": "HedPidNet",
      "suffix_prompt": "Futuristic Concept, Cutting-Edge Technology, Historical Landmark, Timeless Aesthetics, sharp focus, 8k, uhd, file grain, masterpiece",
      "guidance_scale": 7.5,
      "weight_primary": 0.7,
      "generate_square": false,
      "negative_prompt": "deformed, animation, anime, cartoon, comic, cropped, out of frame, low res, draft, cgi, low quality render, thumbnail",
      "weight_secondary": 0.6,
      "erosion_iterations": 2,
      "dilation_iterations": 1,
      "num_inference_steps": 35,
      "adapter_conditioning_scale": 0.9
    }
  }' \
  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/qr2ai/outline@sha256:6f713aeb58eb5034ad353de02d7dd56c9efa79f2214e6b89a790dad8ca67ef49 \
  -i 'seed=0' \
  -i 'image="https://replicate.delivery/pbxt/Lcg9dtkX66pSWfULefLA7UMExDQoonK9lmdA9y1R6kbEF9Q6/outline.png"' \
  -i 'width=1024' \
  -i 'height=1024' \
  -i 'prompt="Modern skyscraper, glass facade, urban skyline, clear day."' \
  -i 'sampler="Euler a"' \
  -i 'blur_size=3' \
  -i 'use_canny=false' \
  -i 'lora_input=""' \
  -i 'lora_scale=""' \
  -i 'kernel_size=3' \
  -i 'num_outputs=1' \
  -i 'sketch_type="HedPidNet"' \
  -i 'suffix_prompt="Futuristic Concept, Cutting-Edge Technology, Historical Landmark, Timeless Aesthetics, sharp focus, 8k, uhd, file grain, masterpiece"' \
  -i 'guidance_scale=7.5' \
  -i 'weight_primary=0.7' \
  -i 'generate_square=false' \
  -i 'negative_prompt="deformed, animation, anime, cartoon, comic, cropped, out of frame, low res, draft, cgi, low quality render, thumbnail"' \
  -i 'weight_secondary=0.6' \
  -i 'erosion_iterations=2' \
  -i 'dilation_iterations=1' \
  -i 'num_inference_steps=35' \
  -i 'adapter_conditioning_scale=0.9'

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/qr2ai/outline@sha256:6f713aeb58eb5034ad353de02d7dd56c9efa79f2214e6b89a790dad8ca67ef49
curl -s -X POST \
  -H "Content-Type: application/json" \
  -d $'{
    "input": {
      "seed": 0,
      "image": "https://replicate.delivery/pbxt/Lcg9dtkX66pSWfULefLA7UMExDQoonK9lmdA9y1R6kbEF9Q6/outline.png",
      "width": 1024,
      "height": 1024,
      "prompt": "Modern skyscraper, glass facade, urban skyline, clear day.",
      "sampler": "Euler a",
      "blur_size": 3,
      "use_canny": false,
      "lora_input": "",
      "lora_scale": "",
      "kernel_size": 3,
      "num_outputs": 1,
      "sketch_type": "HedPidNet",
      "suffix_prompt": "Futuristic Concept, Cutting-Edge Technology, Historical Landmark, Timeless Aesthetics, sharp focus, 8k, uhd, file grain, masterpiece",
      "guidance_scale": 7.5,
      "weight_primary": 0.7,
      "generate_square": false,
      "negative_prompt": "deformed, animation, anime, cartoon, comic, cropped, out of frame, low res, draft, cgi, low quality render, thumbnail",
      "weight_secondary": 0.6,
      "erosion_iterations": 2,
      "dilation_iterations": 1,
      "num_inference_steps": 35,
      "adapter_conditioning_scale": 0.9
    }
  }' \
  http://localhost:5000/predictions

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

Output

{
  "completed_at": "2024-11-08T00:18:12.061381Z",
  "created_at": "2024-11-08T00:17:57.145000Z",
  "data_removed": false,
  "error": null,
  "id": "q5c36e43k5rj60ck14hatd9324",
  "input": {
    "seed": 0,
    "image": "https://replicate.delivery/pbxt/Lcg9dtkX66pSWfULefLA7UMExDQoonK9lmdA9y1R6kbEF9Q6/outline.png",
    "width": 1024,
    "height": 1024,
    "prompt": "Modern skyscraper, glass facade, urban skyline, clear day.",
    "sampler": "Euler a",
    "blur_size": 3,
    "use_canny": false,
    "lora_input": "",
    "lora_scale": "",
    "kernel_size": 3,
    "num_outputs": 1,
    "sketch_type": "HedPidNet",
    "suffix_prompt": "Futuristic Concept, Cutting-Edge Technology, Historical Landmark, Timeless Aesthetics, sharp focus, 8k, uhd, file grain, masterpiece",
    "guidance_scale": 7.5,
    "weight_primary": 0.7,
    "generate_square": false,
    "negative_prompt": "deformed, animation, anime, cartoon, comic, cropped, out of frame, low res, draft, cgi, low quality render, thumbnail",
    "weight_secondary": 0.6,
    "erosion_iterations": 2,
    "dilation_iterations": 1,
    "num_inference_steps": 35,
    "adapter_conditioning_scale": 0.9
  },
  "logs": "INFO:root:Image loaded successfully.\nINFO:root:HED sketch generated successfully.\nINFO:root:PidiNet sketch generated successfully.\nINFO:root:Combined HED and PidiNet sketches with weights 0.5384615384615385/0.46153846153846156.\nINFO:root:Random seed generated: 707529904\n  0%|          | 0/35 [00:00<?, ?it/s]\n  3%|▎         | 1/35 [00:00<00:05,  5.67it/s]\n  6%|▌         | 2/35 [00:00<00:04,  7.09it/s]\n  9%|▊         | 3/35 [00:00<00:04,  7.71it/s]\n 11%|█▏        | 4/35 [00:00<00:03,  8.05it/s]\n 14%|█▍        | 5/35 [00:00<00:03,  8.25it/s]\n 17%|█▋        | 6/35 [00:00<00:03,  8.38it/s]\n 20%|██        | 7/35 [00:00<00:03,  8.46it/s]\n 23%|██▎       | 8/35 [00:00<00:03,  8.50it/s]\n 26%|██▌       | 9/35 [00:01<00:03,  8.52it/s]\n 29%|██▊       | 10/35 [00:01<00:02,  8.56it/s]\n 31%|███▏      | 11/35 [00:01<00:02,  8.58it/s]\n 34%|███▍      | 12/35 [00:01<00:02,  8.60it/s]\n 37%|███▋      | 13/35 [00:01<00:02,  8.61it/s]\n 40%|████      | 14/35 [00:01<00:02,  8.62it/s]\n 43%|████▎     | 15/35 [00:01<00:02,  8.62it/s]\n 46%|████▌     | 16/35 [00:01<00:02,  8.63it/s]\n 49%|████▊     | 17/35 [00:02<00:02,  8.63it/s]\n 51%|█████▏    | 18/35 [00:02<00:01,  8.63it/s]\n 54%|█████▍    | 19/35 [00:02<00:01,  8.63it/s]\n 57%|█████▋    | 20/35 [00:02<00:01,  8.63it/s]\n 60%|██████    | 21/35 [00:02<00:01,  8.63it/s]\n 63%|██████▎   | 22/35 [00:02<00:01,  8.64it/s]\n 66%|██████▌   | 23/35 [00:02<00:01,  8.64it/s]\n 69%|██████▊   | 24/35 [00:02<00:01,  8.64it/s]\n 71%|███████▏  | 25/35 [00:02<00:01,  8.64it/s]\n 74%|███████▍  | 26/35 [00:03<00:01,  8.64it/s]\n 77%|███████▋  | 27/35 [00:03<00:00,  8.64it/s]\n 80%|████████  | 28/35 [00:03<00:00,  8.64it/s]\n 83%|████████▎ | 29/35 [00:03<00:00,  8.63it/s]\n 86%|████████▌ | 30/35 [00:03<00:00,  8.63it/s]\n 89%|████████▊ | 31/35 [00:03<00:00,  8.64it/s]\n 91%|█████████▏| 32/35 [00:03<00:00,  8.64it/s]\n 94%|█████████▍| 33/35 [00:03<00:00,  8.54it/s]\n 97%|█████████▋| 34/35 [00:04<00:00,  8.57it/s]\n100%|██████████| 35/35 [00:04<00:00,  8.59it/s]\n100%|██████████| 35/35 [00:04<00:00,  8.49it/s]\nINFO:root:Image generation completed successfully.",
  "metrics": {
    "predict_time": 9.275241489999999,
    "total_time": 14.916381
  },
  "output": [
    "https://replicate.delivery/yhqm/bUwt9mqcX9ZENB2KyEUWxGqEm2HCKNtReiC7tqjTcUeErwuTA/output_0.png"
  ],
  "started_at": "2024-11-08T00:18:02.786140Z",
  "status": "succeeded",
  "urls": {
    "stream": "https://stream.replicate.com/v1/files/qoxq-46kfyl4iue3l43swtbhbyrmlfblbz235obwiy7tfhfwc3tuwbkwq",
    "get": "https://api.replicate.com/v1/predictions/q5c36e43k5rj60ck14hatd9324",
    "cancel": "https://api.replicate.com/v1/predictions/q5c36e43k5rj60ck14hatd9324/cancel"
  },
  "version": "6f713aeb58eb5034ad353de02d7dd56c9efa79f2214e6b89a790dad8ca67ef49"
}

Generated in

9.3 seconds

Tweak it ShareReport

INFO:root:Image loaded successfully.
INFO:root:HED sketch generated successfully.
INFO:root:PidiNet sketch generated successfully.
INFO:root:Combined HED and PidiNet sketches with weights 0.5384615384615385/0.46153846153846156.
INFO:root:Random seed generated: 707529904
  0%|          | 0/35 [00:00<?, ?it/s]
  3%|▎         | 1/35 [00:00<00:05,  5.67it/s]
  6%|▌         | 2/35 [00:00<00:04,  7.09it/s]
  9%|▊         | 3/35 [00:00<00:04,  7.71it/s]
 11%|█▏        | 4/35 [00:00<00:03,  8.05it/s]
 14%|█▍        | 5/35 [00:00<00:03,  8.25it/s]
 17%|█▋        | 6/35 [00:00<00:03,  8.38it/s]
 20%|██        | 7/35 [00:00<00:03,  8.46it/s]
 23%|██▎       | 8/35 [00:00<00:03,  8.50it/s]
 26%|██▌       | 9/35 [00:01<00:03,  8.52it/s]
 29%|██▊       | 10/35 [00:01<00:02,  8.56it/s]
 31%|███▏      | 11/35 [00:01<00:02,  8.58it/s]
 34%|███▍      | 12/35 [00:01<00:02,  8.60it/s]
 37%|███▋      | 13/35 [00:01<00:02,  8.61it/s]
 40%|████      | 14/35 [00:01<00:02,  8.62it/s]
 43%|████▎     | 15/35 [00:01<00:02,  8.62it/s]
 46%|████▌     | 16/35 [00:01<00:02,  8.63it/s]
 49%|████▊     | 17/35 [00:02<00:02,  8.63it/s]
 51%|█████▏    | 18/35 [00:02<00:01,  8.63it/s]
 54%|█████▍    | 19/35 [00:02<00:01,  8.63it/s]
 57%|█████▋    | 20/35 [00:02<00:01,  8.63it/s]
 60%|██████    | 21/35 [00:02<00:01,  8.63it/s]
 63%|██████▎   | 22/35 [00:02<00:01,  8.64it/s]
 66%|██████▌   | 23/35 [00:02<00:01,  8.64it/s]
 69%|██████▊   | 24/35 [00:02<00:01,  8.64it/s]
 71%|███████▏  | 25/35 [00:02<00:01,  8.64it/s]
 74%|███████▍  | 26/35 [00:03<00:01,  8.64it/s]
 77%|███████▋  | 27/35 [00:03<00:00,  8.64it/s]
 80%|████████  | 28/35 [00:03<00:00,  8.64it/s]
 83%|████████▎ | 29/35 [00:03<00:00,  8.63it/s]
 86%|████████▌ | 30/35 [00:03<00:00,  8.63it/s]
 89%|████████▊ | 31/35 [00:03<00:00,  8.64it/s]
 91%|█████████▏| 32/35 [00:03<00:00,  8.64it/s]
 94%|█████████▍| 33/35 [00:03<00:00,  8.54it/s]
 97%|█████████▋| 34/35 [00:04<00:00,  8.57it/s]
100%|██████████| 35/35 [00:04<00:00,  8.59it/s]
100%|██████████| 35/35 [00:04<00:00,  8.49it/s]
INFO:root:Image generation completed successfully.

Examples

View more examples

Run time and cost

This model costs approximately $0.059 to run on Replicate, or 16 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 43 seconds. The predict time for this model varies significantly based on the inputs.

Readme

This model doesn't have a readme.