Input

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

npx create-replicate --model=kvfrans/clipdraw

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

const output = await replicate.run(
  "kvfrans/clipdraw:5797a99edc939ea0e9242d5e8c9cb3bc7d125b1eac21bda852e5cb79ede2cd9b",
  {
    input: {
      prompt: "a 3D rendering of a temple",
      num_paths: 256,
      num_iterations: 500,
      display_frequency: 20
    }
  }
);

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

output = replicate.run(
    "kvfrans/clipdraw:5797a99edc939ea0e9242d5e8c9cb3bc7d125b1eac21bda852e5cb79ede2cd9b",
    input={
        "prompt": "a 3D rendering of a temple",
        "num_paths": 256,
        "num_iterations": 500,
        "display_frequency": 20
    }
)
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 kvfrans/clipdraw 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": "kvfrans/clipdraw:5797a99edc939ea0e9242d5e8c9cb3bc7d125b1eac21bda852e5cb79ede2cd9b",
    "input": {
      "prompt": "a 3D rendering of a temple",
      "num_paths": 256,
      "num_iterations": 500,
      "display_frequency": 20
    }
  }' \
  https://api.replicate.com/v1/predictions

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

Output

{
  "completed_at": "2021-10-13T21:40:08.192057Z",
  "created_at": "2021-10-13T21:34:30.421702Z",
  "data_removed": false,
  "error": null,
  "id": "ypn6hovmkzbthcpt6f4f5jxrqe",
  "input": {
    "prompt": "a 3D rendering of a temple",
    "num_paths": 256,
    "num_iterations": 500,
    "display_frequency": 20
  },
  "logs": "/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torchvision/transforms/functional_tensor.py:876: UserWarning: Argument fill/fillcolor is not supported for Tensor input. Fill value is zero\n  warnings.warn(\"Argument fill/fillcolor is not supported for Tensor input. Fill value is zero\")\niteration: 0, render:loss: -0.767578125\niteration: 20, render:loss: -1.2333984375\niteration: 40, render:loss: -1.380859375\niteration: 60, render:loss: -1.46875\niteration: 80, render:loss: -1.5146484375\niteration: 100, render:loss: -1.5537109375\niteration: 120, render:loss: -1.5546875\niteration: 140, render:loss: -1.541015625\niteration: 160, render:loss: -1.587890625\niteration: 180, render:loss: -1.5869140625\niteration: 200, render:loss: -1.640625\niteration: 220, render:loss: -1.5888671875\niteration: 240, render:loss: -1.63671875\niteration: 260, render:loss: -1.681640625\niteration: 280, render:loss: -1.646484375\niteration: 300, render:loss: -1.736328125\niteration: 320, render:loss: -1.6767578125\niteration: 340, render:loss: -1.708984375\niteration: 360, render:loss: -1.7412109375\niteration: 380, render:loss: -1.796875\niteration: 400, render:loss: -1.8154296875\niteration: 420, render:loss: -1.8125\niteration: 440, render:loss: -1.7978515625\niteration: 460, render:loss: -1.8515625\niteration: 480, render:loss: -1.80859375\niteration: 499, render:loss: -1.8671875",
  "metrics": {
    "total_time": 337.770355
  },
  "output": [
    {
      "file": "https://replicate.delivery/mgxm/fa8ce451-566d-4491-b4df-f9505ab714e7/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/463c4a3b-310b-4efb-bd0b-f7a3aa4562fe/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/3a865840-74a6-4a27-b251-a75afd736552/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/47e7c5db-e4e5-4fc3-85b9-596788a0a2bf/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/d57c035e-cb9e-43af-be15-828d1c5627ed/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/84176b96-ea36-4a73-aea9-5e5fc3cf55e2/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/f53a007b-a70f-4f10-9ea7-c5d62015397e/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/4499f1ed-238f-493c-9fd7-5409fbe40287/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/2ff1e676-1ebf-4c8a-81da-9cc86a81be83/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/8f20c69c-8a08-48b7-91cf-a8ba9f0b9054/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/0da177c0-9b54-4667-a145-3aca9f9549cc/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/54ba8903-758f-41e3-9a5c-9862972eec05/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/64c07c44-2e61-4651-91e9-2ef9e7f198ca/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/5a98b89d-113f-433c-9791-d21affaec94d/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/91d1ec5e-a1cb-4ed8-a505-ed3adc58cf78/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/e7e8dcab-ba8b-4f9a-9a15-de38ac0cc12c/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/3a3ff7b0-be6e-4efb-83a5-bcf0a4a890c7/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/5dffdab7-1023-4d61-8422-006ac1ffeb30/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/6bdf5057-fa1d-466b-8757-4328e62e5eaf/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/df3f74ab-e8a1-4e00-b2de-40349cf476be/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/a876114c-3a21-49b8-9931-9e42bc234ac5/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/dae4a353-89a9-4344-aac1-72afec76c88f/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/5ab4dfdb-f6a1-400b-93a0-bfef454f1058/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/701146c9-7f99-4b64-a519-bdb54d2334ab/out.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/0986e041-7be2-4d63-9388-dd1cc6154961/out.png"
    }
  ],
  "started_at": "2021-12-01T09:05:06.207880Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/ypn6hovmkzbthcpt6f4f5jxrqe",
    "cancel": "https://api.replicate.com/v1/predictions/ypn6hovmkzbthcpt6f4f5jxrqe/cancel"
  },
  "version": "1bbb18ae07d3ac65c2ae031ac759fd5354a84dcaf0fa7b838334bb048a2cafb8"
}

Tweak itReport

This output was created using a different version of the model, kvfrans/clipdraw:1bbb18ae.

Examples

View more examples

Run time and cost

This model costs approximately $0.094 to run on Replicate, or 10 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 T4 GPU hardware. Predictions typically complete within 7 minutes. The predict time for this model varies significantly based on the inputs.

Readme

CLIPDraw: Synthesize drawings to match a text prompt!

This work presents CLIPDraw, an algorithm that synthesizes novel drawings based on natural language input. CLIPDraw does not require any training; rather a pre-trained CLIP language-image encoder is used as a metric for maximizing similarity between the given description and a generated drawing. Crucially, CLIPDraw operates over vector strokes rather than pixel images, a constraint that biases drawings towards simpler human-recognizable shapes. Results compare between CLIPDraw and other synthesis-through-optimization methods, as well as highlight various interesting behaviors of CLIPDraw, such as satisfying ambiguous text in multiple ways, reliably producing drawings in diverse artistic styles, and scaling from simple to complex visual representations as stroke count is increased.