pixray/text2image – Run with an API on Replicate

pixray / text2image

Uses pixray to generate an image from text prompt

Cold

Public
1.4M runs
T4
GitHub
License

Run with an API

Playground API Examples README Versions

Input

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

npx create-replicate --model=pixray/text2image

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 pixray/text2image using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "pixray/text2image:5c347a4bfa1d4523a58ae614c2194e15f2ae682b57e3797a5bb468920aa70ebf",
  {
    input: {
      drawer: "vqgan",
      prompts: "Manhattan skyline at sunset. #artstation 🌇",
      settings: "\n"
    }
  }
);
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 pixray/text2image using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "pixray/text2image:5c347a4bfa1d4523a58ae614c2194e15f2ae682b57e3797a5bb468920aa70ebf",
    input={
        "drawer": "vqgan",
        "prompts": "Manhattan skyline at sunset. #artstation 🌇",
        "settings": "\n"
    }
)

# The pixray/text2image model can stream output as it's running.
# The predict method returns an iterator, and you can iterate over that output.
for item in output:
    # https://replicate.com/pixray/text2image/api#output-schema
    print(item)

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 pixray/text2image 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": "5c347a4bfa1d4523a58ae614c2194e15f2ae682b57e3797a5bb468920aa70ebf",
    "input": {
      "drawer": "vqgan",
      "prompts": "Manhattan skyline at sunset. #artstation 🌇",
      "settings": "\\n"
    }
  }' \
  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/pixray/text2image@sha256:5c347a4bfa1d4523a58ae614c2194e15f2ae682b57e3797a5bb468920aa70ebf \
  -i 'drawer="vqgan"' \
  -i 'prompts="Manhattan skyline at sunset. #artstation 🌇"' \
  -i $'settings="\\n"'

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/pixray/text2image@sha256:5c347a4bfa1d4523a58ae614c2194e15f2ae682b57e3797a5bb468920aa70ebf
curl -s -X POST \
  -H "Content-Type: application/json" \
  -d $'{
    "input": {
      "drawer": "vqgan",
      "prompts": "Manhattan skyline at sunset. #artstation 🌇",
      "settings": "\\n"
    }
  }' \
  http://localhost:5000/predictions

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

Output

{
  "completed_at": "2022-01-04T06:40:21Z",
  "created_at": "2022-01-04T06:35:16.453894Z",
  "data_removed": false,
  "error": "",
  "id": "bpmogp2wprabndhuyyu73522gi",
  "input": {
    "prompts": "Manhattan skyline at sunset. #artstation 🌇",
    "settings": "\n"
  },
  "logs": "---> BasePixrayPredictor Predict\r\nUsing seed:\r\n10184741873048389411\r\nUsing device:\r\ncuda:0\r\nOptimising using:\r\nAdam\r\nUsing text prompts:\r\n['Manhattan skyline at sunset. #artstation 🌇']\r\n\r\n0it [00:00, ?it/s]\r\n/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/nn/functional.py:3609: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.\r\n  warnings.warn(\r\niter: 0, loss: 3.15, losses: 1.06, 0.0803, 0.967, 0.0479, 0.951, 0.0481 (-0=>3.154)\r\n\r\n0it [00:01, ?it/s]\r\n/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/nn/functional.py:3609: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.\r\n  warnings.warn(\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 10, loss: 3.09, losses: 1.07, 0.0771, 0.942, 0.0435, 0.917, 0.0449 (-0=>3.093)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 20, loss: 3.09, losses: 1.06, 0.0797, 0.929, 0.0453, 0.928, 0.0467 (-4=>3.032)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 30, loss: 2.96, losses: 1.01, 0.0821, 0.889, 0.0488, 0.886, 0.0474 (-0=>2.959)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 40, loss: 2.94, losses:  1, 0.0814, 0.882, 0.0487, 0.872, 0.0482 (-1=>2.891)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 50, loss: 2.85, losses: 0.975, 0.0859, 0.85, 0.0508, 0.841, 0.0474 (-4=>2.83)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 60, loss: 2.74, losses: 0.943, 0.0875, 0.813, 0.0516, 0.797, 0.0477 (-0=>2.74)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 70, loss: 2.72, losses: 0.945, 0.085, 0.804, 0.0501, 0.789, 0.0481 (-4=>2.692)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 80, loss: 2.64, losses: 0.895, 0.0883, 0.785, 0.0516, 0.769, 0.0486 (-2=>2.636)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 90, loss: 2.61, losses: 0.884, 0.0883, 0.778, 0.0509, 0.759, 0.0484 (-1=>2.597)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 100, loss: 2.63, losses: 0.899, 0.0887, 0.781, 0.0492, 0.764, 0.0486 (-11=>2.597)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 110, loss: 2.6, losses: 0.884, 0.0895, 0.771, 0.0495, 0.753, 0.049 (-5=>2.559)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 120, loss: 2.56, losses: 0.865, 0.0886, 0.76, 0.0495, 0.746, 0.0485 (-2=>2.557)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 130, loss: 2.6, losses: 0.884, 0.0881, 0.771, 0.049, 0.758, 0.0481 (-7=>2.556)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 140, loss: 2.58, losses: 0.875, 0.0883, 0.765, 0.0488, 0.755, 0.0479 (-6=>2.519)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 150, loss: 2.58, losses: 0.87, 0.0887, 0.764, 0.0487, 0.761, 0.048 (-1=>2.514)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 160, loss: 2.53, losses: 0.841, 0.09, 0.756, 0.0504, 0.75, 0.0483 (-11=>2.514)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 170, loss: 2.57, losses: 0.862, 0.089, 0.764, 0.0498, 0.752, 0.0488 (-1=>2.513)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 180, loss: 2.55, losses: 0.859, 0.0889, 0.758, 0.0493, 0.747, 0.0488 (-11=>2.513)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 190, loss: 2.56, losses: 0.862, 0.0878, 0.763, 0.0491, 0.755, 0.0485 (-21=>2.513)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 200, loss: 2.58, losses: 0.875, 0.0885, 0.765, 0.0484, 0.76, 0.0482 (-31=>2.513)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 210, loss: 2.58, losses: 0.87, 0.0894, 0.765, 0.0493, 0.757, 0.0489 (-5=>2.508)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 220, loss: 2.58, losses: 0.867, 0.0883, 0.765, 0.0491, 0.762, 0.049 (-15=>2.508)\r\n\r\n0it [00:00, ?it/s]\r\nDropping learning rate\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 230, loss: 2.53, losses: 0.846, 0.0904, 0.748, 0.05, 0.748, 0.0493 (-1=>2.506)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:10, ?it/s]\r\n\r\n0it [00:00, ?it/s]\r\niter: 240, loss: 2.59, losses: 0.88, 0.088, 0.764, 0.0487, 0.761, 0.0482 (-11=>2.506)\r\n\r\n0it [00:00, ?it/s]\r\n\r\n0it [00:03, ?it/s]\r\nTraceback (most recent call last):\r\n  File \"/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/cog/server/redis_queue.py\", line 163, in start\r\n    self.handle_message(response_queue, message, cleanup_functions)\r\n  File \"/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/cog/server/redis_queue.py\", line 232, in handle_message\r\n    result = next(return_value)\r\n  File \"/src/cogrun.py\", line 132, in predict\r\n    yield from super().predict(settings=\"pixray_vdiff\", prompts=prompts, **ydict)\r\n  File \"/src/cogrun.py\", line 48, in predict\r\n    run_complete = pixray.do_run(settings, return_display=True)\r\n  File \"/src/pixray.py\", line 1502, in do_run\r\n    keep_going = train(args, cur_iteration)\r\n  File \"/src/pixray.py\", line 1367, in train\r\n    loss.backward()\r\n  File \"/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/_tensor.py\", line 255, in backward\r\n    torch.autograd.backward(self, gradient, retain_graph, create_graph, inputs=inputs)\r\n  File \"/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/autograd/__init__.py\", line 147, in backward\r\n    Variable._execution_engine.run_backward(\r\n  File \"/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/cog/server/redis_queue.py\", line 35, in handle_timeout\r\n    raise TimeoutError(self.error_message)\r\nTimeoutError: Prediction timed out",
  "metrics": {
    "predict_time": 301,
    "total_time": 304.546106
  },
  "output": [
    {
      "file": "https://replicate.delivery/mgxm/7d3e3aa2-9859-4679-9755-5c8c166c9aae/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/ecf18aee-cdd9-49c7-adf9-71b59a04ea56/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/af7c0e33-e940-4ad1-84ef-aac1f7471653/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/9b1e957a-06d2-412f-81ef-4c95a7ea0f75/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/caa6cbf1-2c7e-4955-ba52-8dd96980f476/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/26418e22-c5dd-4903-9c0a-d61fdd2dd681/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/179edad9-4f0c-41ff-a274-a6a169b842d2/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/a0a49e7f-69bb-4fe2-9608-a746841140ee/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/88a80201-135d-4d05-95fd-b2815edbe50c/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/b8a701f5-c676-423e-917b-6b70c7eb9955/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/4467a54f-6afc-4989-b9c3-985774d8839d/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/0055e7f7-0bdf-4feb-bd89-cf7d74a00233/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/02b0ee4d-e586-4100-bb9f-8cca1a2921f9/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/86873983-d611-480a-a2e9-dbd937b2bdfb/tempfile.png"
    },
    {
      "file": "https://replicate.delivery/mgxm/4eb7ad94-3330-470e-8b3f-5b120bfce2e3/tempfile.png"
    }
  ],
  "started_at": "2022-01-04T06:35:20Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/bpmogp2wprabndhuyyu73522gi",
    "cancel": "https://api.replicate.com/v1/predictions/bpmogp2wprabndhuyyu73522gi/cancel"
  },
  "version": "b18734685c2c41ff3ec4e1b6f8fcee1c3f58f8cdbcd11caecfc92a2232d6aeca"
}

Generated in

5 minutes

Tweak itReport

---> BasePixrayPredictor Predict
Using seed:
10184741873048389411
Using device:
cuda:0
Optimising using:
Adam
Using text prompts:
['Manhattan skyline at sunset. #artstation 🌇']

0it [00:00, ?it/s]
/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/nn/functional.py:3609: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.
  warnings.warn(
iter: 0, loss: 3.15, losses: 1.06, 0.0803, 0.967, 0.0479, 0.951, 0.0481 (-0=>3.154)

0it [00:01, ?it/s]
/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/nn/functional.py:3609: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.
  warnings.warn(

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 10, loss: 3.09, losses: 1.07, 0.0771, 0.942, 0.0435, 0.917, 0.0449 (-0=>3.093)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 20, loss: 3.09, losses: 1.06, 0.0797, 0.929, 0.0453, 0.928, 0.0467 (-4=>3.032)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 30, loss: 2.96, losses: 1.01, 0.0821, 0.889, 0.0488, 0.886, 0.0474 (-0=>2.959)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 40, loss: 2.94, losses:  1, 0.0814, 0.882, 0.0487, 0.872, 0.0482 (-1=>2.891)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 50, loss: 2.85, losses: 0.975, 0.0859, 0.85, 0.0508, 0.841, 0.0474 (-4=>2.83)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 60, loss: 2.74, losses: 0.943, 0.0875, 0.813, 0.0516, 0.797, 0.0477 (-0=>2.74)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 70, loss: 2.72, losses: 0.945, 0.085, 0.804, 0.0501, 0.789, 0.0481 (-4=>2.692)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 80, loss: 2.64, losses: 0.895, 0.0883, 0.785, 0.0516, 0.769, 0.0486 (-2=>2.636)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 90, loss: 2.61, losses: 0.884, 0.0883, 0.778, 0.0509, 0.759, 0.0484 (-1=>2.597)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 100, loss: 2.63, losses: 0.899, 0.0887, 0.781, 0.0492, 0.764, 0.0486 (-11=>2.597)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 110, loss: 2.6, losses: 0.884, 0.0895, 0.771, 0.0495, 0.753, 0.049 (-5=>2.559)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 120, loss: 2.56, losses: 0.865, 0.0886, 0.76, 0.0495, 0.746, 0.0485 (-2=>2.557)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 130, loss: 2.6, losses: 0.884, 0.0881, 0.771, 0.049, 0.758, 0.0481 (-7=>2.556)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 140, loss: 2.58, losses: 0.875, 0.0883, 0.765, 0.0488, 0.755, 0.0479 (-6=>2.519)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 150, loss: 2.58, losses: 0.87, 0.0887, 0.764, 0.0487, 0.761, 0.048 (-1=>2.514)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 160, loss: 2.53, losses: 0.841, 0.09, 0.756, 0.0504, 0.75, 0.0483 (-11=>2.514)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 170, loss: 2.57, losses: 0.862, 0.089, 0.764, 0.0498, 0.752, 0.0488 (-1=>2.513)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 180, loss: 2.55, losses: 0.859, 0.0889, 0.758, 0.0493, 0.747, 0.0488 (-11=>2.513)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 190, loss: 2.56, losses: 0.862, 0.0878, 0.763, 0.0491, 0.755, 0.0485 (-21=>2.513)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 200, loss: 2.58, losses: 0.875, 0.0885, 0.765, 0.0484, 0.76, 0.0482 (-31=>2.513)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 210, loss: 2.58, losses: 0.87, 0.0894, 0.765, 0.0493, 0.757, 0.0489 (-5=>2.508)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 220, loss: 2.58, losses: 0.867, 0.0883, 0.765, 0.0491, 0.762, 0.049 (-15=>2.508)

0it [00:00, ?it/s]
Dropping learning rate

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 230, loss: 2.53, losses: 0.846, 0.0904, 0.748, 0.05, 0.748, 0.0493 (-1=>2.506)

0it [00:00, ?it/s]

0it [00:10, ?it/s]

0it [00:00, ?it/s]
iter: 240, loss: 2.59, losses: 0.88, 0.088, 0.764, 0.0487, 0.761, 0.0482 (-11=>2.506)

0it [00:00, ?it/s]

0it [00:03, ?it/s]
Traceback (most recent call last):
  File "/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/cog/server/redis_queue.py", line 163, in start
    self.handle_message(response_queue, message, cleanup_functions)
  File "/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/cog/server/redis_queue.py", line 232, in handle_message
    result = next(return_value)
  File "/src/cogrun.py", line 132, in predict
    yield from super().predict(settings="pixray_vdiff", prompts=prompts, **ydict)
  File "/src/cogrun.py", line 48, in predict
    run_complete = pixray.do_run(settings, return_display=True)
  File "/src/pixray.py", line 1502, in do_run
    keep_going = train(args, cur_iteration)
  File "/src/pixray.py", line 1367, in train
    loss.backward()
  File "/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/_tensor.py", line 255, in backward
    torch.autograd.backward(self, gradient, retain_graph, create_graph, inputs=inputs)
  File "/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/torch/autograd/__init__.py", line 147, in backward
    Variable._execution_engine.run_backward(
  File "/root/.pyenv/versions/3.8.12/lib/python3.8/site-packages/cog/server/redis_queue.py", line 35, in handle_timeout
    raise TimeoutError(self.error_message)
TimeoutError: Prediction timed out

This example was created by a different version, pixray/text2image:b1873468.

Examples

View more examples

Run time and cost

This model costs approximately $0.095 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 8 minutes.

Readme

Alt text

Pixray is an image generation system. It combines previous ideas including:

Perception Engines which uses image augmentation and iteratively optimises images against an ensemble of classifiers
CLIP guided GAN imagery from Ryan Murdoch and Katherine Crowson as well as modifictions such as CLIPDraw from Kevin Frans
Useful ways of navigating latent space from Sampling Generative Networks
(more to come)

pixray is also a python library and command line utility. Take a look at the GitHub repository for more details.

There is currently some documentation on options. Also checkout THE DEMO NOTEBOOKS or join in the discussion on discord.