tstramer/ghibli-diffusion – Run with an API on Replicate

tstramer / ghibli-diffusion

Cold

Public
44.6K runs
A100 (80GB)

Run with an API

Playground API Examples README Versions

Input

prompt

string

Shift + Return to add a new line

Input prompt

Default: "a photo of an astronaut riding a horse on mars"

negative_prompt

string

Shift + Return to add a new line

Specify things to not see in the output

width

integer

Width of output image. Maximum size is 1024x768 or 768x1024 because of memory limits

Default: 768

height

integer

Height of output image. Maximum size is 1024x768 or 768x1024 because of memory limits

Default: 768

prompt_strength

number

Prompt strength when using init image. 1.0 corresponds to full destruction of information in init image

Default: 0.8

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: 50

guidance_scale

number

(minimum: 1, maximum: 20)

Scale for classifier-free guidance

Default: 7.5

scheduler

string

Choose a scheduler.

Default: "DPMSolverMultistep"

seed

integer

Random seed. Leave blank to randomize the seed

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

npx create-replicate --model=tstramer/ghibli-diffusion

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 tstramer/ghibli-diffusion using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "tstramer/ghibli-diffusion:b76aa203ed8b55a2bd69bacdab46b48b981984181476570b7ba75699ae286025",
  {
    input: {
      width: 512,
      height: 512,
      prompt: "ghibli style elf",
      scheduler: "K-LMS",
      num_outputs: 1,
      guidance_scale: 7.5,
      prompt_strength: 0.8,
      num_inference_steps: 150
    }
  }
);
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 tstramer/ghibli-diffusion using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "tstramer/ghibli-diffusion:b76aa203ed8b55a2bd69bacdab46b48b981984181476570b7ba75699ae286025",
    input={
        "width": 512,
        "height": 512,
        "prompt": "ghibli style elf",
        "scheduler": "K-LMS",
        "num_outputs": 1,
        "guidance_scale": 7.5,
        "prompt_strength": 0.8,
        "num_inference_steps": 150
    }
)
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 tstramer/ghibli-diffusion 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": "b76aa203ed8b55a2bd69bacdab46b48b981984181476570b7ba75699ae286025",
    "input": {
      "width": 512,
      "height": 512,
      "prompt": "ghibli style elf",
      "scheduler": "K-LMS",
      "num_outputs": 1,
      "guidance_scale": 7.5,
      "prompt_strength": 0.8,
      "num_inference_steps": 150
    }
  }' \
  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/tstramer/ghibli-diffusion@sha256:b76aa203ed8b55a2bd69bacdab46b48b981984181476570b7ba75699ae286025 \
  -i 'width=512' \
  -i 'height=512' \
  -i 'prompt="ghibli style elf"' \
  -i 'scheduler="K-LMS"' \
  -i 'num_outputs=1' \
  -i 'guidance_scale=7.5' \
  -i 'prompt_strength=0.8' \
  -i 'num_inference_steps=150'

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/tstramer/ghibli-diffusion@sha256:b76aa203ed8b55a2bd69bacdab46b48b981984181476570b7ba75699ae286025
curl -s -X POST \
  -H "Content-Type: application/json" \
  -d $'{
    "input": {
      "width": 512,
      "height": 512,
      "prompt": "ghibli style elf",
      "scheduler": "K-LMS",
      "num_outputs": 1,
      "guidance_scale": 7.5,
      "prompt_strength": 0.8,
      "num_inference_steps": 150
    }
  }' \
  http://localhost:5000/predictions

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

Output

{
  "completed_at": "2022-11-22T02:40:27.132264Z",
  "created_at": "2022-11-22T02:40:15.884998Z",
  "data_removed": false,
  "error": null,
  "id": "vuj42xsehbgcbld4gdy7o5754m",
  "input": {
    "width": 512,
    "height": 512,
    "prompt": "ghibli style elf",
    "scheduler": "K-LMS",
    "num_outputs": 1,
    "guidance_scale": 7.5,
    "prompt_strength": 0.8,
    "num_inference_steps": "150"
  },
  "logs": "Using seed: 14453\n  0%|          | 0/150 [00:00<?, ?it/s]\n  1%|▏         | 2/150 [00:00<00:11, 13.36it/s]\n  3%|▎         | 4/150 [00:00<00:10, 13.77it/s]\n  4%|▍         | 6/150 [00:00<00:10, 13.96it/s]\n  5%|▌         | 8/150 [00:00<00:10, 13.47it/s]\n  7%|▋         | 10/150 [00:00<00:10, 13.69it/s]\n  8%|▊         | 12/150 [00:00<00:09, 13.91it/s]\n  9%|▉         | 14/150 [00:01<00:09, 13.96it/s]\n 11%|█         | 16/150 [00:01<00:09, 13.98it/s]\n 12%|█▏        | 18/150 [00:01<00:09, 14.10it/s]\n 13%|█▎        | 20/150 [00:01<00:09, 14.09it/s]\n 15%|█▍        | 22/150 [00:01<00:09, 14.07it/s]\n 16%|█▌        | 24/150 [00:01<00:09, 13.97it/s]\n 17%|█▋        | 26/150 [00:01<00:08, 14.00it/s]\n 19%|█▊        | 28/150 [00:02<00:08, 14.02it/s]\n 20%|██        | 30/150 [00:02<00:08, 13.99it/s]\n 21%|██▏       | 32/150 [00:02<00:08, 14.00it/s]\n 23%|██▎       | 34/150 [00:02<00:08, 14.05it/s]\n 24%|██▍       | 36/150 [00:02<00:08, 14.07it/s]\n 25%|██▌       | 38/150 [00:02<00:08, 13.94it/s]\n 27%|██▋       | 40/150 [00:02<00:07, 14.00it/s]\n 28%|██▊       | 42/150 [00:03<00:07, 14.06it/s]\n 29%|██▉       | 44/150 [00:03<00:07, 14.12it/s]\n 31%|███       | 46/150 [00:03<00:07, 14.21it/s]\n 32%|███▏      | 48/150 [00:03<00:07, 14.30it/s]\n 33%|███▎      | 50/150 [00:03<00:06, 14.35it/s]\n 35%|███▍      | 52/150 [00:03<00:06, 14.24it/s]\n 36%|███▌      | 54/150 [00:03<00:06, 14.22it/s]\n 37%|███▋      | 56/150 [00:03<00:06, 14.20it/s]\n 39%|███▊      | 58/150 [00:04<00:06, 14.22it/s]\n 40%|████      | 60/150 [00:04<00:06, 14.25it/s]\n 41%|████▏     | 62/150 [00:04<00:06, 14.29it/s]\n 43%|████▎     | 64/150 [00:04<00:06, 14.26it/s]\n 44%|████▍     | 66/150 [00:04<00:06, 13.95it/s]\n 45%|████▌     | 68/150 [00:04<00:05, 13.86it/s]\n 47%|████▋     | 70/150 [00:04<00:05, 13.90it/s]\n 48%|████▊     | 72/150 [00:05<00:05, 14.06it/s]\n 49%|████▉     | 74/150 [00:05<00:05, 14.16it/s]\n 51%|█████     | 76/150 [00:05<00:05, 14.25it/s]\n 52%|█████▏    | 78/150 [00:05<00:05, 14.29it/s]\n 53%|█████▎    | 80/150 [00:05<00:04, 14.05it/s]\n 55%|█████▍    | 82/150 [00:05<00:04, 14.08it/s]\n 56%|█████▌    | 84/150 [00:05<00:04, 14.07it/s]\n 57%|█████▋    | 86/150 [00:06<00:04, 14.10it/s]\n 59%|█████▊    | 88/150 [00:06<00:04, 14.13it/s]\n 60%|██████    | 90/150 [00:06<00:04, 14.14it/s]\n 61%|██████▏   | 92/150 [00:06<00:04, 14.11it/s]\n 63%|██████▎   | 94/150 [00:06<00:04, 13.90it/s]\n 64%|██████▍   | 96/150 [00:06<00:03, 13.96it/s]\n 65%|██████▌   | 98/150 [00:06<00:03, 13.99it/s]\n 67%|██████▋   | 100/150 [00:07<00:03, 13.99it/s]\n 68%|██████▊   | 102/150 [00:07<00:03, 14.00it/s]\n 69%|██████▉   | 104/150 [00:07<00:03, 14.01it/s]\n 71%|███████   | 106/150 [00:07<00:03, 13.96it/s]\n 72%|███████▏  | 108/150 [00:07<00:03, 13.82it/s]\n 73%|███████▎  | 110/150 [00:07<00:02, 13.90it/s]\n 75%|███████▍  | 112/150 [00:07<00:02, 13.99it/s]\n 76%|███████▌  | 114/150 [00:08<00:02, 14.04it/s]\n 77%|███████▋  | 116/150 [00:08<00:02, 14.01it/s]\n 79%|███████▊  | 118/150 [00:08<00:02, 14.05it/s]\n 80%|████████  | 120/150 [00:08<00:02, 13.95it/s]\n 81%|████████▏ | 122/150 [00:08<00:02, 13.89it/s]\n 83%|████████▎ | 124/150 [00:08<00:01, 13.93it/s]\n 84%|████████▍ | 126/150 [00:08<00:01, 13.96it/s]\n 85%|████████▌ | 128/150 [00:09<00:01, 13.95it/s]\n 87%|████████▋ | 130/150 [00:09<00:01, 14.02it/s]\n 88%|████████▊ | 132/150 [00:09<00:01, 14.00it/s]\n 89%|████████▉ | 134/150 [00:09<00:01, 14.01it/s]\n 91%|█████████ | 136/150 [00:09<00:01, 13.88it/s]\n 92%|█████████▏| 138/150 [00:09<00:00, 13.77it/s]\n 93%|█████████▎| 140/150 [00:09<00:00, 13.72it/s]\n 95%|█████████▍| 142/150 [00:10<00:00, 13.74it/s]\n 96%|█████████▌| 144/150 [00:10<00:00, 13.76it/s]\n 97%|█████████▋| 146/150 [00:10<00:00, 13.79it/s]\n 99%|█████████▊| 148/150 [00:10<00:00, 13.81it/s]\n100%|██████████| 150/150 [00:10<00:00, 13.62it/s]\n100%|██████████| 150/150 [00:10<00:00, 13.99it/s]",
  "metrics": {
    "predict_time": 11.211884,
    "total_time": 11.247266
  },
  "output": [
    "https://replicate.delivery/pbxt/EBBNVwxeuQwQZKbuecO6mMgSQ11l1D715fvMrEra3A50E9EgA/out-0.png"
  ],
  "started_at": "2022-11-22T02:40:15.920380Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/vuj42xsehbgcbld4gdy7o5754m",
    "cancel": "https://api.replicate.com/v1/predictions/vuj42xsehbgcbld4gdy7o5754m/cancel"
  },
  "version": "b190224e487827d5fdeb8bf8bcdaa71c41b09a152f175de9f170aa0e088f66e9"
}

Generated in

11.2 seconds

Tweak itReport

Using seed: 14453
  0%|          | 0/150 [00:00<?, ?it/s]
  1%|▏         | 2/150 [00:00<00:11, 13.36it/s]
  3%|▎         | 4/150 [00:00<00:10, 13.77it/s]
  4%|▍         | 6/150 [00:00<00:10, 13.96it/s]
  5%|▌         | 8/150 [00:00<00:10, 13.47it/s]
  7%|▋         | 10/150 [00:00<00:10, 13.69it/s]
  8%|▊         | 12/150 [00:00<00:09, 13.91it/s]
  9%|▉         | 14/150 [00:01<00:09, 13.96it/s]
 11%|█         | 16/150 [00:01<00:09, 13.98it/s]
 12%|█▏        | 18/150 [00:01<00:09, 14.10it/s]
 13%|█▎        | 20/150 [00:01<00:09, 14.09it/s]
 15%|█▍        | 22/150 [00:01<00:09, 14.07it/s]
 16%|█▌        | 24/150 [00:01<00:09, 13.97it/s]
 17%|█▋        | 26/150 [00:01<00:08, 14.00it/s]
 19%|█▊        | 28/150 [00:02<00:08, 14.02it/s]
 20%|██        | 30/150 [00:02<00:08, 13.99it/s]
 21%|██▏       | 32/150 [00:02<00:08, 14.00it/s]
 23%|██▎       | 34/150 [00:02<00:08, 14.05it/s]
 24%|██▍       | 36/150 [00:02<00:08, 14.07it/s]
 25%|██▌       | 38/150 [00:02<00:08, 13.94it/s]
 27%|██▋       | 40/150 [00:02<00:07, 14.00it/s]
 28%|██▊       | 42/150 [00:03<00:07, 14.06it/s]
 29%|██▉       | 44/150 [00:03<00:07, 14.12it/s]
 31%|███       | 46/150 [00:03<00:07, 14.21it/s]
 32%|███▏      | 48/150 [00:03<00:07, 14.30it/s]
 33%|███▎      | 50/150 [00:03<00:06, 14.35it/s]
 35%|███▍      | 52/150 [00:03<00:06, 14.24it/s]
 36%|███▌      | 54/150 [00:03<00:06, 14.22it/s]
 37%|███▋      | 56/150 [00:03<00:06, 14.20it/s]
 39%|███▊      | 58/150 [00:04<00:06, 14.22it/s]
 40%|████      | 60/150 [00:04<00:06, 14.25it/s]
 41%|████▏     | 62/150 [00:04<00:06, 14.29it/s]
 43%|████▎     | 64/150 [00:04<00:06, 14.26it/s]
 44%|████▍     | 66/150 [00:04<00:06, 13.95it/s]
 45%|████▌     | 68/150 [00:04<00:05, 13.86it/s]
 47%|████▋     | 70/150 [00:04<00:05, 13.90it/s]
 48%|████▊     | 72/150 [00:05<00:05, 14.06it/s]
 49%|████▉     | 74/150 [00:05<00:05, 14.16it/s]
 51%|█████     | 76/150 [00:05<00:05, 14.25it/s]
 52%|█████▏    | 78/150 [00:05<00:05, 14.29it/s]
 53%|█████▎    | 80/150 [00:05<00:04, 14.05it/s]
 55%|█████▍    | 82/150 [00:05<00:04, 14.08it/s]
 56%|█████▌    | 84/150 [00:05<00:04, 14.07it/s]
 57%|█████▋    | 86/150 [00:06<00:04, 14.10it/s]
 59%|█████▊    | 88/150 [00:06<00:04, 14.13it/s]
 60%|██████    | 90/150 [00:06<00:04, 14.14it/s]
 61%|██████▏   | 92/150 [00:06<00:04, 14.11it/s]
 63%|██████▎   | 94/150 [00:06<00:04, 13.90it/s]
 64%|██████▍   | 96/150 [00:06<00:03, 13.96it/s]
 65%|██████▌   | 98/150 [00:06<00:03, 13.99it/s]
 67%|██████▋   | 100/150 [00:07<00:03, 13.99it/s]
 68%|██████▊   | 102/150 [00:07<00:03, 14.00it/s]
 69%|██████▉   | 104/150 [00:07<00:03, 14.01it/s]
 71%|███████   | 106/150 [00:07<00:03, 13.96it/s]
 72%|███████▏  | 108/150 [00:07<00:03, 13.82it/s]
 73%|███████▎  | 110/150 [00:07<00:02, 13.90it/s]
 75%|███████▍  | 112/150 [00:07<00:02, 13.99it/s]
 76%|███████▌  | 114/150 [00:08<00:02, 14.04it/s]
 77%|███████▋  | 116/150 [00:08<00:02, 14.01it/s]
 79%|███████▊  | 118/150 [00:08<00:02, 14.05it/s]
 80%|████████  | 120/150 [00:08<00:02, 13.95it/s]
 81%|████████▏ | 122/150 [00:08<00:02, 13.89it/s]
 83%|████████▎ | 124/150 [00:08<00:01, 13.93it/s]
 84%|████████▍ | 126/150 [00:08<00:01, 13.96it/s]
 85%|████████▌ | 128/150 [00:09<00:01, 13.95it/s]
 87%|████████▋ | 130/150 [00:09<00:01, 14.02it/s]
 88%|████████▊ | 132/150 [00:09<00:01, 14.00it/s]
 89%|████████▉ | 134/150 [00:09<00:01, 14.01it/s]
 91%|█████████ | 136/150 [00:09<00:01, 13.88it/s]
 92%|█████████▏| 138/150 [00:09<00:00, 13.77it/s]
 93%|█████████▎| 140/150 [00:09<00:00, 13.72it/s]
 95%|█████████▍| 142/150 [00:10<00:00, 13.74it/s]
 96%|█████████▌| 144/150 [00:10<00:00, 13.76it/s]
 97%|█████████▋| 146/150 [00:10<00:00, 13.79it/s]
 99%|█████████▊| 148/150 [00:10<00:00, 13.81it/s]
100%|██████████| 150/150 [00:10<00:00, 13.62it/s]
100%|██████████| 150/150 [00:10<00:00, 13.99it/s]

This example was created by a different version, tstramer/ghibli-diffusion:b190224e.

Examples

View more examples

Run time and cost

This model costs approximately $0.16 to run on Replicate, or 6 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 115 seconds. The predict time for this model varies significantly based on the inputs.

Readme

…

Model description

…

Intended use

…

Ethical considerations

…

Caveats and recommendations

…