suno-ai / bark

🔊 Text-Prompted Generative Audio Model

Cold

Public
297.5K runs
T4
GitHub
License

Run with an API

Playground API Examples README Versions

Input

prompt

string

Shift + Return to add a new line

Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe.Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe.

Input prompt

Default: "Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe."

history_prompt

string

history choice for audio cloning, choose from the list

custom_history_prompt

file

Provide your own .npz file with history choice for audio cloning, this will override the previous history_prompt setting

text_temp

number

generation temperature (1.0 more diverse, 0.0 more conservative)

Default: 0.7

waveform_temp

number

generation temperature (1.0 more diverse, 0.0 more conservative)

Default: 0.7

output_full

boolean

return full generation as a .npz file to be used as a history prompt

Default: false

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

npx create-replicate --model=suno-ai/bark

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 suno-ai/bark using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

const output = await replicate.run(
  "suno-ai/bark:b76242b40d67c76ab6742e987628a2a9ac019e11d56ab96c4e91ce03b79b2787",
  {
    input: {
      prompt: "Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe.",
      text_temp: 0.7,
      output_full: false,
      waveform_temp: 0.7,
      history_prompt: "announcer"
    }
  }
);
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 suno-ai/bark using Replicate’s API. Check out the model's schema for an overview of inputs and outputs.

output = replicate.run(
    "suno-ai/bark:b76242b40d67c76ab6742e987628a2a9ac019e11d56ab96c4e91ce03b79b2787",
    input={
        "prompt": "Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe.",
        "text_temp": 0.7,
        "output_full": False,
        "waveform_temp": 0.7,
        "history_prompt": "announcer"
    }
)
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 suno-ai/bark 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": "b76242b40d67c76ab6742e987628a2a9ac019e11d56ab96c4e91ce03b79b2787",
    "input": {
      "prompt": "Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe.",
      "text_temp": 0.7,
      "output_full": false,
      "waveform_temp": 0.7,
      "history_prompt": "announcer"
    }
  }' \
  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/suno-ai/bark@sha256:b76242b40d67c76ab6742e987628a2a9ac019e11d56ab96c4e91ce03b79b2787 \
  -i 'prompt="Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe."' \
  -i 'text_temp=0.7' \
  -i 'output_full=false' \
  -i 'waveform_temp=0.7' \
  -i 'history_prompt="announcer"'

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/suno-ai/bark@sha256:b76242b40d67c76ab6742e987628a2a9ac019e11d56ab96c4e91ce03b79b2787
curl -s -X POST \
  -H "Content-Type: application/json" \
  -d $'{
    "input": {
      "prompt": "Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe.",
      "text_temp": 0.7,
      "output_full": false,
      "waveform_temp": 0.7,
      "history_prompt": "announcer"
    }
  }' \
  http://localhost:5000/predictions

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

Output

Video Player is loading.

Current Time 00:00:000

Duration 00:00:000

Loaded: 0%

Stream Type LIVE

Remaining Time 00:00:000

{
  "completed_at": "2023-04-25T22:18:59.625638Z",
  "created_at": "2023-04-25T22:11:26.774980Z",
  "data_removed": false,
  "error": null,
  "id": "ngk3yp5omvdsvkcdoljxs2m4ra",
  "input": {
    "prompt": "Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe."
  },
  "logs": "0%|          | 0/100 [00:00<?, ?it/s]\n  1%|          | 1/100 [00:00<00:27,  3.65it/s]\n  3%|▎         | 3/100 [00:00<00:13,  7.12it/s]\n  5%|▌         | 5/100 [00:00<00:10,  8.90it/s]\n  7%|▋         | 7/100 [00:00<00:09,  9.98it/s]\n  9%|▉         | 9/100 [00:00<00:08, 10.54it/s]\n 11%|█         | 11/100 [00:01<00:08, 11.09it/s]\n 13%|█▎        | 13/100 [00:01<00:07, 11.16it/s]\n 15%|█▌        | 15/100 [00:01<00:07, 11.32it/s]\n 17%|█▋        | 17/100 [00:01<00:07, 11.66it/s]\n 19%|█▉        | 19/100 [00:01<00:06, 11.69it/s]\n 21%|██        | 21/100 [00:01<00:06, 12.05it/s]\n 23%|██▎       | 23/100 [00:02<00:06, 12.17it/s]\n 25%|██▌       | 25/100 [00:02<00:06, 11.89it/s]\n 27%|██▋       | 27/100 [00:02<00:06, 11.64it/s]\n 29%|██▉       | 29/100 [00:02<00:06, 11.51it/s]\n 31%|███       | 31/100 [00:02<00:05, 11.61it/s]\n 33%|███▎      | 33/100 [00:02<00:05, 11.88it/s]\n 35%|███▌      | 35/100 [00:03<00:05, 11.82it/s]\n 37%|███▋      | 37/100 [00:03<00:05, 11.75it/s]\n 39%|███▉      | 39/100 [00:03<00:05, 11.63it/s]\n 41%|████      | 41/100 [00:03<00:05, 11.64it/s]\n 43%|████▎     | 43/100 [00:03<00:04, 11.79it/s]\n 45%|████▌     | 45/100 [00:04<00:04, 11.94it/s]\n 47%|████▋     | 47/100 [00:04<00:04, 11.83it/s]\n 49%|████▉     | 49/100 [00:04<00:04, 11.89it/s]\n 51%|█████     | 51/100 [00:04<00:04, 11.70it/s]\n 53%|█████▎    | 53/100 [00:04<00:04, 11.56it/s]\n 55%|█████▌    | 55/100 [00:04<00:03, 11.64it/s]\n 57%|█████▋    | 57/100 [00:05<00:03, 11.63it/s]\n 59%|█████▉    | 59/100 [00:05<00:03, 11.38it/s]\n 61%|██████    | 61/100 [00:05<00:03, 11.38it/s]\n 63%|██████▎   | 63/100 [00:05<00:03, 11.07it/s]\n 65%|██████▌   | 65/100 [00:05<00:03, 11.19it/s]\n 67%|██████▋   | 67/100 [00:05<00:02, 11.28it/s]\n 69%|██████▉   | 69/100 [00:06<00:02, 11.08it/s]\n 71%|███████   | 71/100 [00:06<00:02, 11.11it/s]\n 73%|███████▎  | 73/100 [00:06<00:02, 10.91it/s]\n 75%|███████▌  | 75/100 [00:06<00:02, 10.78it/s]\n 77%|███████▋  | 77/100 [00:06<00:02, 10.83it/s]\n 79%|███████▉  | 79/100 [00:07<00:01, 10.86it/s]\n 81%|████████  | 81/100 [00:07<00:01, 10.64it/s]\n 83%|████████▎ | 83/100 [00:07<00:01, 10.67it/s]\n 85%|████████▌ | 85/100 [00:07<00:01, 10.57it/s]\n 87%|████████▋ | 87/100 [00:07<00:01, 10.34it/s]\n 89%|████████▉ | 89/100 [00:08<00:01, 10.39it/s]\n 91%|█████████ | 91/100 [00:08<00:00, 10.10it/s]\n 93%|█████████▎| 93/100 [00:08<00:00, 10.06it/s]\n100%|██████████| 100/100 [00:08<00:00, 20.32it/s]\n100%|██████████| 100/100 [00:08<00:00, 11.69it/s]\n  0%|          | 0/36 [00:00<?, ?it/s]\n  3%|▎         | 1/36 [00:00<00:23,  1.48it/s]\n  6%|▌         | 2/36 [00:01<00:22,  1.48it/s]\n  8%|▊         | 3/36 [00:02<00:22,  1.45it/s]\n 11%|█         | 4/36 [00:02<00:22,  1.43it/s]\n 14%|█▍        | 5/36 [00:03<00:21,  1.41it/s]\n 17%|█▋        | 6/36 [00:04<00:22,  1.36it/s]\n 19%|█▉        | 7/36 [00:05<00:21,  1.34it/s]\n 22%|██▏       | 8/36 [00:05<00:21,  1.30it/s]\n 25%|██▌       | 9/36 [00:06<00:21,  1.25it/s]\n 28%|██▊       | 10/36 [00:07<00:21,  1.23it/s]\n 31%|███       | 11/36 [00:08<00:21,  1.19it/s]\n 33%|███▎      | 12/36 [00:09<00:20,  1.15it/s]\n 36%|███▌      | 13/36 [00:10<00:20,  1.13it/s]\n 39%|███▉      | 14/36 [00:11<00:19,  1.12it/s]\n 42%|████▏     | 15/36 [00:12<00:18,  1.11it/s]\n 44%|████▍     | 16/36 [00:13<00:18,  1.10it/s]\n 47%|████▋     | 17/36 [00:14<00:17,  1.10it/s]\n 50%|█████     | 18/36 [00:14<00:16,  1.10it/s]\n 53%|█████▎    | 19/36 [00:15<00:15,  1.09it/s]\n 56%|█████▌    | 20/36 [00:16<00:14,  1.08it/s]\n 58%|█████▊    | 21/36 [00:17<00:13,  1.08it/s]\n 61%|██████    | 22/36 [00:18<00:12,  1.08it/s]\n 64%|██████▍   | 23/36 [00:19<00:11,  1.09it/s]\n 67%|██████▋   | 24/36 [00:20<00:11,  1.08it/s]\n 69%|██████▉   | 25/36 [00:21<00:10,  1.08it/s]\n 72%|███████▏  | 26/36 [00:22<00:09,  1.08it/s]\n 75%|███████▌  | 27/36 [00:23<00:08,  1.08it/s]\n 78%|███████▊  | 28/36 [00:24<00:07,  1.08it/s]\n 81%|████████  | 29/36 [00:25<00:06,  1.08it/s]\n 83%|████████▎ | 30/36 [00:26<00:05,  1.08it/s]\n 86%|████████▌ | 31/36 [00:26<00:04,  1.08it/s]\n 89%|████████▉ | 32/36 [00:27<00:03,  1.08it/s]\n 92%|█████████▏| 33/36 [00:28<00:02,  1.08it/s]\n 94%|█████████▍| 34/36 [00:29<00:01,  1.08it/s]\n 97%|█████████▋| 35/36 [00:30<00:00,  1.08it/s]\n100%|██████████| 36/36 [00:31<00:00,  1.08it/s]\n100%|██████████| 36/36 [00:31<00:00,  1.14it/s]",
  "metrics": {
    "predict_time": 44.949506,
    "total_time": 452.850658
  },
  "output": "https://replicate.delivery/pbxt/HuWYFtJyyH50BxruGu1XfUleB3kC2NfbTy2fmHbeEwKS6BsGC/audio.wav",
  "started_at": "2023-04-25T22:18:14.676132Z",
  "status": "succeeded",
  "urls": {
    "get": "https://api.replicate.com/v1/predictions/ngk3yp5omvdsvkcdoljxs2m4ra",
    "cancel": "https://api.replicate.com/v1/predictions/ngk3yp5omvdsvkcdoljxs2m4ra/cancel"
  },
  "version": "f23937d7c80b3c0f06c5a01ec55154388647292cb9398bd7d117678bc930791a"
}

Generated in

45.0 seconds

Tweak itReport

0%|          | 0/100 [00:00<?, ?it/s]
  1%|          | 1/100 [00:00<00:27,  3.65it/s]
  3%|▎         | 3/100 [00:00<00:13,  7.12it/s]
  5%|▌         | 5/100 [00:00<00:10,  8.90it/s]
  7%|▋         | 7/100 [00:00<00:09,  9.98it/s]
  9%|▉         | 9/100 [00:00<00:08, 10.54it/s]
 11%|█         | 11/100 [00:01<00:08, 11.09it/s]
 13%|█▎        | 13/100 [00:01<00:07, 11.16it/s]
 15%|█▌        | 15/100 [00:01<00:07, 11.32it/s]
 17%|█▋        | 17/100 [00:01<00:07, 11.66it/s]
 19%|█▉        | 19/100 [00:01<00:06, 11.69it/s]
 21%|██        | 21/100 [00:01<00:06, 12.05it/s]
 23%|██▎       | 23/100 [00:02<00:06, 12.17it/s]
 25%|██▌       | 25/100 [00:02<00:06, 11.89it/s]
 27%|██▋       | 27/100 [00:02<00:06, 11.64it/s]
 29%|██▉       | 29/100 [00:02<00:06, 11.51it/s]
 31%|███       | 31/100 [00:02<00:05, 11.61it/s]
 33%|███▎      | 33/100 [00:02<00:05, 11.88it/s]
 35%|███▌      | 35/100 [00:03<00:05, 11.82it/s]
 37%|███▋      | 37/100 [00:03<00:05, 11.75it/s]
 39%|███▉      | 39/100 [00:03<00:05, 11.63it/s]
 41%|████      | 41/100 [00:03<00:05, 11.64it/s]
 43%|████▎     | 43/100 [00:03<00:04, 11.79it/s]
 45%|████▌     | 45/100 [00:04<00:04, 11.94it/s]
 47%|████▋     | 47/100 [00:04<00:04, 11.83it/s]
 49%|████▉     | 49/100 [00:04<00:04, 11.89it/s]
 51%|█████     | 51/100 [00:04<00:04, 11.70it/s]
 53%|█████▎    | 53/100 [00:04<00:04, 11.56it/s]
 55%|█████▌    | 55/100 [00:04<00:03, 11.64it/s]
 57%|█████▋    | 57/100 [00:05<00:03, 11.63it/s]
 59%|█████▉    | 59/100 [00:05<00:03, 11.38it/s]
 61%|██████    | 61/100 [00:05<00:03, 11.38it/s]
 63%|██████▎   | 63/100 [00:05<00:03, 11.07it/s]
 65%|██████▌   | 65/100 [00:05<00:03, 11.19it/s]
 67%|██████▋   | 67/100 [00:05<00:02, 11.28it/s]
 69%|██████▉   | 69/100 [00:06<00:02, 11.08it/s]
 71%|███████   | 71/100 [00:06<00:02, 11.11it/s]
 73%|███████▎  | 73/100 [00:06<00:02, 10.91it/s]
 75%|███████▌  | 75/100 [00:06<00:02, 10.78it/s]
 77%|███████▋  | 77/100 [00:06<00:02, 10.83it/s]
 79%|███████▉  | 79/100 [00:07<00:01, 10.86it/s]
 81%|████████  | 81/100 [00:07<00:01, 10.64it/s]
 83%|████████▎ | 83/100 [00:07<00:01, 10.67it/s]
 85%|████████▌ | 85/100 [00:07<00:01, 10.57it/s]
 87%|████████▋ | 87/100 [00:07<00:01, 10.34it/s]
 89%|████████▉ | 89/100 [00:08<00:01, 10.39it/s]
 91%|█████████ | 91/100 [00:08<00:00, 10.10it/s]
 93%|█████████▎| 93/100 [00:08<00:00, 10.06it/s]
100%|██████████| 100/100 [00:08<00:00, 20.32it/s]
100%|██████████| 100/100 [00:08<00:00, 11.69it/s]
  0%|          | 0/36 [00:00<?, ?it/s]
  3%|▎         | 1/36 [00:00<00:23,  1.48it/s]
  6%|▌         | 2/36 [00:01<00:22,  1.48it/s]
  8%|▊         | 3/36 [00:02<00:22,  1.45it/s]
 11%|█         | 4/36 [00:02<00:22,  1.43it/s]
 14%|█▍        | 5/36 [00:03<00:21,  1.41it/s]
 17%|█▋        | 6/36 [00:04<00:22,  1.36it/s]
 19%|█▉        | 7/36 [00:05<00:21,  1.34it/s]
 22%|██▏       | 8/36 [00:05<00:21,  1.30it/s]
 25%|██▌       | 9/36 [00:06<00:21,  1.25it/s]
 28%|██▊       | 10/36 [00:07<00:21,  1.23it/s]
 31%|███       | 11/36 [00:08<00:21,  1.19it/s]
 33%|███▎      | 12/36 [00:09<00:20,  1.15it/s]
 36%|███▌      | 13/36 [00:10<00:20,  1.13it/s]
 39%|███▉      | 14/36 [00:11<00:19,  1.12it/s]
 42%|████▏     | 15/36 [00:12<00:18,  1.11it/s]
 44%|████▍     | 16/36 [00:13<00:18,  1.10it/s]
 47%|████▋     | 17/36 [00:14<00:17,  1.10it/s]
 50%|█████     | 18/36 [00:14<00:16,  1.10it/s]
 53%|█████▎    | 19/36 [00:15<00:15,  1.09it/s]
 56%|█████▌    | 20/36 [00:16<00:14,  1.08it/s]
 58%|█████▊    | 21/36 [00:17<00:13,  1.08it/s]
 61%|██████    | 22/36 [00:18<00:12,  1.08it/s]
 64%|██████▍   | 23/36 [00:19<00:11,  1.09it/s]
 67%|██████▋   | 24/36 [00:20<00:11,  1.08it/s]
 69%|██████▉   | 25/36 [00:21<00:10,  1.08it/s]
 72%|███████▏  | 26/36 [00:22<00:09,  1.08it/s]
 75%|███████▌  | 27/36 [00:23<00:08,  1.08it/s]
 78%|███████▊  | 28/36 [00:24<00:07,  1.08it/s]
 81%|████████  | 29/36 [00:25<00:06,  1.08it/s]
 83%|████████▎ | 30/36 [00:26<00:05,  1.08it/s]
 86%|████████▌ | 31/36 [00:26<00:04,  1.08it/s]
 89%|████████▉ | 32/36 [00:27<00:03,  1.08it/s]
 92%|█████████▏| 33/36 [00:28<00:02,  1.08it/s]
 94%|█████████▍| 34/36 [00:29<00:01,  1.08it/s]
 97%|█████████▋| 35/36 [00:30<00:00,  1.08it/s]
100%|██████████| 36/36 [00:31<00:00,  1.08it/s]
100%|██████████| 36/36 [00:31<00:00,  1.14it/s]

This example was created by a different version, suno-ai/bark:f23937d7.

Run time and cost

This model costs approximately $0.033 to run on Replicate, or 30 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 148 seconds. The predict time for this model varies significantly based on the inputs.

Readme

🐶 Bark

Original repo: https://github.com/suno-ai/bark

Bark is a transformer-based text-to-audio model created by Suno. Bark can generate highly realistic, multilingual speech as well as other audio - including music, background noise and simple sound effects. The model can also produce nonverbal communications like laughing, sighing and crying. To support the research community, we are providing access to pretrained model checkpoints ready for inference.

🙏 Appreciation

nanoGPT for a dead-simple and blazing fast implementation of GPT-style models
EnCodec for a state-of-the-art implementation of a fantastic audio codec
AudioLM for very related training and inference code
Vall-E, AudioLM and many other ground-breaking papers that enabled the development of Bark

© License

Bark is licensed under the MIT License.

Please contact us at bark@suno.ai to request access to a larger version of the model.