Music Generation

Music Generation using the meta/musicgen AI Model.

POST http://localhost:9000/apiv1/replicate/musicgeneration

Try it

Music Generation

cURL

curl --request POST \
  --url http://localhost:9000/api/v1/replicate/musicgeneration \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: multipart/form-data' \
  --form 'prompt=<string>' \
  --form 'model_version=<string>' \
  --form duration=123 \
  --form continuation=true \
  --form continuation_start=123 \
  --form continuation_end=123 \
  --form multi_band_diffusion=true \
  --form 'normalization_strategy=<string>' \
  --form top_k=123 \
  --form top_p=123 \
  --form temperature=123 \
  --form classifier_free_guidance=123 \
  --form 'output_format=<string>' \
  --form input_audio='@example-file'

Query

model · string · default: "meta/musicgen" · required

The model to be used;

Body

prompt · string · required

A description of the music you want to generate.

model_version · string · default: "stereo-melody-large"

Model to use for generation

input_audio · file

An audio file that will influence the generated music. If continuation is True, the generated music will be a continuation of the audio file. Otherwise, the generated music will mimic the audio file’s melody.

duration · int · default: "8"

Duration of the generated audio in seconds.

continuation · bool · default: "False"

If True, generated music will continue from input_audio. Otherwise, generated music will mimic input_audio’s melody.

continuation_start · int · default: "0"

Start time of the audio file to use for continuation.

continuation_end · int · default: "0"

End time of the audio file to use for continuation. If -1 or None, will default to the end of the audio clip.

multi_band_diffusion · boolean · default: "False"

If True, the EnCodec tokens will be decoded with MultiBand Diffusion. Only works with non-stereo models.

normalization_strategy · string · default: "peak"

Strategy for normalizing audio.

top_k · int · default: "250"

Reduces sampling to the k most likely tokens.

top_p · float · default: "0.0"

Reduces sampling to tokens with cumulative probability of p. When set to 0 (default), top_k sampling is used.

temperature · float · default: "1.0"

Controls the ‘conservativeness’ of the sampling process. Higher temperature means more diversity.

classifier_free_guidance · int · default: "3"

Increases the influence of inputs on the output. Higher values produce lower-varience outputs that adhere more closely to inputs.

output_format · string · default: "mp3"

Output format for generated audio.