MusicGen

This is your private demo for MusicGen, a simple and controllable model for music generation presented at: "Simple and Controllable Music Generation"
Model
Decoder
1 120
Examples
Input Text Condition on a melody (optional) Model Decoder

More details

The model will generate a short music extract based on the description you provided. The model can generate up to 30 seconds of audio in one pass.

The model was trained with description from a stock music catalog, descriptions that will work best should include some level of details on the instruments present, along with some intended use case (e.g. adding "perfect for a commercial" can somehow help).

Using one of the melody model (e.g. musicgen-melody-*), you can optionally provide a reference audio from which a broad melody will be extracted. The model will then try to follow both the description and melody provided. For best results, the melody should be 30 seconds long (I know, the samples we provide are not...)

It is now possible to extend the generation by feeding back the end of the previous chunk of audio. This can take a long time, and the model might lose consistency. The model might also decide at arbitrary positions that the song ends.

WARNING: Choosing long durations will take a long time to generate (2min might take ~10min). An overlap of 12 seconds is kept with the previously generated chunk, and 18 "new" seconds are generated each time.

We present 10 model variations:

  1. facebook/musicgen-melody -- a music generation model capable of generating music condition on text and melody inputs. Note, you can also use text only.
  2. facebook/musicgen-small -- a 300M transformer decoder conditioned on text only.
  3. facebook/musicgen-medium -- a 1.5B transformer decoder conditioned on text only.
  4. facebook/musicgen-large -- a 3.3B transformer decoder conditioned on text only.
  5. facebook/musicgen-melody-large -- a 3.3B transformer decoder conditioned on and melody.
  6. facebook/musicgen-stereo-*: same as the previous models but fine tuned to output stereo audio.

We also present two way of decoding the audio tokens

  1. Use the default GAN based compression model. It can suffer from artifacts especially for crashes, snares etc.
  2. Use MultiBand Diffusion. Should improve the audio quality, at an extra computational cost. When this is selected, we provide both the GAN based decoded audio, and the one obtained with MBD.

See github.com/facebookresearch/audiocraft for more details.