Gradio

upload some audio (will be randomly trimmed to max of 100s)

input audio

periodic prompt

0 13

onset mask width (multiplies with the periodic mask, 1 step ~= 10milliseconds) does not affect mask preview

0 100

beat mask width (milliseconds) does not affect mask preview

1 200

compression prompt

1 14

mask dropout

0 1

feedback steps (token telephone) -- turn it up for better timbre/rhythm transfer quality, but it's slower!

1 8

preset

mask image

seed (0 for random)

model choice

output audio 1

output audio 2

Input Audio

Randomness

Controls sampling diversity- higher values yield more novel variations.

0.1 10

Rhythmic Reference

Determines how frequently rhythmic cues are repeated in generation.

0 13

Variation Amount

How much of the input the model changes. 0 = close copy, higher = greater remix.

0 1

Timbral Reference

Controls the richness of timbre retained from the source audio. Lower = simpler, higher = richer timbre.

1 14

Model Choice

Select which fine-tuned VampNet model to use.

Time Stretch Factor

Scales the output duration relative to the input. 1 = same tempo, 0.5 = half speed.

0 8

Generated Audio

Generated Labels

Controls Data