search

YuanGongND / ast

Code for the Interspeech 2021 paper "AST: Audio Spectrogram Transformer".
BSD 3-Clause "New" or "Revised" License
1.06k stars 203 forks source link

For own data #113

Open ShadowVicky opened 9 months ago

ShadowVicky commented 9 months ago

How can i use this model for my own dataset.

YuanGongND commented 9 months ago

You can check https://github.com/YuanGongND/ast#use-pretrained-model-for-downstream-tasks.

Also, you can modify the ESC-50 recipe to your own dataset. https://github.com/YuanGongND/ast/blob/master/egs/esc50/run_esc.sh

The ESC-50 recipe is one click after you set up the environment, it does everything you would need for a new dataset. I always recommend first run that and if success, then moving to your own data.

-Yuan