Intent (audio) Classification Challenge

Hello everyone,

We’re launching the Intent Classification challenge in the ML for Audio study group this week.

Challenge - Build the most accurate audio classification model for the fluent.ai intent classification data.

Scoring criteria - Accuracy (the model with highest accuracy wins)

Dataset - http://140.112.21.28:9000/fluent.tar.gz

Models - Any model/ architecture available on the HF hub (be as creative as you want to be)

Deadline - 15th February 2022

Submission format - Tag @osanseviero or I in the #ml-4-audio-study-group discord channel (Discord (tester)) with your colab/ hf-hub model link.

Here’s some starter code to get you started - Google Colab

Feel free to put down your questions below, all the best!

P.S. Oh and the winner gets some sweet HF swag! xD