adavirro's picture
Create README.md
84a060c
|
raw
history blame
239 Bytes

DistilHubert fine-tuned on gtzan dataset as part of HF Audio Course:

https://huggingface.co/learn/audio-course/chapter4/fine-tuning

I changed the train-test seed like a rebel so final eval is different:

Loss: 0.604328 Accuracy: 0.800000