Model card contadicts information on Github.
#2
by
ayymen
- opened
GitHub repo says Phoneme error rate, not WER:
https://github.com/huggingface/transformers/tree/main/examples/pytorch/speech-recognition#common-voice-ctc