On the CommonVoice benchmark, XLSR shows a relative phoneme error rate reduction | |
of 72% compared to the best known results. |
On the CommonVoice benchmark, XLSR shows a relative phoneme error rate reduction | |
of 72% compared to the best known results. |