neonbjb / tts-scores

Scripts for computing the Intelligibility and CLVP scores for evaluating TTS models
Apache License 2.0
142 stars 15 forks source link

Missing Scores in README #10

Closed fakerybakery closed 11 months ago

fakerybakery commented 11 months ago

Hi, In the readme, it says:

Scores from common models A metric is only good if there are benchmarks which can be used as points of comparison. To this end, I computed all of the scores in this repo on two high-performance TTS models:

Tacotron2+waveglow from NVIDIA's repo FastSpeech2+hifigan from ming024's repo See the scores below:

However, no scores are provided. Are these available? Thank you!

fakerybakery commented 11 months ago

Whoops, duplicate of #3 - sorry!