4 ms·
Does anyone know ho this compares with Mozilla's DeepSpeech? https://github.com/mozilla/DeepSpeech https://github.com/mozilla/DeepSpeech
by Yoric 8y ago
Does anyone know ho this compares with Mozilla's DeepSpeech? https://github.com/mozilla/DeepSpeech https://github.com/mozilla/DeepSpeech
- yorwba 8y agoFrom https://arxiv.org/abs/1812.06864 https://arxiv.org/abs/1812.06864 "On Librispeech, we report state-of-the-art performance among end-to-end models, including Deep Speech 2 trained with 12 times more acoustic data and significantly more linguistic data." Specifically they claim word error rates that are 1 to 2 percentage points lower, 3.44% on "clean" and 11.24% on "other".
- StudentStuff 8y agoIt'll be interesting to see if anyone can reproduce their results, thus far its been troublesome: https://github.com/facebookresearch/wav2letter/issues/88 https://github.com/facebookresearch/wav2letter/issues/88
- m0zg 8y agoProbably positively. I brought up DeepSpeech in a docker container the other day, it couldn't understand me at all, using Mozilla's pre-trained model. _Some_ words were right, but the output did not make sense.
- StudentStuff 8y agoDid you use one of the examples (eg the Python GUI or CLI client)? WebRTCVad (which is used in most of those) made mincemeat of my audio files, versus feeding them into DeepSpeech got me usable results. Just need VAD for files nearing or over 1 minute in length (unless you have 4+GB of RAM to spare).