Hacker News Reader

Comment by ftreml | Hacker News Reader

good point, will add this information. in short, german and english, because those are the languages i am comfortable with. i hope to find native speakers contributing more languages. deepspeech was evaluated, but right now kaldi provides better performance, thats why we stick to kaldi by default.

View on HN

Mozilla DeepSpeech has seen significant improvements in the quality of results it provides (and RAM/CPU usage) since early December, the accuracy of transcriptions has substantially improved in our testing.

Can you provide some quantifiable results, like word error rate or something? Do you think it is better than what Kaldi offers currently? (In terms of accuracy, not in terms of computational performance.)

Every release mentions the word error rate, its interesting to watch the WER improve over time: https://github.com/mozilla/DeepSpeech/releases

> deepspeech was evaluated, but right now kaldi provides better performance

Was this before the streaming API was added to DeepSpeech? I recently did some testing with it, and it provides text within ~100ms of last audio block on my PC.

edit: that is, the most significant latency I had was from having to wait a bit to detect end of speech.

with "performance" i meant the error rate, not the speed. speed was not a criteria for us, so we didnt evaluate it. i read that deepspeech is way quicker in training phase as it is smarter in using gpu computing power.

Ahh gotcha. Yes I got rather poor "hit rate" until I made my own language model.

Fortunately for me I only needed it for command recognition, so the process was quite quick, and results were very good. No need to retrain the net.

we used a context-free data set, augmented with additional domain-specific sound samples and it worked out fine, although the additiinal samples made nearly no difference

Reply on news.ycombinator.com