Paper of this repository. about forwardtacotron HOT 3 OPEN

We will do a comparison soon, hopefully. The latencies in the FastSpeech paper are measured with a batch size of 1, which is an unrealistic setting for production systems, so we probably will recompute the latencies for FastSpeech as well.

from forwardtacotron.

Hi, we unfortunately do not have the (time) resources currently to write papers, the closest would be FastSpeech, which the current implementation is based on: https://arxiv.org/abs/1905.09263

We did some non-scientific internal comparisons of Tacotron and ForwardTacotron with WaveRNN where ForwardTacotron performed slightly better, but that's anecdotal. We would of course be happy about a real comparison if anyone would like to contribute.

from forwardtacotron.

Hi, we unfortunately do not have the (time) resources currently to write papers, the closest would be FastSpeech, which the current implementation is based on: https://arxiv.org/abs/1905.09263

We did some non-scientific internal comparisons of Tacotron and ForwardTacotron with WaveRNN where ForwardTacotron performed slightly better, but that's anecdotal. We would of course be happy about a real comparison if anyone would like to contribute.

Do you compared inference latency of forward tacotron with orther parallel tacotron like fastspeech?

from forwardtacotron.