Skip to content
This repository has been archived by the owner on Sep 18, 2021. It is now read-only.

Server performance #89

Open
jwhite2000 opened this issue May 7, 2019 · 3 comments
Open

Server performance #89

jwhite2000 opened this issue May 7, 2019 · 3 comments

Comments

@jwhite2000
Copy link

Is there any way to speed up the server's response time. I am calling the api here https://speaktome-2.services.mozilla.com/ using the speaktome javascript modules developed by Mozilla. It is taking between 3 and 6 seconds to get a response.

@milescrabill
Copy link
Contributor

Yeah, that's pretty abysmal. There are a few issues with the current server side architecture that we're looking at addressing. We're also moving towards a new speech solution based on Deepspeech that will run on GPUs instead of CPUs. The hope is that that will speed things up significantly.

@flozi00
Copy link

flozi00 commented Apr 4, 2020

how can we speed up the transition to deepspeech?
Any possibilitys to support ?

@flozi00
Copy link

flozi00 commented Apr 4, 2020

Yeah, that's pretty abysmal. There are a few issues with the current server side architecture that we're looking at addressing. We're also moving towards a new speech solution based on Deepspeech that will run on GPUs instead of CPUs. The hope is that that will speed things up significantly.

What about clustering like solution ?

Client asks --> server returns IP --> client use ip instead of https://speaktome-2.services.mozilla.com/ as static server

Then you have a load balancing option, with only one more Handshake

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants