Toward a DeepSpeech Modelers Meeting?

One of us from Te Hiku plan on giving a talk. I’m stepping back to listen but if there’s enough interest from others on an overview of what we’re doing from an indigenous perspective I’m happy to take no more than 5 minutes to do that… unless it’s pretty chill and we all don’t mind hanging out for a couple of hours :sweat_smile:

1 Like

How will we find the zoom/matrix details? I’ll check this thread in the morning. Talk soon!

Yeah, I’ll share that. I’m unsure matrix will work, I’m unable to share my screen there :confused: so I guess it will be Zoom.

DeepSpeech Modelers Meeting
29 juil. 2020 10:00 PM Paris

Participer à la réunion Zoom
https://mozilla.zoom.us/j/98646762351

ID de réunion : 986 4676 2351
Une seule touche sur l’appareil mobile
+13462487799,,98646762351# États-Unis d’Amérique (Houston)
+16465588656,,98646762351# États-Unis d’Amérique (New York)

Composez un numéro en fonction de votre emplacement
        +1 346 248 7799 États-Unis d’Amérique (Houston)
        +1 646 558 8656 États-Unis d’Amérique (New York)
        +1 669 900 9128 États-Unis d’Amérique (San Jose)
        +1 253 215 8782 États-Unis d’Amérique (Tacoma)
        +1 301 715 8592 États-Unis d’Amérique (Germantown)
        +1 312 626 6799 États-Unis d’Amérique (Chicago)
        877 853 5257 États-Unis d’Amérique Numéro gratuit
        +61 8 7150 1149 Australia
        +61 2 8015 6011 Australia
        +61 3 7018 2005 Australia
        +61 731 853 730 Australia
        +61 861 193 900 Australia
        1800 893 423 Australia Numéro gratuit
        +1 647 374 4685 Canada
        +1 647 558 0588 Canada
        +1 778 907 2071 Canada
        +1 204 272 7920 Canada
        +1 438 809 7799 Canada
        +1 587 328 1099 Canada
        855 703 8985 Canada Numéro gratuit
        +33 1 8699 5831 France
        +33 1 7037 2246 France
        +33 1 7037 9729 France
        +33 1 7095 0103 France
        +33 1 7095 0350 France
        0 805 082 588 France Numéro gratuit
        +49 69 3807 9883 Allemagne
        +49 695 050 2596 Allemagne
        +49 69 7104 9922 Allemagne
        +49 30 5679 5800 Allemagne
        0 800 724 3138 Allemagne Numéro gratuit
        +44 203 901 7895 Royaume-Uni
        +44 208 080 6591 Royaume-Uni
        +44 208 080 6592 Royaume-Uni
        +44 330 088 5830 Royaume-Uni
        +44 131 460 1196 Royaume-Uni
        +44 203 481 5237 Royaume-Uni
        +44 203 481 5240 Royaume-Uni
        0 800 031 5717 Royaume-Uni Numéro gratuit
ID de réunion : 986 4676 2351
Trouvez votre numéro local : https://mozilla.zoom.us/u/aeqpmOaYQi

Participer à l’aide d’un protocole SIP
98646762351@zoomcrc.com

Participer à l’aide d’un protocole H.323
162.255.37.11 (États-Unis (Ouest))
162.255.36.11 (États-Unis (Est))
115.114.131.7 (Mumbai - Inde)
115.114.115.7 (Hyderabad - Inde)
213.19.144.110 (EMEA)
103.122.166.55 (Australie)
64.211.144.160 (Brésil)
69.174.57.160 (Canada)
207.226.132.110 (Japon)
ID de réunion : 986 4676 2351

Hmm no audio in the browser version of Zoom (tried both FF and chromium), so much for webapps in 2020 :frowning_face:.

Sorry to hear that :confused: Have you tried the native version?

Nope, never had to use Zoom and not looking forward to installing binaries from a firm with a questionable track record. Mostly used jitsi instead.

Right, I can understand that. Matrix is supposed to use Jitsi in the background, unfortunately it was not working appropriately when I tried, but thanks for the feedback, I hope we can use something else than Zoom next time.

Hmm screen sharing indeed doesn’t seem entirely stable with jitsi, just tested and it resulted in one freeze of FF, after starting with a fresh FF it seemed to be working. sigh

Thx @lissyx for organizing and hosting this. Enjoyed hearing about other’s work. Here’s my feedback:

  • Frequency - 1 every two months sounds good, 1 ever 3 month okay.
  • Recorded - Yes, and we could also live stream it for those who just want to watch but don’t necessarily want to be on the Zoom call.
  • Topics:
    • I enjoyed hearing about the upcoming features e.g. 0.8 - that’s useful to know
    • I enjoyed hearing about other language efforts especially indigenous languages and languages that are low resourced or in a state of revitalization
    • Also having more of a deep dive on technical things I think would be helpful. Maybe we could alternate between technical and then more community focus e.g. what projects are people doing and how are you using DeepSpeech
  • Length / Times: I was happy with the length… 1.5 - 2 hours seems like a good duration. It was 8am New Zealand time for me which was also fine… but I’d be willing to trade time so that more people have the opportunity to join in. So I’d be okay with 6am our time or 12/1am knowing that we’d have more participation. Maybe based on who’s raised their hands to present then their time choices might have more weighting or if a particular group hasn’t had the opportunity to participate then a time that works for them gets more weighting.

I was happy with your hosting @lissyx. We’ll always have connection issues so I think just reminding people in the beginning to do things like mute or turn off their video especially when we notice that a speaker might have connection issues.

2 Likes

Hello;
Thank you @lissyx for hosting this meeting. I was pleased to meet everyone and hear about their work.

Thank you for sharing with us the new features of the latest version of DeepSpeech. I would have liked to have more information on the steps taken by each in their work with a more in-depth technical content.

I’m up for a meeting of around 3 hours once every 3 months. You can record it and stream it.

I agree with @keoni 's comments. And it would be great to have like a 12min.me format for presentations where you present for 12 minutes and then discuss it for 12 minutes. That way you get some participation and can have multiple topics with equal length.

I meant to comment this on the Problems Training on RTX3080 post but the page doesnt allow more than 3 replies per topic for new users but here is the comment “That is exactly the conclusion I was coming to after weeks of trying to train it on Windows with Gpu. I felt defeated. Interestingly enough, before trying to train with my RTX 3090 on my Windows PC, I was able to train it when using a CPU and not a GPU. So my best option is to just use my CPU on Windows? Or else, what is your recommendation?” also maybe we can continue chatting on another service that doesnt restrict replies? do you have Facebook, IG, or whatsapp?