Firefox Voice
161–166 of 166 posts
Re: Firefox Voice
#162Earlier quoted context omitted.
There's tens of millions of hours of video content out there that has been subtitled pretty well, and I'd wager a lot of it is under usable licenses for Mozilla. Has that been considered?
If you know such sources, file an issue, and better yet, download the video content yourself and publish a dataset. But note that raw video content is not training data. It has to be segmented to be in short enough parts for training (few seconds), the subtitles have to be aligned to match what's said precisely, and one needs to balance the data, e.g. when 90% of speakers are men and 10% are women, you have a problem…
Gotcha, all excellent points.
Re: Firefox Voice
#163I see a lot of skeptical voices here, (somewhat warranted, given it's a voice assistant technology), but the fact remains that if we want open, on-device voice recognition, we'll have to do the work and donate sample data. This extension is trying to provide some useful functionality in the hopes that Mozilla gets more data for https://commonvoice.mozilla.org I'd at least consider recording your voice, especially if…
Nokia n800 was crushed by internal fighting.
Re: Firefox Voice
#164I see a lot of skeptical voices here, (somewhat warranted, given it's a voice assistant technology), but the fact remains that if we want open, on-device voice recognition, we'll have to do the work and donate sample data. This extension is trying to provide some useful functionality in the hopes that Mozilla gets more data for https://commonvoice.mozilla.org I'd at least consider recording your voice, especially if…
You have wrong premise here. First of all, current Mozilla data is almost useless for training because it is carefully read speech. You do not need much of it, even accents. If you add 1000 hours of CV data to 1000 hours of random data the improvement in accuracy will be minimal. Same for the speech collected with Firefox Voice, it will be mostly a set of short commands, most likely it will not be very useful for generic transcription of random people speech.
Second, you can build models much better than Mozilla ones simply from public data, there is no need to collect user voices. We at Vosk https://alphacephei.com/vosk/ support 10+ languages for example without any user data. Everyone creates very good models from augmented text-to-speech data these days (Microsoft demonstrated in the last paper you can get almost as good as domain-specific data https://arxiv.org/abs/2007.15188).
Given that it surprises me that Mozilla continues to insist they need the voices of their userbase.
Re: Firefox Voice
#165Earlier quoted context omitted.
I want hands free for CAD. Imagine being able to vocalise and build a model. I did have a HN user who said they be happy to collaborate with me to build it but I dropped the ball and have since killed that email address.
If you manage to get into the GPT-3 beta, I'd love to work on that with you :D. For simple models, English -> OpenSCAD sounds like it's doable given the things I've seen on Twitter and for normal modeling, GPT-3 would probably make an excellent intent recognizer for voice commands.
How would I go about looking in to this?
I'm a metal fabricator by trade but also technically minded, AutoHotKey scripts for a few things and can very basic Python and C# if I need to.
I tend to learn in a very solutions oriented.
And would be keen to collaborate / learn / skill share.
Re: Firefox Voice
#166Earlier quoted context omitted.
If you manage to get into the GPT-3 beta, I'd love to work on that with you :D. For simple models, English -> OpenSCAD sounds like it's doable given the things I've seen on Twitter and for normal modeling, GPT-3 would probably make an excellent intent recognizer for voice commands.
How would I go about looking in to this? I'm a metal fabricator by trade but also technically minded, AutoHotKey scripts for a few things and can very basic Python and C# if I need to. I tend to learn in a very solutions oriented . And would be keen to collaborate / learn / skill share.
The GPT-3 beta is something every programmer and their dog wants to get into these days and most of us can't. It's a really impressive new language processing neural network that people have managed to train to (among many other things) generate code from an English description of the program. If it can do that, it might be able to generate some reasonably complex Constructive Solid Geometry models and even something like MEL commands in Maya.
Greg, OpenAI's CTO, occasionally manually lets people in if they convince him of their use case (or, as one guy did, plant a bunch of trees in his name) in an email (gdb@openai.com). It might be worth shooting him a message explaining the idea. From what I've heard, once you have a key, it's mostly a matter of feeding the model examples until it does what you want.