ftyers

9 ms·

121.
by ftyers 5y ago
The Norwegian government and Sámi parliament put a lot of effort into language technology for the Sámo languages. A big problem is lack of openness in platform support. E.g. Google and Apple make it very difficult for external developers to
122.
by ftyers 5y ago
Common Voice is mostly used for STT not TTS. TTS requires single speaker, clean audio. STT requires multi speaker, noisy audio.
123.
by ftyers 5y ago
One of the most noticeable additions in my opinion is Guarani, the first Indigenous language of the Americas to be added. Indigenous languages are extremely poorly supported and forgotten by all of the major platforms and companies, and it&
124.
by ftyers 5y ago
That's what happens when people have the opportunity and tools to support their own languages and not just rely on hand outs from big tech :)
125.
by ftyers 5y ago
Do you have JavaScript turned on? No paywall here.
126.
by ftyers 5y ago
Yeah, which is a good step forward, but it's definitely not usable yet in my experience. There is definitely a vibe of "why would anyone want to use anything other than Go" in the community. (I actually really like IPFS, I j
127.
by ftyers 5y ago
An additional problem is the lack of language bindings/implementations. There is no fully-fledged implementation apart from the Go one, which makes getting many developers onboard a challenge.
128.
by ftyers 5y ago
Sqlite to the rescue. Here is a list ordered by delay. https://dpaste.com/6JRADR63T Looks like Guayaquil is a hot ticket!
129.
by ftyers 5y ago
Would be cool to see a list of consulates ordered by speed for a given visa class. In the US ATM but have to leave the country to renew the H1B (ugh), but don't care where I do it. The thing has been authorised, just need to do the DS-
130.
by ftyers 5y ago
Better in what respect? Larger? That's true, but is largely a reflection of land prices, and if you're able to steal the land it makes it a lot cheaper. In Europe most land was stolen a long time ago and there isn't so much t
131.
by ftyers 5y ago
Dunbar-Ortiz has a good take on this, for her the US American identity is basically a European settler/covenant identity, much like the whites in other settler colonies: Australia and South Africa or Rhodesia (as they were). Unlike in
132.
by ftyers 5y ago
There is another article about the TTS: https://news.ycombinator.com/item?id=26790951
133.
by ftyers 5y ago
You don't need _no_ errors, you just need low errors. Aside from Common Voice, there are also a lot of resources at openslr. Also, the amount of data you need is often vastly overestimated, with advances in pretraining and transfer lea
134.
by ftyers 5y ago
Iirc it is realtime, but check out the Matrix channel: https://app.element.io/#/room/#coqui-ai_TTS:gitter.im
135.
by ftyers 5y ago
I believe Espeak is only used for the grapheme to phoneme conversion (if at all). The rest is all new.
136.
by ftyers 5y ago
I wasted a week trying to replace the scorer component with a NN-based language model. Every time I made a change the whole codebase, including Tensorflow recompiled, so the turnaround time was about an hour per change. It was awful. I mean
137.
by ftyers 5y ago
You might want something like LocalSTT if it's on mobile: https://github.com/ccoreilly/LocalSTT Otherwise this code does streaming on a websocket: https://github.com/coqui-ai/STT-examples&#x2F
138.
by ftyers 5y ago
Yeah, ROCm is a bit of a mess. I actually have an AMD GPU in a server, but the drivers in the mainline kernel don't work properly, so have never been able to use it. If I were into conspiracy theories I'd say that AMDs failure to
139.
by ftyers 5y ago
There is a lot of example code here: https://github.com/coqui-ai/STT-examples If you have any more specific requirements then we can point you in the right direction. Or just join us on Matrix: https://app.e
140.
by ftyers 5y ago
They do, the issue is with Tensorflow support iirc and with NVIDIA drivers. So for the English model they use mostly free/open-source data, but some non-free data: - train_files Fisher, LibriSpeech, Switchboard, Common Voice English, a
141.
by ftyers 5y ago
When are you going to do Chuvash ? ;)
142.
by ftyers 6y ago
They are traditionally reindeer herders. There is a page on Wikipedia about Sámi cuisine you can check out. If you are interested, you could also check out the films Kukuška (2002) and Ofelaš (1987).
143.
Ask HN: Deep Learning with Free Software
1 points by ftyers 6y ago|1 comments
144.
by ftyers 6y ago
Exactly, or should I say nemleg :)
145.
by ftyers 6y ago
James Maffie's "Aztec Philosophy" is pretty great, as is "Filosofia Náhuatl" by Miguel Leon Portilla. (If you are interested in process metaphysics) Oh and the latter seems to be available online: http://
146.
by ftyers 6y ago
Wrong. Norway.
147.
by ftyers 7y ago
Mozilla DeepSpeech trained on the Common Voice dataset for English. You can get pretrained models too. They have a nice matrix channel where you can get help, and pretty good documentation. It is also actively developed by several engineers
148.
by ftyers 7y ago
why is this even a thing. just use the NHS.