[flagged]
Show HN: Three new Kitten TTS models – smallest less than 25MB
171–180 of 201 posts
Re: Show HN: Three new Kitten TTS models – smallest less than 25MB
#172Earlier quoted context omitted.
why you don't use some kind of environment, Conda or something like that?
Because I need a new version of python very rarely (years go by). I don't remember all the arcane incantations to set everything up. I did eventually do that though, and I'm pretty sure I had to mess about with installing and uninstalling torch. I dread using anything made in python because of this. It's always annoying and never just works (if the version of python is incompatible, otherwise it's fine) .
I don't know, I'm pretty happy with Conda. I just create a new environment and install on it. It normally works.
Even if you have to install using pip it just affect the active environment.
Maybe I'm only trying simple things.
Re: Show HN: Three new Kitten TTS models – smallest less than 25MB
#173How noticeable is the difference in quality between the 4M model and the 80M model?
Re: Show HN: Three new Kitten TTS models – smallest less than 25MB
#174They sound like cartoon voices... but I really like them I could listen to a book with those.
Yeah, I was wondering if it's all helium voices. Should maybe try and see or find more demos.
Re: Show HN: Three new Kitten TTS models – smallest less than 25MB
#175[dead]
Re: Show HN: Three new Kitten TTS models – smallest less than 25MB
#176the dependency chain issue is a real barrier for edge deployment. i've been running tts models on a raspberry pi for a home automation project and anything that pulls torch + cuda makes the whole thing a non-starter. 25MB is genuinely exciting for that use case.
curious about the latency characteristics though. 1.5x realtime on a 9700 is fine for batch processing but for interactive use you need first-chunk latency under 200ms or the conversation feels broken. does anyone know if it supports streaming output or is it full-utterance only?
the phoneme-based approach should help with pronunciation consistency too. the models i've tried that work on raw text tend to mispronounce technical terms unpredictably — same word pronounced differently across runs.
Re: Show HN: Three new Kitten TTS models – smallest less than 25MB
#177Earlier quoted context omitted.
not v long. until then you start running tts on phones, wearables and r pis. at the model level, we'll have a model for this kind of mcu's later this year.
You can (just about?) already run on a pi zero, right? That's not literally a chip, but in practical utility it can't be very different
A CPU will probably consume much more power.
Re: Show HN: Three new Kitten TTS models – smallest less than 25MB
#178[dead]
Re: Show HN: Three new Kitten TTS models – smallest less than 25MB
#179the dependency chain issue is a real barrier for edge deployment. i've been running tts models on a raspberry pi for a home automation project and anything that pulls torch + cuda makes the whole thing a non-starter. 25MB is genuinely exciting for that use case. curious about the latency characteristics though. 1.5x realtime on a 9700 is fine for batch processing but for interactive use you need first-chunk latency u…
Could you share what you're currently using?
Re: Show HN: Three new Kitten TTS models – smallest less than 25MB
#180Only American voices? For some reason I'm only interested in Irish, British or Welsh accents. American is a no
minor nit to pick: Welsh accents are British accents as Wales is in Britain. In fact by some definitions it's the most British part. People from outside the UK often use British as synonymous with English, and in the context of accents, often a South East English accent or some sort of Received Pronunciation (RP) accent. Technically a "British" accent could be from anywhere in England, Scotland, or Wales, and therefo…
I am actually English but I'm so used to speaking with international people I instinctively say British instead of English - because that's what people expect.
So being factually correct doesn't really matter. Nobody cares and nobody wants to learn so I adapt for them.
In the same way I almost exclusively write with American spelling now. Life is just easier when you stop fighting.