I don't want to just spew AI-hate, but is an LLM actually necessary for this? I haven't worked with, but came across loads of non-AI TTS tools. Are these now exceptionally better to justify the overhead? Genuinely asking.
Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro
81–90 of 107 posts
Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro
#82Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro
#83Off-topic, re: “kokoro” Article refers to: https://huggingface.co/hexgrad/Kokoro-82M Caught my eye for the related name to my book of Kakuro puzzles for sale at https://www.kakurokokoro.com Kokoro comes from the Japanese word meaning something like heart or spirit, and not the literal ones.
Also from this: https://terminator.fandom.com/wiki/Kokoro
Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro
#84Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro
#85I love Kokoro. I use it to read ebooks that don't have audiobooks and it works pretty good for that. I have a Python script that reads epub/html, runs the model and writes mp3s.
Kokoro supports a limited number of languages. I've had to resort to other models to support Dutch, and they just aren't as good. And usually a lot slower and bigger (too big for my 8GB VRAM). I've had to resort to the Windows built-in TTS. It doesn't sound natural at all, but at least it can pronounce most words and is very fast.
Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro
#86TTS has come incredible long way, there are so many options. There is Kokoro of course, then there is Pocket TTS which is also a tiny 100M model that allows voice cloning. There is also Chatterbox Turbo, which is bit bigger but also allows for more emotional control of the voice. And then finally there is the Fish Audio S2, which is even bigger but allows even larger and essentially unbounded finegrained control of tone and emotion. And all of these can easily run on your Macbook.
Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro
#87I've used Piper for local TTS before, and Kokoro looks like an interesting
Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro
#88I recently created a video and didn't want to use my voice on it.
Rather than TTS I did STS (speech to speech) through Eleven Labs (you upload the video it alters the audio track to voice of choosing with some parameters you can control).
I showed it to a bunch of people and they could not believe it was non-human. The bit that got most was all the pausing, broken sentences ...essentially elements you don't put into text if your doing T2S.
An LLM over the original text could perhaps inject these elements though if prompted correctly.
Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro
#89TTS has come incredible long way, there are so many options. There is Kokoro of course, then there is Pocket TTS which is also a tiny 100M model that allows voice cloning. There is also Chatterbox Turbo, which is bit bigger but also allows for more emotional control of the voice. And then finally there is the Fish Audio S2, which is even bigger but allows even larger and essentially unbounded finegrained control of t…
Dunno why people still write about Kokoro, there are better alternatives. Pocket TTS has more natural intonation and voice cloning. Supertonic 3 is another great model in the same weight class, which is the only one that can autodetect language and make a mix of different languages sound good.
Re: Local, CPU-Friendly, High-Quality TTS (Text-to-Speech) with Kokoro
#90Supertonic 3 is pretty amazing too. I’ve got time to first sound down to well under a second by streaming chunks and stitching together from CPU with goonx (non c-go onnx) - happy to upload model if any go devs interested.