Amarsingh Thakur's picture

1 2

Amarsingh Thakur

spawn08

·

spawn08

AI & ML interests

None yet

Recent Activity

reacted to hexgrad's post with 👍 17 days ago

I wrote an article about G2P: https://hf.co/blog/hexgrad/g2p G2P is an underrated piece of small TTS models, like offensive linemen who do a bunch of work and get no credit. Instead of relying on explicit G2P, larger speech models implicitly learn this task by eating many thousands of hours of audio data. They often use a 500M+ parameter LLM at the front to predict latent audio tokens over a learned codebook, then decode these tokens into audio. Kokoro instead relies on G2P preprocessing, is 82M parameters, and thus needs less audio to learn. Because of this, we can cherrypick high fidelity audio for training data, and deliver solid speech for those voices. In turn, this excellent audio quality & lack of background noise helps explain why Kokoro is very competitive in single-voice TTS Arenas.

new activity 17 days ago

hexgrad/Kokoro-82M:Nice ~90x real-time generation on 3090TI. Quickstart provided.

liked a dataset 25 days ago

speechbrain/LargeScaleASR

View all activity

Organizations

Collections 1

models 1

spawn08/llama-3b-jpw-faq

Updated Dec 8, 2023 • 3

datasets

None public yet