@VikParuchuri When identifying the book ”choose the right word“, the index word in the upper left corner will be mistakenly identified as the title.
English
chengbo
39 posts
Surya OCR doesn't have a tokenizer (UTF-16 bytes). We've been experimenting with predicting multiple bytes simultaneously.
Seeing linear speedups with minimal accuracy loss. Our tokenizer-free version is faster than our tokenized version of the same model.
English
@cocktailpeanut @realmrfakename github.com/RVC-Boss/GPT-S…
Did you consider adding this to 1 Click Launcher?
English
MeloTTS 1 Click Launcher
Ultra fast TTS! Took this Huggingface Gradio App from @realmrfakename & added:
1. Multi Language: English, Spanish, French, Chinese, Korean, Japanese! (Examples below)
2. 1 Click Launcher: Mac, Win, Linux
A whole paragraph in just 3~6s on M1 Max 64G!
Gradio@Gradio
𝐌𝐞𝐥𝐨𝐓𝐓𝐒 demo on Spaces is🚀🚀 A permissively licensed (MIT) high-quality multi-speaker multi-lingual TTS library. 🤯Translated @karpathy's Intro to LLMs (~12k words) transcript to audio in <60 sec, producing a 250mb+ file as output. [🔊Sound: On]
English
chengbo retweetledi
xdite.gitbooks.io/learn-somethin… 我的學習文集。把我這五年來的各式各樣關於「學習」的文集，整理好放在 gitbook了 .....
中文
Xdite 學習歷程經驗 · Issue #23 · softnshare/careergrowth github.com/softnshare/car…
中文
Check out Dash, the Documentation Browser for 150+ APIs kapeli.com/dash
English
