正在加载视频...

视频加载失败

Open Source NotebookLM Alternative on Your PC Wow Dia is REALLY good. 1. Voice clone 2. Non-verbal generation: (laughs), (coughs), .. 3. Dialogues: Just upload 1 clip with 2 voices to generate a NotebookLM style dialogue 1-Click Gradio Launcher. Mac, Windows, Linux.

118,830 次观看 • 1 年前 •via X (Twitter)

10 条评论

cocktail peanut 的头像
cocktail peanut1 年前

Available on Again, this works on ALL platforms: - Mac - Windows - Linux

cocktail peanut 的头像
cocktail peanut1 年前

Original github repo

cocktail peanut 的头像
cocktail peanut1 年前

Note that there's an issue with voice cloning right now. Just to be clear, everything working fine if you just use the TTS alone, but when you try to use the voice cloning feature AND use a long reference audio clip, it will generate gibberish sound or cut short because of the way it works. This is NOT the model's fault but just the way the gradio app logic is written currently. There's a related Github issue here It's unfortunate that many people will try this model through the local installation or the huggingface space, and in BOTH cases, the voice cloning feature is broken right now if you try 1. slightly long reference audio 2. try to generate slightly longer audio hopefully this gets addressed soon, the model deserves better, cc: @_doyeob_

cocktail peanut 的头像
cocktail peanut1 年前

if you want this to generate longer than 25 sec with voice cloning, please go ask the author on this thread, the more people ask the more likely that they will implement this in the app

ILikeToasters 的头像
ILikeToasters1 年前

Perfect timing. Was just about to install without Pinokio.

cocktail peanut 的头像
cocktail peanut1 年前

you should try learning gepeto, because if you can install without pinokio, you can easily generate a pinokio script that does the same thing. that way, once you get it working you can share it back with the community.

David Harvey 的头像
David Harvey1 年前

I don't know man. They are cherry picking what to show ☠️

Don Jose Valle 的头像
Don Jose Valle1 年前

its available in spanish?

nico oliver 的头像
nico oliver1 年前

Great work, but needs a bit of fine tuning. Wont produce clips over 30 secs when I tested. This may have been fixed by now. The speed of the generated voice is also very fast. If I try to adjust it, it just makes the voice deeper but the rate of speech is still to fast.

Mark Gadala-Maria 的头像
Mark Gadala-Maria1 年前

Dude you're a legend thank you for adding this so quickly!!

相关视频