Awesome!!!! 6GB vram worked with Large-Quant model

[agusliem]

Hi there, thanks alot many many alot for your hardworking for this repo and thanks for the man who provide Large quantize model , and i can confirmed that Large-Quant-7B work with low 6GB vram and no issue except the generating speed.

And for the result? AWESOME!!
For man voice i use Jackie chan sample voice no issue at all and it can replicate very well and for the woman i used the very noise/bad sample just for my experiment to see how far the vibe voice handles it, and it's out of my expectation, it's very very good even the sample are bads and not clear as Jackie Chan.

And for last thing i want to share is it can handle other language (in my case i use Indonesia) and the results very acceptable especially with Indonesia pronunciation.

and here the sample https://drive.google.com/file/d/1RNjens2ae-WT3Djr5KFpYdqk9dvzTwBu/view?usp=sharing

My device spec
Proc i5 Gen 13
RAM 44GB
GPU RTX 3050 6GB