KwaiVGI / LivePortrait

Bring portraits to life!
https://liveportrait.github.io
Other
12.09k stars 1.27k forks source link

The animal model is really cool, I've incorporated it into my TensorRT Windows integration package #301

Open warmshao opened 1 month ago

warmshao commented 1 month ago

https://github.com/user-attachments/assets/385b86e8-1688-4838-9849-0ff965b1f343

You can directly get it from my project at: https://github.com/warmshao/FasterLivePortrait, and it also supports MediaPipe.

https://github.com/user-attachments/assets/dc37e2dd-551a-43b0-8929-fc5d5fe16ec5

FurkanGozukara commented 1 month ago

Amazing

I hope TensorRT can be merged into main repo

warmshao commented 1 month ago

@warmshao bro , do you have guide to complete install , I've tried many methods, but none of them worked.

i use windows 11

thankyou

You can use the Windows integration package I provided. It is portable, so you can simply unzip it and use it without installation.

warmshao commented 1 month ago

@warmshao bro , do you have guide to complete install , I've tried many methods, but none of them worked. i use windows 11 thankyou

You can use the Windows integration package I provided. It is portable, so you can simply unzip it and use it without installation.

Bro , you mean i unzip and just run

For web usage: python app.py --mode trt --mp or python app.py --mode onnx --mp

No need install any ?

Yes, you don't need to install Python packages like TensorRT. However, you need to double-click a BAT file to convert ONNX to TRT. I've written detailed instructions in the README, please refer to that.

tivikoff commented 1 month ago

Hey, nice. @warmshao Is it possible in your implementation to make v2v with animal model? Because it looks like in the LivePortrait it is currently possible only for human models

warmshao commented 1 month ago

Hey, nice. @warmshao Is it possible in your implementation to make v2v with animal model? Because it looks like in the LivePortrait it is currently possible only for human models

yes, it supports

https://github.com/user-attachments/assets/fef41fd0-8321-4357-869c-373ae6395d6a