search

harrisonvanderbyl / rwkv-cpp-accelerated

A torchless, c++ rwkv implementation using 8bit quantization, written in cuda/hip/vulkan for maximum compatibility and minimum dependencies
MIT License
306 stars 19 forks source link

Added support for both GUI and CLI to exporter #15

Closed Murugurugan closed 1 year ago

Murugurugan commented 1 year ago

I added support for both GUI and CLI to convert_model.py so that people can use it via CLI only without the need of installing tkinter. I also added progress bar for the CLI and some security lines so that it checks if we are truly loading the weights or something else. Also fixed the relative path issue and made it a bit faster.