marlin
marlin copied to clipboard
comparison with SmoothQuant
I wonder whether you can compared with SmoothQuant, it is a great quantization method that many people tend to use.
I wonder whether you can compared with SmoothQuant, it is a great quantization method that many people tend to use.
smoothquant is a method to quantize models, this repo is the implementation of cuda kernel. You cant compare these two things.