TensorRT-LLM

TensorRT-LLM copied to clipboard

I followed the steps provided in the tutorial to use awq-4bit quantization, but found that the accuracy was seriously lost after quantization. https://github.com/NVIDIA/TensorRT-LLM/blob/main/examples/llama/README.md#awq My model is the llama after SFT. Therefore, I want to know whether I need to use my own training set for calibration. When using awq-4bit result in dropped points, are there any parameters that can be adjusted?

+1 I have met the same problem with the accuracy drop after quantization, how can i debug with it or how to choose the calibration dataset? @Tracin @byshiue

+1 I also wonder if there is any reference for choosing an appropriate calibration dataset.

@Hukongtao Do you still have the problem? If not, we will close it soon.