When converting quantized model to tensorrt, it says “TRT only supports symmetric quantization - zeroPt must be all zeros”. Is there a way to set PyTorch’s built-in quantization tool’s quantization parameters(symmetric quantization) so that it can be converted to a TensorRT model?
tensorrt-issue
It seems it is possible: TensorRT/test_quant_trt.py at main · pytorch/TensorRT · GitHub
in fx mode right now