Does pytorch support tools to get a quant performance with a fp32 model

quantization
2

This thread might be useful for you Speed benchmarking on android?

Please reach out to the mobile team if this script doesn’t work as expected.