Runs a bunch of models using pytorch's post training quantization and see how fast it runs.