Closed IsidoraR closed 2 years ago
Why do you say: why isn't TIDL_delegate used to calibrate the models for 8-bit fixed-point inference By default 8-bit tidl_offload is used.
If you wish to change it, these flags can be changed: https://github.com/TexasInstruments/edgeai-benchmark/blob/master/settings_base.yaml#L79 https://github.com/TexasInstruments/edgeai-benchmark/blob/master/settings_base.yaml#L12
Hi, In the tutorial_classification and tutorial_detection notebooks, why isn't TIDL_delegate used to calibrate the models for 8-bit fixed-point inference?