fix inference with latest tensorflow by introducing evaltensors into generated code

cpetig / tflite_micro_compiler

generate tflite micro code which bypasses the interpreter (directly calls into kernels)

Apache License 2.0

77 stars 26 forks source link

fix inference with latest tensorflow by introducing evaltensors into generated code #56

Closed rafzi closed 4 years ago

rafzi commented 4 years ago

This adds more overhead for the additional structures. Currently it seems like the operators require both tensor structs to exist. Maybe this will change in the future?