from this forum. I know that it is impossible to use Tensorflow with OpenMV. but how does Edge Impulse convert model to tflite that support OpenMV?. is that related with Tensorflow Post Training Quantization
Hi @denuradhan ,
Edge impulse uses multiple optimization techniques to make ML run on embedded devices including the OpenMV cam. Among them is the TensorFlow’s Model Optimization Toolkit to quantize models, reducing their weights’ precision from float32 to int8 with minimal impact on accuracy.
To learn more on how Edge impulse utilizes Tensorflow to make ML accessible to embedded Engineers, you can have a look at this blogpost: How TensorFlow helps Edge Impulse make ML accessible to embedded engineers — The TensorFlow Blog