Tool/software: Code Composer Studio
Dear Sir,
We have a Tensorflow based model which has no post-processing included in it. For post-processing, we have a separate code to generate the Bounding boxes on the test image.
For post-processing script, we need three layers of output.
for your reference:
The model has 127 hidden layers(after importing) through TIDL out which we need the Hidden Layer-109, Hidden Layer-120 and Hidden Layer-127 dump floating values for Post-processing.
Currently, we are reading these hidden layers dumps as floating data type from "float.bin" files generated after each layer dumps.
We need your support to understand few arguments listed below:
1) In which part of code we can find the buffers which hold these hidden layer dumps? (for test_bench)
2) Do buffers allocated to hold the hidden-layer dumps has a life equal to the single-frame processing?
3) If Yes, do we have sufficient internal buffer for storing hidden layers dumps for next frame?
Kindly help us and do the needful.
Thanks and Regards,
Vyom Mishra
