Tool/software:
Hello,
When performing inference on the Cloud or EVM, I consistently observe a very high DDR usage per image, while my colleagues reported only 0 MB two years ago.
I have tried various models on both the Cloud and EVM, and I always receive very high results.
On cloud
On EVM
Old results with same model :
Are these results typical, or has the calculation method of edgeai_tidl_tools not been updated?
Thanks and regards,
Azer
