Intel® Distribution of OpenVINO™ Toolkit
Community support and discussions about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all things computer vision-related on Intel® platforms.
5961 Discussions

Combination of fp32 and int8 inference using OpenVINO

Sengodan__Nathiyaa
159 Views

I would like to run the inference for a network using OpenVINO, but some layers in the network using int8 and some in fp32.

We know that we can run the inference in int8 using Calibration tool. But is there any way to manually make some of the layers in the network run in fp32 mode itself after we calibrate the model ?. Like making changes in .xml file generated from Calibration tool , etc.,

0 Kudos
1 Reply
Shubha_R_Intel
Employee
159 Views

Dear Nathiyaa:

Looks like this is a duplicate question which I have answered i here:

https://software.intel.com/en-us/forums/computer-vision/topic/806590

Thanks,

Shubha

Reply