inference an uff file using int8

Autonomous Machines Jetson & Embedded Systems Jetson TX2
1

I convert an tensorflow model to uff. How can I do inference in Int8 model using the uff file?

2

Hi,

INT8 is a feature for GPU architecture 6.1 and 7.x.
Jetson TX2 is sm=6.2 which cannot support INT8 inference.

Thanks.

3

Hi AastaLLL:

I want to do it on 1060

4

Hi,

Suppose you can control it with FieldType.
Check our document for details:
[url]API Reference :: NVIDIA Deep Learning TensorRT Documentation

Thanks.