VX Delegate issue - WARNING: Fallback unsupported op 32 to TfLite

[dpcivl]

Hello. I'm doing my toy project with i.MX 8M PLUS evk board.

I'm in trouble executing my python file which runs object detection on vx delegate.

As I used vx delegate, the following message came up.

"WARNING: Fallback unsupported op 32 to TfLite"

Please tell me what I should do solve this problem.

Should I fix the model? or python code? I don't know at all.

I want to run my program on vx delegate. (with no using CPU)

Here is chunk of logs when I run my python code.

//

Vx delegate: allowed_cache_mode set to 0.
Vx delegate: device num set to 0.
Vx delegate: allowed_builtin_code set to 0.
Vx delegate: error_during_init set to 0.
Vx delegate: error_during_prepare set to 0.
Vx delegate: error_during_invoke set to 0.
WARNING: Fallback unsupported op 32 to TfLite
INFO: Created TensorFlow Lite XNNPACK delegate for CPU.

//

[Mohan51]

Hi @dpcivl  have you able to build Vx delegate successfully, I am trying to build Vx delegate but unable to build it by following the official NXP IMX-machine-learning-UG. if you have any sources could you please attach here then i can reference the same for building the Vx delegate.and my os is  Linux

[Zhiming_Liu · Accepted solution]

You need make uint8/int8 quantization with eIQ tool or yourself tool.

https://www.nxp.com/design/software/eiq-ml-development-environment/eiq-toolkit-for-end-to-end-model-...