Object Detection Error : Faster R-CNN ResNet50 W/ GPU

General Discussion
,
#1

My first posting on this forum. Please forgive me if I’m in the wrong place, and any suggestions for the proper way to ask would be appreciated.

I’m self-teaching Tensorflow Object Detection in hopes of being able to identify and count nesting birds from UAV images using available tutorials.

I am able to train and detect with all the various models in the Tensorflow Zoo using my tutorial base setup except the Faster R-CNN versions.

On any Faster R-CNN model I first get a new message :

“This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX AVX2
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.”

Followed by a failure :

“raise ValueError(’{} is not supported. See model_builder.py for features '”

I suspect this comes up from some sort of a version error between the tutorial environment and the Faster R-CNN pre-trained model, but I don’t know where to start to find a path forward.

Here is the pre-trained model I’m using…

‘http://download.tensorflow.org/models/object_detection/tf2/20200711/faster_rcnn_resnet50_v1_640x640_coco17_tpu-8.tar.gz’

Any suggestions would be appreciated, including any recommendations of where I might find a consulting service to assist.

Thanks in advance

#4

Windows 10, NVidia 3080TI

#6

Hi @Hank_Arnold , Welcome to the forum. The hyper - link shared by you is not working.

#7

It’s not a hyper-link, its the download link for the pre-trained model from the Tensoflow Model Zoo. If my problem is some sort of a version or interface problem, that link is what’s going to have to change.

Thanks for responding