# Use TF2 saved\_model.pb with OpenCV dnn module

**URL:** <https://forum.opencv.org/t/use-tf2-saved-model-pb-with-opencv-dnn-module/4425>\
**Category:** Python\
**Tags:** dnn\
**Created:** [July 22, 2021, 12:14pm UTC](https://forum.opencv.org/t/use-tf2-saved-model-pb-with-opencv-dnn-module/4425 "2021-07-22T12:14:02Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![mata\_mare](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.opencv.org/mata_mare/32/2709_2.png) [@mata\_mare](https://forum.opencv.org/u/mata_mare)\
**Post date:** [July 22, 2021, 12:14pm UTC](https://forum.opencv.org/t/use-tf2-saved-model-pb-with-opencv-dnn-module/4425/1 "2021-07-22T12:14:02Z")

</div>

Recentryl I have trained mask\_rcnn\_inception\_resnet\_v2\_1024x1024\_coco17 with TF2 Object Detection API and exported the model in saved\_model.pb. Than the challenge was to actually use this model with OpenCV dnn module which is practically impossible. Is there any support in this area? What can we do so we can get support for TF2 saved\_model.pb ?