TensorRT UFF parser warning keepdims is ignored by the UFF Parser and defaults to True

I am working on TensorRT and I am converting my model to uff format. I am able to convert the model with warning message as “keepdims is ignored by the UFF Parser and defaults to True”. When I try to run inference using C++ API, at runtime I get the error as “add_1/add: elementwise inputs must have same dimensions or follow broadcast rules (input dimensions were [13,20,31] and [13,20,30])”.

I am aware that all the warning from uff converter can not be ignored. But I would like to know how to resolve this issue ?

(Note: I use Transpose convolution layers in the model)

1 Like

I am facing the same issue. I am using a Resnet implementation provided here: https://github.com/raghakot/keras-resnet

Tensorflow version:1.14.0
Keras version: 2.2.4
Tensorrt version:

I am using older version of tensorflow for legacy reasons.

Your help is appreciated.