[Arrow/i.MX8MN/Linux L5.4.3_1.0.0]Extension feature support from TensorFlow lite Engine

cancel
Showing results for 
Search instead for 
Did you mean: 

[Arrow/i.MX8MN/Linux L5.4.3_1.0.0]Extension feature support from TensorFlow lite Engine

Jump to solution
210 Views
Lambert
NXP Employee
NXP Employee

Project info are below shown:

Software: BSP 5.4.3_1.0.0  (tensorflow + keras)

Hardware: imx8m nano, pico zense Dcam710(USB)

Image process workflow:

TOF frame (640*480 UMat) à resize (64*64 UMat) à gray (64*64 UMat) à MSE(UMat) à tflite engine process(UMat) à result (float)

 

  1. Extension feature support from TensorFlow lite Engine

We use Convolution 3D, Max-pooling 3D, UpSampling 3D and Deconvolution 3D layers for DSTCAE model. The model file is h5 format.

Convert the h5 model to C3D_AE.tflite by tflite_convert function. The C3D_AE.tflite is running well in PC platform with tensorflow lite c++.

But when use the same tensorflow lite c++ code with C3D_AE.tflite in the imx8mn. The imx8mn said Tensorflow ops are not supported.

Detail logs are below shown:

 

sh-5.0# ./bin tfliteModel/C3D_AE.tflite

Loading TFLite model

[2020-04-10 03:23:33.682525] [0x0000ffff962149b0] [info]    Allocate tensor buffer

Regular TensorFlow ops are not supported by this interpreter. Make sure you invoke the Flex delegate before inference.

Node number 0 (Flex) failed to prepare.

 

terminate called after throwing an instance of 'std::runtime_error'

  what():  Failed to allocate tensor buffer

Aborted

 

 

Labels (1)
0 Kudos
1 Solution
197 Views
DevinJiao
NXP Employee
NXP Employee

Hi Lambert,

What's the TFLite version customer used on PC? Can the model be shared with us?

View solution in original post

0 Kudos
1 Reply
198 Views
DevinJiao
NXP Employee
NXP Employee

Hi Lambert,

What's the TFLite version customer used on PC? Can the model be shared with us?

View solution in original post

0 Kudos