Intel® Distribution of OpenVINO™ Toolkit
Community support and discussions about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all things computer vision-related on Intel® platforms.

Model Optimizations for EfficientNet Lite?

automationbestest
233 Views

Hi,

 

I apologize in advance, I'm a bit new to the scene. I recently purchased an iNCS2 for a project I'm working on, because I have a few tflite models I would like to run faster on my raspberry pi.

 

I believe I have correctly set up my raspberry pi by following the directions on the Intel website, and I have also installed the OpenVINO toolkit for my Windows 11 PC through Anaconda. Now, I'm wondering what are the next steps to optimizing my models? I've been looking around the forums and online and I'm seeing a lot of support for TF models, but not so many TFLite models. Help would be greatly appreciated! As I said, I am new to this, so please bear with me.

 

Thanks!

0 Kudos
2 Replies
Hairul_Intel
Moderator
189 Views

Hi automationbestest,

Thank you for reaching out to us.

 

Here is the comprehensive guide on how to Convert TensorFlow Model using Model Optimizer. Additionally, here is the list of Supported Topologies for TensorFlow framework for OpenVINO.

 

Model Optimizer does not support .tflite file to IR format conversion natively. Model Optimizer currently only supports few frozen quantized topologies hosted on TensorFlow Lite site. This information is available in the subsection "Supported Frozen Quantized Topologies" under the section Supported Topologies.

 

However, the frozen model file (.pb file) is required for conversion to IR format by Model Optimizer. It is also necessary to use these specific command line parameters for the Model Optimizer (HEIGHT and WIDTH are the input images height and width for which the model was trained):

  • --input input
  • --input_shape [1,HEIGHT,WIDTH,3]

 

These information are also mentioned at the above link.

 

 

Regards,

Hairul


Hairul_Intel
Moderator
129 Views

Hi automationbestest,

This thread will no longer be monitored since we have provided a solution. If you need any additional information from Intel, please submit a new question.

 

 

Regards,

Hairul


Reply