Intel® Distribution of OpenVINO™ Toolkit
Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.
6447 Discussions

Creating an IR model from a tensorflow (1.15.0) model fails

Yumi_ques
Novice
588 Views

I am trying to generate an IR model from a tensorflow(1.15.0) model provided by Indoor-segmentation.
(Referring to Converting a TensorFlow* Model)
However, an error is output and the IR model fails to generate.

Could you please tell us how to solve this problem?
I am trying to create from a METAGRAPH.

Thank you in advance.

◇environment
Ubuntu:Ubuntu 18.04.5 LTS
OpenVINO 2022.1.0.643
tensorflow:2.9.1


[Execution details]
mo --input_meta_graph ./pretrained_models/ResNet101/model.ckpt-100000.meta

 

[Error Description]
[ ERROR ] Exception occurred during running replacer "REPLACEMENT_ID" (<class 'openvino.tools.mo.load.tf.loader.TFLoader'>): Unexpected exception happened during extracting attributes for node LessEqual/y.
Original exception message: uint8_val

 

◇Indoor-segmentation

https://github.com/hellochick/Indoor-segmentation

◇Tensorflow models are downloaded from

https://drive.google.com/file/d/1o7QrlNxH6BX6uYatlR06-A_cutWD9sNg/view

◇Converting a TensorFlow* Model
https://docs.openvino.ai/latest/openvino_docs_MO_DG_prepare_model_convert_model_Convert_Model_From_TensorFlow.html#convert-tensorflow-1-models

 

 

0 Kudos
1 Solution
Yumi_ques
Novice
530 Views


Thanks for the reply.

I understand that it is an error when it is not NCDHW.

Because there is no mention of output layout in indoor segmentation
I don't know if the output layout is NCDHW. However, since it is an error, I don't think it is NCDHW.

I close this question as resolved.
Thank you for your response.

 

Yumi_ques

View solution in original post

0 Kudos
4 Replies
IntelSupport
Community Manager
562 Views

Hi Yumi_ques,

 

Thanks for reaching out.

We are checking on this and will update you with the information soon.


 

Regards,

Aznie


0 Kudos
IntelSupport
Community Manager
548 Views

Hi Yumi_ques,

 

Our documentation on the supported output layout [Supported Devices — OpenVINO™ documentation — Version(latest)] indicated that LessEqual (Eltwise-LessEqual) is supported for CPU if the layout is NCDHW (refer to the bottom of the page).

 

If the model is not NCDHW layout, the error is expected.

 

 

Regards,

Aznie


0 Kudos
Yumi_ques
Novice
531 Views


Thanks for the reply.

I understand that it is an error when it is not NCDHW.

Because there is no mention of output layout in indoor segmentation
I don't know if the output layout is NCDHW. However, since it is an error, I don't think it is NCDHW.

I close this question as resolved.
Thank you for your response.

 

Yumi_ques

0 Kudos
IntelSupport
Community Manager
520 Views

Hi Yumi_ques,


This thread will no longer be monitored since this issue has been resolved. If you need any additional information from Intel, please submit a new question. 



Regards,

Aznie


0 Kudos
Reply