I'm handling Openvino 2021 and overall mechanism is quite changed from 2019. So I reallly need to get a kind of new example. Can you please provide more recent one?

Hi,

first of all, thanks for this tutorial. While following the instructions, I ran into one problem at the very last step when executing the optimization with openvino.

[ ERROR ]  Cannot pre-process TensorFlow graph after reading from model file "None". File is corrupt or has unsupported format. Details: 'Graph' object has no attribute 'node'.
 For more information please refer to Model Optimizer FAQ (https://docs.openvinotoolkit.org/latest/_docs_MO_DG_prepare_model_Model_Optimizer_FAQ.html), question #44.

I executed the command as described in the instructions:
python %OV%\deployment_tools\model_optimizer\mo_tf.py --input_meta_graph model.ckpt.meta --batch 1 --output "ModCosh/Activation_8/softmax_output" --extensions %CLWS%\cl_cosh\user_mo_extensions --output_dir %CLWS%\cl_ext_cosh

Model files are presented in the folder as described in the instructions.

I am pretty new to openvino. So, what is the reason for this error and how can I solve it?

I appreciate any help on this.

I want to implement a LSTM layer with peephole in OpenVINO R1, the formula is here:
https://www.paddlepaddle.org.cn/documentation/api/zh/0.11.0/v2/config/layer.html#lstmemory
I try to use the Model Extension Generator but have no idea to fill in these parameters

Hi! Nicely written tutorial for CPU and GPU. I wanted to ask that is there also a tutorial for implementing custom layers for NCS2 or VPU. I there is any can you point me yo such resource? thanks.