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.

supported layers

idata
Employee
1,955 Views

Hi,

 

I'm testing few models with NCS at the moment. I found out the layers supported by mvNCCompile is quite limited.

 

Can you provide what full list of layers supported by your product in each framework (caffe/tensorflow)?

 

I wasn't able to find it in the dev documents.

 

Also, I see 'concat' layer is not supported atm, Is there any time-plan when it is going to be updated?

 

Thanks
0 Kudos
13 Replies
idata
Employee
1,584 Views

Supported layers are listed in the latest release notes: https://github.com/movidius/ncsdk/releases/tag/v1.10.01.01

 

Currently includes (according to the release notes)

 

     

  1. Convolutions

       

    • NxN Convolution with Stride S.
    •  

    • The following cases have been extensively tested: 1x1s1,3x3s1,5x5s1,7x7s1, 7x7s2, 7x7s4
    •  

    • Group convolution
    •  

    • Depth Convolution
    •  

    • Dilated convolution
  2.  

  3. Max Pooling Radix NxM with Stride S
  4.  

  5. Average Pooling: Radix NxM with Stride S, Global average pooling
  6.  

  7. Local Response Normalization
  8.  

  9. Relu, Relu-X, Prelu (see erattum #10)
  10.  

  11. Softmax
  12.  

  13. Sigmoid
  14.  

  15. Tanh (see erratum #10)
  16.  

  17. Deconvolution
  18.  

  19. Slice
  20.  

  21. Scale
  22.  

  23. ElmWise unit : supported operations - sum, prod, max
  24.  

  25. Fully Connected Layers (limited support -- see erratum #8)
  26.  

  27. Reshape
  28.  

  29. Flatten
  30.  

  31. Power
  32.  

  33. Crop
  34.  

  35. ELU
  36.  

  37. Batch Normalization
  38.  

0 Kudos
idata
Employee
1,584 Views

@tf_movidius Thank you very much!

0 Kudos
idata
Employee
1,584 Views

Hi,

 

My net uses a tf.slice operator on a tensor and I get [Error 5] Toolkit Error: Stage Details Not Supported: Slice type not supported while running mvNCCompile on it.

 

Could you help me with this?

0 Kudos
idata
Employee
1,584 Views

@gokul_uf Yes. We're working to add as much support as we can for all of the different neural networks out there, so if you can provide a link to your network, we can help with debugging issues.

0 Kudos
idata
Employee
1,584 Views

@Tome_at_Intel The net is at model.py here https://github.com/SpaceML/GalaxyGAN_python/

 

I had modified the net to have both image and cond passed in as single tensor and slice it inside to get around the only one input limitation. That's where I got the Slice issue.

 

I later realized that I needed only cond so I reverted back to the original code and I get the following error.

 

[Error 5] Toolkit Error: Stage Details Not Supported: Unsupported Mean operation

 

It'd be great if you could help me with this, because I have a demo coming up :smile:

0 Kudos
idata
Employee
1,584 Views

@gokul_uf Since I cannot find the network's meta file from the link., it looks like I am required to download their data and train to obtain the network. It would save me a lot of time if you could provide me the model's meta file.

0 Kudos
idata
Employee
1,584 Views

@Tome_at_Intel Apologies, my bad! The meta files are at https://drive.google.com/drive/folders/1yoqOVZIAfm3UpVAqWq3Ps-6cmpgd_1Cr?usp=sharing

 

the non_bnorm one has the bnorm op removed (because we use just 1 image per training so no point having bnorm)

 

the non_unet one has the skip links removed in the generator method (this uses concat and I saw that concat is buggy in ncsdk)

 

the input node name is cond and output is gen/gen_op.

 

Thanks!

0 Kudos
idata
Employee
1,584 Views

@gokul_uf I need the checkpoint file also if you could provide it. I am getting an error saying : Failed to find any matching files for non_bnorm_model_1.ckpt. Thanks!

0 Kudos
idata
Employee
1,584 Views

Hi, are you planning to release support for 3dconvolutions/ 3d Max pooling any time soon?

 

Thanks!
0 Kudos
idata
Employee
1,584 Views

@gokul_uf Can you try your network with the latest SDK (1.11) and let me know if this fixes your issue?

 

@JMat I can't give a specific timeline for you regarding the support for 3d convolutions, but as of now, it is something we are considering and looking into. Thanks.
0 Kudos
idata
Employee
1,584 Views

@gokul_uf I had issues too, according to /opt/movidius/NCSDK/ncsdk-x86_64/tk/Controllers/TensorFlowParser.py, line 1489:

 

if (len(input_shape) != 4 or len(slicingbegin) != 4 or len(slicingsize) != 4 or

 

slicingbegin[0] != 0 or slicingbegin[1] != 0 or slicingbegin[2] != 0 or

 

input_shape[0] != slicingsize[0] or input_shape[1] != slicingsize[1] or input_shape[2] != slicingsize[2]):

 

You can slice only on 4th dimension (image channels). Moreover, you can't use "-1" indexing you can use on base Tensorflow.

0 Kudos
idata
Employee
1,584 Views

Is GRU supported?

0 Kudos
idata
Employee
1,584 Views

@sreerag_ibtl,

 

I don't think GRU is supported as currently the NCSDK does not have support for RCNNs, hope this helps.

 

Regards,

 

@Luis_at_Intel
0 Kudos
Reply