Custom layer TanH detected but not introduced to ModelOptimizer

[Murugappan__Indrajit]

Hi

I'm trying to convert a caffe model to CVSDK IR file. My network has the following layers: Convolution, Inner Product, Pooling, ReLU, Softmax and TanH. As per your documentation, all the above mentioned layers are recognized by the Model Optimizer. But when I try to convert it, I get this message:

"Warning, custom layer TanH detected but not introduced to ModelOptimizer

Falling back to default layer definitions..."

Could you please tell me what might be wrong?

Thanks

[Zhen_Z_Intel · Accepted solution]

Hi,

For current beta version of inference engine, we still do not support tanh and sigmoid activation. 

[Murugappan__Indrajit]

Thank you for your reply