TIDL MUL/ADD operator support

Hello,

Which device is this for? For example, TDA4VM, TDA4VL/VE/AL, J784S4, etc.

Regards,

Erick

TDA4VM 

Hi Abhilash,

     The two warnings are independent warnings. 

abhilash sk said:

For mul/add with a constant tensor as one of the inputs, the shape and size of constant tensor and input tensor are different

    We only support constant tensor of same size

abhilash sk said:

“WARNING: Batch Norm Layer coeff cannot be found(or not match) in coef file”

This is an independent warning indicating that for a given layer corresponding weights are not available in model. Are you using Caffe models?

Regards,

Anshu

I am using onnx model and the whole warning is for one operator only 

This warning comes for the batchnorm layer of the number of parameter doesn't match the tensor size.

Would it be possible to share a representative model

Apologies sir couldnot share the model but can share the model part details pls find below