torch.eq Not Supported, Ignored During Conversion and Causes Runtime Error

[lufficc]

Hi,

I encountered an issue when converting a PyTorch model to NCNN. The model contains the torch.eq operation, which is ignored during conversion, as indicated by the following message:

ignore torch.eq torch.eq_28 param other=1.000000e+00

At runtime, this leads to the following error:

layer Tensor.eq not exists or registered

It seems that NCNN doesn't support torch.eq operation . I also tried replacing torch.eq with the equality operator == (e.g., tensor == 1.00), but it resulted in the same issue. Could you please provide support for these operations or suggest a workaround to handle this issue?

Thank you for your help!

### Tasks

[Baiyuetribe]

#5890