unsupported operator ReduceL2

[chouxscream]

Hi, I have trouble converting my swin transformer v2 model from ONNX, as this model has ReduceL2 operator. I found out that ReduceL2 operator is not supported. In rknn-toolkit 1, there was a function named custom operator. But I could not find its equivalent function in rknn-toolkit 2. So, how can I run this model? My hardware is RK3568. Thanks in advance!

[BigerPatax]

I have the same problem, the current rknn-toolkit2 does not support ReduceL2 operator, did you solve it?

[chouxscream]

@BigerPatax Sadly, no. We decided to use resnet-based models instead.