Problem with combing quantisation and pruning

[DoubleZZeta]

Hi, I was trying to extend the tutorial6 code to support all types of linear layers and then pass the quantised model to the next NAS search to find the best pruning configuration. However, when I try to pass a model that is quantised using torch.nn.linear and LinearMinifloatDenorm to the pruning NAS search, I got some errors as shown in the picture.

For reference, here is how I extend the code to make it support LinearMinifloatDenorm

Thanks.