lewis

Results 2 issues of lewis

as I was running this ![image](https://github.com/user-attachments/assets/405a2e94-74dd-45a1-a6ce-79cee5ba1d00) I got ![image](https://github.com/user-attachments/assets/8fbdc1cc-a0cc-4ec3-afa5-fe6bcff0aadc) and I try to upgrade my torch_geometric ![image](https://github.com/user-attachments/assets/bafc3351-4494-4985-a00b-8c55143fecee) and it sill doesn't work. How can I get rid of this? help!

In your paper section 3.3 you said you calculate the regularization term using During sparsity training, you use the unnormalized importance. You only use the max normalized importance during pruning....