Fly-Pluche

model_attack is Resnet18 and the type is DataParallel. when I try this. ``` target_layer = model_attack.module.layer4 e = shap.GradientExplainer((model_attack, target_layer), background, local_smoothing=0) ``` ``` Exception has occurred: AttributeError 'Sequential' object...

Not training enough times

1

Stopped training after only running 6 times, although it displayed as RUING, there was no issue with memory usage. Ubuntu 22.04.3 LTS python 3.8.18 nni 3.0b2    ...

The time spent on each epoch is increasing

1

I added the flow block in the PointFlow to my model. Then, the bug happened. I tried `torch.cuda.empty_cache()` and `del recon_loss` , but not work. 

Don't have dgn_agent.py and gcn_agent.py.