Open OliverZijia opened 2 years ago
Hi Oliver - that's weird, let's see if we can figure out what's going on. It may take some poking around to see exactly what the problem is given that there's no error message, but I have some ideas.
First, just to be sure, can you confirm:
imputer.train(...)
UNet(...)
and the FastSHAP wrapper with FastSHAP(...)
If all of the above is fine, I'll get to my first guess. In my experience, problems that lead to the kernel dying without an error message typically involve some incorrect operation on the GPU. Would you be able to try moving the two DNNs (the model
and explainer
objects) to CPU and see what happens when you train? It will be quite slow, but maybe we'll see an error message within the first couple training steps.
Hi!
I got this notification while running this line in cifar_single_model.ipynb:
fastshap.train( fastshap_train, fastshap_val, batch_size=128, num_samples=2, max_epochs=200, eff_lambda=1e-2, validation_samples=1, lookback=10, bar=True, verbose=True)
I reinstall the fastshap, but the issue remains there.
Does anyone know what should I do?
Thanks!