'SWA' object has no attribute '_optimizer_step_pre_hooks'

[wanzijing12]

When I run into ''model.fit(epochs=n_epochs, losscurve=True)" code, the AttributeError: 'SWA' object has no attribute '_optimizer_step_pre_hooks' occured. And 'data/ExperimentalHoldoutPreds.csv' was not found.

[SurgeArrester]

Hi @wanzijing12, apologies I didn't have the capacity to look into this when you commented. The ExperimentalHoldouts I've added to the data folder, but were accessible in the original paper. I think the predictions you will have to generate yourself using the dataset from the website given in the paper, due to licensing reasons from the university of liverpool.

Did you fix the SWA issue or is that something you still require assistance with?