High Accuracy but F1 Score is 0.0 on UK-DALE Dataset Using Provided Code

[Badar-e-Alam]

I am using the code for the UK-DALE dataset to reproduce the results. From the initial experiments, the accuracy is quite high, but the F1 score is consistently 0.0. I have used both the provided code, which calculates the F1 score manually per batch, and the torch.metrics.F1Score class. Any hint what could be reasons.

FYI, I've also modified the following code for the in the load_data function.

                  if house_id == self.house_indicies[0]:
                              entire_data = house_data
                    else:
                             entire_data = pd.concat([entire_data, house_data], axis=0)

Please provide any insights or suggestions on this issue. Thank you