Closed islam-nassar closed 1 year ago
Hi there,
I hit a bug while trying to evaluate our system. Here, if ref dataframe is empty, the return value does not include "type". However, here, it is expected to have "type" so it throws an error. To fix this, I just edited L80 to the following:
ref
return pd.DataFrame(data=[[row.Class, row.type, row.file_id, np.nan, np.nan, row.start, row.end, row.llr, 0.0]], columns=['Class', 'type', 'file_id', 'start_ref', 'end_ref', 'start_hyp', 'end_hyp', 'llr', 'IoU'])
Thanks for the report. I've fixed this in the upcoming release. The commit id is 3c12a5660fb42b8765bb2edf74e2e5525b792d4b.
Hi there,
I hit a bug while trying to evaluate our system. Here, if
ref
dataframe is empty, the return value does not include "type". However, here, it is expected to have "type" so it throws an error. To fix this, I just edited L80 to the following:return pd.DataFrame(data=[[row.Class, row.type, row.file_id, np.nan, np.nan, row.start, row.end, row.llr, 0.0]], columns=['Class', 'type', 'file_id', 'start_ref', 'end_ref', 'start_hyp', 'end_hyp', 'llr', 'IoU'])