Description
When performing hyperparmeter tuning with a list as val_data (for example to monitor the metrics on the training set as well as the val set during training), get an attribute error ("copy").
Proposed fix
In the objective_with_param function of tune.py, replace
score = scorer(nlp, val_data)
with :
score = scorer(nlp, val_docs)
as in trainer.py.