How to compute metrics for multitask learning?

Hi! I’m new to DeepChem and to deep learning in general. I am trying to compute evaluation metrics for multitask learning such as accuracy, precision, recall, f1 score and MCC using the following code:

y_true = test_dataset.y
y_pred = model.predict(test_dataset)
n_tasks = len(tasks)
for i in range(n_tasks):
score = dc.metrics.f1_score(y_true, y_pred)
print(tasks[i], score)

But got this error:
ValueError: Classification metrics can’t handle a mix of multilabel-indicator and unknown targets

Here are my questions:

  1. How to go around with the error in order to calculate previously mentioned metrics?
  2. How to create plots for precision-recall curve and ROC curve for multitask learning in DeepChem?