Define output format of evaluate function

[Wenshansilvia]

During evaluation of all metrics, detail log info should be written to local file using logger. Two variable should be returned by the evaluate function:

1. average score of each metric

2. instance level score of each metric

   >>> import rageval as rl
   >>> ds = load_dataset("testset_name")
   >>> result, instance_level_result = evaluate(
               ds.select(range(3)),
               metrics=[ContextRecall(), AnswerGroundedness()],
               models = [cr_model, ag_model]
           )
   >>> result
   Dataset({
       features: ['context_recall', 'answer_groundedness'],
       num_rows: 2
    })
   >>> instance_level_result
   Dataset({
       features: ['questions', 'gt_answers', 'answers', 'contexts', 'context_recall', 'answer_groundedness'],
       num_rows: 9
   }))

[faneshion]

Maybe dict is enough for result, hahh...