Migrate metric to Evaluate library for tensorflow examples (#18327)

* Migrate metric to Evaluate library in tf examples

Currently tensorflow examples use `load_metric` function from Datasets
library , commit migrates function call to `load` function to
Evaluate library.

Fix for #18306

* Migrate metric to Evaluate library in tf examples

Currently tensorflow examples use `load_metric` function from Datasets
library , commit migrates function call to `load` function to
Evaluate library.

Fix for #18306

* Migrate `metric` to Evaluate for all tf examples

Currently tensorflow examples use `load_metric` function from Datasets
library , commit migrates function call to `load` function to
Evaluate library.

examples/tensorflow/question-answering/requirements.txt

@@ -1,2 +1,3 @@
 datasets >= 1.4.0
 tensorflow >= 2.3.0
+evaluate >= 0.2.0

examples/tensorflow/question-answering/run_qa.py

@@ -26,8 +26,9 @@ from pathlib import Path
 from typing import Optional
 
 import tensorflow as tf
-from datasets import load_dataset, load_metric
+from datasets import load_dataset
 
+import evaluate
 import transformers
 from transformers import (
     AutoConfig,
@@ -600,7 +601,7 @@ def main():
         references = [{"id": ex["id"], "answers": ex[answer_column_name]} for ex in examples]
         return EvalPrediction(predictions=formatted_predictions, label_ids=references)
 
-    metric = load_metric("squad_v2" if data_args.version_2_with_negative else "squad")
+    metric = evaluate.load("squad_v2" if data_args.version_2_with_negative else "squad")
 
     def compute_metrics(p: EvalPrediction):
         return metric.compute(predictions=p.predictions, references=p.label_ids)