diff --git a/examples/run_trainier_seq2seq_huggingface.py b/examples/run_trainier_seq2seq_huggingface.py
index e800076..6ebfda8 100644
--- a/examples/run_trainier_seq2seq_huggingface.py
+++ b/examples/run_trainier_seq2seq_huggingface.py
@@ -30,7 +30,8 @@
 
 import nltk  # Here to have a nice missing dependency error message early on
 import numpy as np
-from datasets import load_dataset, load_metric
+from datasets import load_dataset
+import evaluate
 from sklearn.metrics import accuracy_score, f1_score, recall_score, precision_score, classification_report, confusion_matrix
 
 import transformers
@@ -562,7 +563,7 @@ def preprocess_function(examples):
     metric_name = regex.sub("eval_","",training_args.metric_for_best_model)
 #     metric_name = "f1" #"rouge" if data_args.task.startswith("summarization") else "sacrebleu"
     print ("[INFO] evlaute using ", metric_name, "score", "task name:", data_args.task)
-    # metric = load_metric("f1", cache_dir=cache_dir)
+    # metric = evaluate.Metric("f1", cache_dir=cache_dir)
 
     def postprocess_text(preds, labels):
         preds = [pred.strip() for pred in preds]