add some examples

app.py

@@ -1,14 +1,30 @@
 import gradio as gr
 import gem_metrics
+from datasets import load_dataset
+import random
 
-def calc_rouge_score(pred, gold):
+random.seed(42)
+
+raw_datasets = load_dataset("eli5")
+validation_len = len(raw_datasets["validation"])
+random_insts = [random.randint(validation_len)] 
+questions = []
+q2ref = {}
+for inst in random_insts:
+  question = raw_datasets["validation_eli5"][inst]["title"]
+  ref = raw_datasets["validation_eli5"][inst]["answers"]["text"]
+  questions.append(question)
+  q2ref[question] = ref
+
+def calc_rouge_score(og_question, pred):
     preds = gem_metrics.texts.Predictions(pred)
+    gold = q2ref[og_question]
     refs = gem_metrics.texts.References(gold)
     result = gem_metrics.compute(preds, refs, metrics_list=['bleu', 'rouge'])
     return result
 
 iface = gr.Interface(
     fn=calc_rouge_score,
-    inputs=["text", "text"],
+    inputs=[gr.Dropdown(questions), "text"],
     outputs=["text"])
 iface.launch()

requirements.txt

@@ -1 +1,2 @@
 git+https://github.com/GEM-benchmark/GEM-metrics.git
+datasets