Update app.py

app.py

@@ -1,16 +1,20 @@
 import gradio as gr
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 import torch
+import datasets
 
 # Load your fine-tuned model and tokenizer
 model_name = "legacy107/flan-t5-large-bottleneck-adapter-cpgQA-unique"
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
 model.set_active_adapters("question_answering")
-
 max_length = 512
 max_target_length = 128
 
+# Load your dataset
+dataset = datasets.load_dataset("minh21/cpgQA-v1.0-unique-context-test-10-percent", split="test[:10]")
+
+
 # Define your function to generate answers
 def generate_answer(question, context):
     # Combine question and context
@@ -22,7 +26,7 @@ def generate_answer(question, context):
         return_tensors="pt",
         padding="max_length",
         truncation=True,
-        max_length=512,
+        max_length=max_length,
     ).input_ids
 
     # Generate the answer
@@ -34,6 +38,17 @@ def generate_answer(question, context):
 
     return generated_answer
 
+
+# Define a function to list examples from the dataset
+def list_examples():
+    examples = []
+    for example in dataset:
+        context = example["context"]
+        question = example["question"]
+        examples.append([context, question])
+    return examples
+
+
 # Create a Gradio interface
 iface = gr.Interface(
     fn=generate_answer,
@@ -41,7 +56,8 @@ iface = gr.Interface(
         gr.inputs.Textbox(label="Question"),
         gr.inputs.Textbox(label="Context")
     ],
-    outputs=gr.outputs.Textbox(label="Generated Answer")
+    outputs=gr.outputs.Textbox(label="Generated Answer"),
+    examples=list_examples()
 )
 
 # Launch the Gradio interface