Update app.py

app.py

@@ -52,13 +52,13 @@ def chunk_text(text, chunk_size=500, overlap=100):
         chunks.append(current_chunk.strip())
     return chunks
 
-def generate_with_gpt(prompt, max_length=150):
+def generate_with_gpt(prompt, max_new_tokens=100):
     """Generate text with GPT model"""
-    inputs = gpt_tokenizer(prompt, return_tensors="pt")
+    inputs = gpt_tokenizer(prompt, return_tensors="pt", truncation=True, max_length=512)
     with torch.no_grad():
         outputs = gpt_model.generate(
             inputs.input_ids,
-            max_length=max_length,
+            max_new_tokens=max_new_tokens,   # FIXED
             num_return_sequences=1,
             no_repeat_ngram_size=2,
             do_sample=True,
@@ -76,7 +76,7 @@ def refine_answer_with_gpt(context, question, answer):
         f"Answer: {answer}\n\n"
         f"Please provide a clearer and more complete answer in simple language."
     )
-    return generate_with_gpt(prompt, max_length=120)
+    return generate_with_gpt(prompt, max_new_tokens=120)
 
 def extract_direct_definition(text, term):
     """Find a direct definition of a term in the text"""
@@ -169,4 +169,4 @@ with gr.Blocks() as demo:
         [question, chatbot]
     )
 
-demo.launch()
+demo.launch()