Update app.py

app.py

@@ -1,3 +1,4 @@
+import os
 import pandas as pd
 import torch
 from sentence_transformers import SentenceTransformer, util
@@ -6,6 +7,8 @@ import json
 from transformers import AutoTokenizer, AutoModelForCausalLM, AutoModelForSequenceClassification
 import spaces
 
+CUDA_LAUNCH_BLOCKING=1
+
 # Ensure you have GPU support
 device = 'cuda' if torch.cuda.is_available() else 'cpu'
 
@@ -50,7 +53,7 @@ def generate_response(query):
     if intent == 0:  # Handle intent 0 (e.g., informational query)
         input_text = f"Document: {relevant_doc}\n\nQuestion: {query}\n\nAnswer:"
         inputs = tokenizer(input_text, return_tensors="pt").to(device)
-        outputs = model_response.generate(inputs["input_ids"], max_length=500)
+        outputs = model_response.generate(inputs["input_ids"], max_length=150)
         response = tokenizer.decode(outputs[0], skip_special_tokens=True)
     elif intent == 1:  # Handle intent 1 (e.g., opinion-based query)
         # Generate a response based on the detected intent