Update app.py

app.py

@@ -18,9 +18,9 @@ embeddings = torch.tensor(df['embedding'].tolist(), device=device)
 # Load the Sentence Transformer model
 model = SentenceTransformer('all-MiniLM-L6-v2', device=device)
 
-# Load the LLaMA model for response generation
-llama_tokenizer = AutoTokenizer.from_pretrained("openai-community/gpt2")
-llama_model = AutoModelForCausalLM.from_pretrained("openai-community/gpt2").to(device)
+# Load the ai model for response generation
+ai_tokenizer = AutoTokenizer.from_pretrained("openai-community/gpt2")
+ai_model = AutoModelForCausalLM.from_pretrained("openai-community/gpt2").to(device)
 
 # Define the function to find the most relevant document
 @spaces.GPU(duration=120)
@@ -35,9 +35,9 @@ def retrieve_relevant_doc(query):
 def generate_response(query):
     relevant_doc = retrieve_relevant_doc(query)
     input_text = f"Document: {relevant_doc}\n\nQuestion: {query}\n\nAnswer:"
-    inputs = llama_tokenizer(input_text, return_tensors="pt").to(device)
-    outputs = llama_model.generate(inputs["input_ids"], max_length=1024)
-    response = llama_tokenizer.decode(outputs[0], skip_special_tokens=True)
+    inputs = ai_tokenizer(input_text, return_tensors="pt").to(device)
+    outputs = ai_model.generate(inputs["input_ids"], max_length=1024)
+    response = ai_tokenizer.decode(outputs[0], skip_special_tokens=True)
     return response
 
 # Create a Gradio interface