Turning off memory

app.py

@@ -32,11 +32,11 @@ def chat_with_distilgpt2(input_text):
     conversation_history = memory.load_memory_variables({})['history']
     
     # Combine the (possibly summarized) history with the current user input
-    full_input = f"{conversation_history}\nUser: {input_text}\nAssistant:"
+    #full_input = f"{conversation_history}\nUser: {input_text}\nAssistant:"
     no_memory_input = f"Question: {input_text}\nAnswer:"
     
     # Tokenize the input and convert to tensor
-    input_ids = tokenizer.encode(full_input, return_tensors="pt").to(device)
+    input_ids = tokenizer.encode(no_memory_input, return_tensors="pt").to(device)
     
     # Generate the response using the model with adjusted parameters
     outputs = model.generate(