Update app.py

app.py

@@ -15,7 +15,7 @@ from langchain.memory import ConversationBufferMemory
 from langchain.chains import ConversationalRetrievalChain
 
 
-llm = Llama(model_path= hf_hub_download(repo_id="TheBloke/airoboros-l2-7B-gpt4-m2.0-GGUF", filename="airoboros-l2-7B-gpt4-m2.0.Q6_K.gguf"), n_ctx=2048) #download model from hf/ n_ctx=2048 for high ccontext length
+llm = Llama(model_path= hf_hub_download(repo_id="TheBloke/Llama-2-7B-32K-Instruct-GGUF", filename="llama-2-7b-32k-instruct.Q2_K.gguf"), n_ctx=2048) #download model from hf/ n_ctx=2048 for high ccontext length
 
 history = []
 
@@ -46,10 +46,10 @@ def generate_text(input_text, history):
     print("input ", input_text)
     temp =""
     if history == []:
-        input_text_with_history = f"INSTRUCTIONS:{pre_prompt}"+ "\n" + f"HR: {input_text} " + "\n" +" ISAAC:"
+        input_text_with_history = f"INSTRUCTIONS:{pre_prompt}"+ "\n" + f"HR: {input_text} " + "\n" +"ISAAC:"
     else:
         input_text_with_history = f"{history[-1][1]}"+ "\n"
-        input_text_with_history += f"HR: {input_text}" + "\n" +" ISAAC:"
+        input_text_with_history += f"HR: {input_text}" + "\n" +"ISAAC:"
     print("new input", input_text_with_history)
     output = llm(input_text_with_history, max_tokens=1024, stop=["<|prompter|>", "<|endoftext|>", "<|endoftext|> \n","ISAAC:","HR:","INSTRUCTIONS:","TASK:","BACKGROUND_INFORMATION:"], stream=True)
     for out in output: