Update app.py

app.py

@@ -17,7 +17,6 @@ from langchain.prompts import PromptTemplate
 from langchain import hub
 import os
 import glob
-import gc
 
 
 # TEXT LOADERS
@@ -103,15 +102,15 @@ def get_vectorstore(text_chunks, embeddings):
 
 def get_conversation_chain(vectorstore):
 
-    model_path = "models/llama-2-13b-chat.Q4_K_S.gguf"
+    model_path = st.session_state.model
     callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])
 
     llm = LlamaCpp(model_path= model_path, 
                    n_ctx=4000, 
                    max_tokens= 500,
-                   n_gpu_layers = 50, 
+                   fp = 50, 
                    n_batch = 512,
-                   callback_manager = callback_manager
+                   callback_manager = callback_manager,
                    verbose=True)
     
     memory = ConversationBufferMemory(
@@ -123,18 +122,10 @@ def get_conversation_chain(vectorstore):
     Try to summarise the content and keep the answer to the point.
     If you don't know the answer, just say that you don't know, don't try to make up an answer.
 
-    Followe the template below
-    Example: 
-    Question : how many paid leaves do i have ?
-    Answer : The number of paid leaves varies depending on the type of leave, like privilege leave you're entitled to a maximum of 21 days in a calendar year. Other leaves might have different entitlements. thanks for asking!
-    make sure to add "thanks for asking!" after every answer 
 
-    {context}
-    
+    Context: {context}
     Question: {question}
     Answer:
-
-    Just answer to the point! 
     """
 
     rag_prompt_custom = PromptTemplate.from_template(template)
@@ -189,11 +180,11 @@ def add_rounded_edges(image_path="./randstad_featuredimage.png", radius=30):
 
 def main():
     load_dotenv()
-    gc.collect()
     st.set_page_config(page_title="Chat with multiple Files",
                        page_icon=":books:")
     st.write(css, unsafe_allow_html=True)
 
+
     if "conversation" not in st.session_state:
         st.session_state.conversation = None
     if "chat_history" not in st.session_state:
@@ -201,20 +192,29 @@ def main():
 
     st.title("💬 Randstad HR Chatbot")
     st.subheader("🚀 A HR powered by Generative AI")
+
+    # default model 
+    st.session_state.model = "./models/mistral-7b-instruct-v0.2.Q5_K_M.gguf"
     # user_question = st.text_input("Ask a question about your documents:")
 
     st.session_state.embeddings = FastEmbedEmbeddings( model_name= "BAAI/bge-small-en-v1.5", 
                                                                     cache_dir="./embedding_model/")
 
-    if len(glob.glob("./vectordb/*.sqlite3")) > 0:
+    if len(glob.glob("./vectordb/*.sqlite3")) > 0 :
         
         vectorstore = Chroma(persist_directory="./vectordb/", embedding_function=st.session_state.embeddings)
         st.session_state.conversation = get_conversation_chain(vectorstore)
         handle_userinput()
 
     with st.sidebar:
+
+        # calling a 
         add_rounded_edges()
 
+        st.subheader("Select Your Embedding Model Model")
+        st.session_state.model = st.selectbox( 'Models', tuple( glob.glob('./models/*.gguf') ) )
+
+
         st.subheader("Your documents")
         docs = st.file_uploader(
             "Upload File (pdf,text,csv...) and click 'Process'", accept_multiple_files=True)
@@ -257,6 +257,4 @@ if __name__ == '__main__':
         print("Command executed successfully.")
     except subprocess.CalledProcessError as e:
         print(f"Error: {e}")
-        
-    main()
-    
+    main()