Update app.py

app.py

@@ -20,22 +20,23 @@ def embed_text(texts):
         embeddings = model(**inputs).last_hidden_state.mean(dim=1)
     return embeddings.numpy()
 
+def embedding_function(texts):
+    # This function converts texts to embeddings using the Hugging Face model
+    embeddings = embed_text(texts)
+    return embeddings
+
 if "vector" not in st.session_state:
     st.session_state.loader = WebBaseLoader("https://docs.nvidia.com/cuda/")
     st.session_state.docs = st.session_state.loader.load()
 
     st.session_state.text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)
     documents = st.session_state.text_splitter.split_documents(st.session_state.docs[:50])
-    
-    # Generate embeddings for documents
-    doc_texts = [doc.page_content for doc in documents]
-    embeddings = embed_text(doc_texts)
-    
-    # Convert embeddings to list of lists for FAISS
-    embeddings = embeddings.tolist()
 
-    # Create FAISS index from documents and embeddings
-    st.session_state.vectors = FAISS.from_documents(documents, embeddings)
+    # Create FAISS index using the custom embedding function
+    st.session_state.vectors = FAISS.from_texts(
+        [doc.page_content for doc in documents], 
+        embedding_function
+    )
 
 st.title("ChatGroq Demo")
 groq_api_key = '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'