updated LLM model

app.py

@@ -66,7 +66,7 @@ def initialize_pinecone(index_name: str):
 
 # Initialize Pinecone index and BM25 encoder
 pinecone_index = initialize_pinecone("updated-mbzuai-policies-17112024")
-bm25 = BM25Encoder().load("./new_mbzuai-policies.json")
+bm25 = BM25Encoder().load("./mbzuai-policies.json")
 
 ##################################################
 ##################################################
@@ -74,7 +74,7 @@ bm25 = BM25Encoder().load("./new_mbzuai-policies.json")
 # old_embed_model = HuggingFaceEmbeddings(model_name="sentence-transformers/gte-multilingual-base")
 
 # Initialize models and retriever
-embed_model = HuggingFaceEmbeddings(model_name="GameScribes/stella_en_400M_v5", model_kwargs={"trust_remote_code":True})
+embed_model = HuggingFaceEmbeddings(model_name="jinaai/jina-embeddings-v3", model_kwargs={"trust_remote_code":True})
 retriever = PineconeHybridSearchRetriever(
     embeddings=embed_model, 
     sparse_encoder=bm25, 
@@ -86,7 +86,7 @@ retriever = PineconeHybridSearchRetriever(
 
 # Initialize LLM
 # llm = ChatGroq(model="llama-3.1-70b-versatile", temperature=0, max_tokens=1024, max_retries=2)
-llm = ChatPerplexity(temperature=0, pplx_api_key=GROQ_API_KEY, model="llama-3.1-sonar-large-128k-online", max_tokens=1024, max_retries=2)
+llm = ChatPerplexity(temperature=0, pplx_api_key=GROQ_API_KEY, model="llama-3.1-sonar-large-128k-chat", max_tokens=1024, max_retries=2)
 
 
 # Initialize Reranker