Update app-llama.py

app-llama.py

@@ -60,14 +60,14 @@ from langchain.memory import ConversationBufferMemory
 from langchain_community.llms import HuggingFaceEndpoint
 from langchain_community.embeddings import HuggingFaceEmbeddings 
 
-"""
+
 # for hugging face llm
 from transformers import AutoTokenizer
 import transformers
 import torch
 import tqdm 
 import accelerate
-"""
+
 
 import pymupdf
 from PIL import Image
@@ -189,7 +189,9 @@ def build_qa_chain(collection_name, vector_db, file: str):
     documents, file_name = process_file2(file)
     # Load embeddings model
     #embeddings = OpenAIEmbeddings(openai_api_key=self.OPENAI_API_KEY)
-    
+
+    #vincent for old LLM
+    """
     embeddings = AzureOpenAIEmbeddings(
     model="text-embedding-ada-002",
     # dimensions: Optional[int] = None, # Can specify dimensions with new text-embedding-3 models   
@@ -198,11 +200,11 @@ def build_qa_chain(collection_name, vector_db, file: str):
     #openai_api_version="2023-05-15", # If not provided, will read env variable AZURE_OPENAI_API_VERSION
     openai_api_version="2023-05-15", # If not provided, will read env variable AZURE_OPENAI_API_VERSION
     )
+    """    
     
-    """
     #vincent for new LLM
     embeddings = HuggingFaceEmbeddings()
-    """
+    
     
     #vincent added to handle the tenant problem
     new_client = chromadb.EphemeralClient()
@@ -237,8 +239,8 @@ def build_qa_chain(collection_name, vector_db, file: str):
     )
     """
 
-    #vincent added self.chain
- 
+    #vincent added for old LLM
+    """
     chain = ConversationalRetrievalChain.from_llm(
     #ChatOpenAI(temperature=0.0, openai_api_key=self.OPENAI_API_KEY),
     
@@ -255,7 +257,7 @@ def build_qa_chain(collection_name, vector_db, file: str):
     
     """
     #vincent for new LLM
-    llm_model = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
+    #llm_model = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
     llm_model = "meta-llama/Llama-2-7b-chat-hf"
     llm = HuggingFaceEndpoint(
             repo_id=llm_model, 
@@ -276,7 +278,7 @@ def build_qa_chain(collection_name, vector_db, file: str):
         #return_generated_question=False,
         verbose=False,
     )
-    """
+    
     
     return collection_name, vector_db, chain