Update app.py

app.py

@@ -23,14 +23,16 @@ import torch
 
 
 def get_vector_store_from_url(url):
-    model_name = "BAAI/bge-large-en"
-    model_kwargs = {'device': 'cpu'}
-    encode_kwargs = {'normalize_embeddings': False}
-    embeddings = HuggingFaceBgeEmbeddings(
-        model_name=model_name,
-        model_kwargs=model_kwargs,
-        encode_kwargs=encode_kwargs
-    )
+    # model_name = "BAAI/bge-large-en"
+    # model_kwargs = {'device': 'cpu'}
+    # encode_kwargs = {'normalize_embeddings': False}
+    # embeddings = HuggingFaceBgeEmbeddings(
+    #     model_name=model_name,
+    #     model_kwargs=model_kwargs,
+    #     encode_kwargs=encode_kwargs
+    # )
+    embeddings = HuggingFaceEmbeddings(model_name='thenlper/gte-large',
+                                   model_kwargs={'device': 'cpu'})
     
     loader = WebBaseLoader(url)
     document = loader.load()
@@ -114,17 +116,23 @@ def get_response(user_input):
     #                 lib="avx2", # for CPU
     # )
 
-    model_name = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
-    # llm = HuggingFaceHub(
-    #     repo_id=llm_model, 
-    #     model_kwargs={"temperature": 0.3, "max_new_tokens": 250, "top_k": 3}
+    # model_name = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
+    # # llm = HuggingFaceHub(
+    # #     repo_id=llm_model, 
+    # #     model_kwargs={"temperature": 0.3, "max_new_tokens": 250, "top_k": 3}
+    # # )
+
+    # llm = transformers.AutoModelForCausalLM.from_pretrained(
+    #             model_name,
+    #             trust_remote_code=True,
+    #             torch_dtype=torch.bfloat16,
+    #             device_map='auto'
     # )
 
-    llm = transformers.AutoModelForCausalLM.from_pretrained(
-                model_name,
-                trust_remote_code=True,
-                torch_dtype=torch.bfloat16,
-                device_map='auto'
+    llm = HuggingFacePipeline.from_model_id(
+        model_id="google/flan-t5-base",
+        task="text2text-generation",
+        # model_kwargs={"temperature": 0.2},
     )
     retriever_chain = get_context_retriever_chain(st.session_state.vector_store,llm)
     conversation_rag_chain = get_conversational_rag_chain(retriever_chain,llm)