Update app.py

app.py

@@ -17,30 +17,30 @@ DB_FAISS_PATH = 'vectorstore/db_faiss'
 
 # Load the model of choice
 def load_llm():
-    url = "https://huggingface.co/TheBloke/Llama-2-7B-Chat-GGML/blob/main/llama-2-7b-chat.ggmlv3.q4_K_M.bin"  # 2.87G
+    # url = "https://huggingface.co/TheBloke/Llama-2-7B-Chat-GGML/blob/main/llama-2-7b-chat.ggmlv3.q4_K_M.bin"  # 2.87G
 
-    model_loc, file_size = dl_hf_model(url)
+    # model_loc, file_size = dl_hf_model(url)
     
-    llm = CTransformers(
-        model=model_loc,
-        temperature=0.2,
-        model_type="llama",
-        top_k=10,
-        top_p=0.9,
-        repetition_penalty=1.0,
-        max_new_tokens=512,  # adjust as needed
-        seed=42,
-        reset=True,  # reset history (cache)
-        stream=False,
-        # threads=cpu_count,
-        # stop=prompt_prefix[1:2],
+    # llm = CTransformers(
+    #     model=model_loc,
+    #     temperature=0.2,
+    #     model_type="llama",
+    #     top_k=10,
+    #     top_p=0.9,
+    #     repetition_penalty=1.0,
+    #     max_new_tokens=512,  # adjust as needed
+    #     seed=42,
+    #     reset=True,  # reset history (cache)
+    #     stream=False,
+    #     # threads=cpu_count,
+    #     # stop=prompt_prefix[1:2],
                
         
-    )
-    # llm = G4FLLM(
-    #     model=models.gpt_35_turbo,
-    #     provider=Provider.Aichatos,
     # )
+    llm = G4FLLM(
+        model=models.gpt_35_turbo,
+        provider=Provider.Aichatos,
+    )
     # OPENAI_API_KEY = 'sk-nAqoChT9cmkAxALwMLdWT3BIbkFJcNHsH5Z5LN2ixPcDAopT'
     # llm =ChatOpenAI(openai_api_key=OPENAI_API_KEY,temperature=0)
     return llm