Update app.py

app.py

@@ -12,12 +12,12 @@ from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
 import os
 from dotenv import load_dotenv
 from langchain_huggingface import ChatHuggingFace, HuggingFacePipeline
-
+from langchain_community.llms import HuggingFaceEndpoint
 
 
 warnings.filterwarnings("ignore")
 load_dotenv()
-os.environ["HUGGINGFACEHUB_API_TOKEN"] = os.getenv("HUGGINGFACEHUB_API_TOKEN")
+api_token = os.getenv("HUGGINGFACEHUB_API_TOKEN")
 
 # Constants and configurations
 APP_TITLE = "💊 Asisten Kesehatan Feminacare"
@@ -68,15 +68,22 @@ def create_llm():
     
     # llm = HuggingFacePipeline(pipeline=text_generation_pipeline)
 
-    return HuggingFaceHub(
-        repo_id=MODEL_NAME,
-        model_kwargs={
-            "temperature": 0.7,  # Balanced between creativity and accuracy
-            "max_new_tokens": 1024,
-            "top_p": 0.9,
-            "frequency_penalty": 0.5
-        }
-    )
+    # return HuggingFaceHub(
+    #     repo_id=MODEL_NAME,
+    #     model_kwargs={
+    #         "temperature": 0.7,  # Balanced between creativity and accuracy
+    #         "max_new_tokens": 1024,
+    #         "top_p": 0.9,
+    #         "frequency_penalty": 0.5
+    #     }
+    # )
+    llm = HuggingFaceEndpoint(
+            repo_id=MODEL_NAME,
+            huggingfacehub_api_token = api_token,
+            temperature = 0.7,
+            max_new_tokens = 1024,
+            top_k = 0.9,
+        )
     # llm = HuggingFacePipeline.from_model_id(
     #     model_id=MODEL_NAME,
     #     task="text-generation",
@@ -86,7 +93,7 @@ def create_llm():
     #         repetition_penalty=1.03,
     #     ),
     # )
-    # return llm
+    return llm
     
     
     # chat_model = ChatHuggingFace(llm=llm)