Update app.py

app.py

@@ -18,7 +18,7 @@ APP_TITLE = "💊 Asisten Kesehatan Feminacare"
 INITIAL_MESSAGE = """Halo! 👋 Saya adalah asisten kesehatan feminacare yang siap membantu Anda dengan informasi seputar kesehatan wanita. 
 Silakan ajukan pertanyaan apa saja dan saya akan membantu Anda dengan informasi yang akurat."""
 
-MODEL_NAME = "SeaLLMs/SeaLLMs-v3-7B-Chat"
+MODEL_NAME = "SeaLLMs/SeaLLM-13B-Chat"
 EMBEDDING_MODEL = "sentence-transformers/all-MiniLM-L6-v2"
 TOP_K_DOCS = 5
 
@@ -34,34 +34,43 @@ def initialize_models():
   
 def create_llm():
     """Initialize the language model with auto device mapping"""
-    model = AutoModelForCausalLM.from_pretrained(
-        MODEL_NAME,
-        device_map="auto",
-        trust_remote_code=True
-    )
-    tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+    # model = AutoModelForCausalLM.from_pretrained(
+    #     MODEL_NAME,
+    #     device_map="auto",
+    #     trust_remote_code=True
+    # )
+    # tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
     
-    # Get terminators for the model
-    terminators = [tokenizer.eos_token_id]
-    if hasattr(tokenizer, 'convert_tokens_to_ids'):
-        try:
-            terminators.append(tokenizer.convert_tokens_to_ids("<|eot_id|>"))
-        except:
-            pass
+    # # Get terminators for the model
+    # terminators = [tokenizer.eos_token_id]
+    # if hasattr(tokenizer, 'convert_tokens_to_ids'):
+    #     try:
+    #         terminators.append(tokenizer.convert_tokens_to_ids("<|eot_id|>"))
+    #     except:
+    #         pass
     
-    text_generation_pipeline = pipeline(
-        model=model,
-        tokenizer=tokenizer,
-        task="text-generation",
-        temperature=0.2,
-        do_sample=True,
-        repetition_penalty=1.1,
-        return_full_text=False,
-        max_new_tokens=200,
-        eos_token_id=terminators,
-    )
+    # text_generation_pipeline = pipeline(
+    #     model=model,
+    #     tokenizer=tokenizer,
+    #     task="text-generation",
+    #     temperature=0.2,
+    #     do_sample=True,
+    #     repetition_penalty=1.1,
+    #     return_full_text=False,
+    #     max_new_tokens=200,
+    #     eos_token_id=terminators,
+    # )
     
-    return HuggingFacePipeline(pipeline=text_generation_pipeline)
+    # return 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
+        }
+    )
 
 PROMPT_TEMPLATE = """
 Anda adalah asisten kesehatan profesional dengan nama Feminacare.