Update model_setup.py

model_setup.py

@@ -24,14 +24,13 @@ def load_model():
     # model = PeftModel.from_pretrained(base_model, "anhvv200053/Vinallama-2-7B-updated1-instruction-v2", token = token)
     
     model = AutoModelForCausalLM.from_pretrained(
-        "anhvv200053/Vinallama-2-7B-updated1-instruction-v2",
-        # quantization_config=bnb_config,
+        "anhvv200053/Vinallam_vssai_model",
         device_map={"": "cpu"},  # Đặt sử dụng CPU
         token = token
     )
     # model.config.pretraining_tp = 1
 
-    tokenizer = AutoTokenizer.from_pretrained('anhvv200053/Vinallama-2-7B-updated1-instruction-v2', trust_remote_code=True, use_fast=True, token = token)
+    tokenizer = AutoTokenizer.from_pretrained('anhvv200053/Vinallam_vssai_model', trust_remote_code=True, use_fast=True, token = token)
     tokenizer.pad_token = tokenizer.eos_token
 
     return model, tokenizer