Update app.py

app.py

@@ -68,14 +68,14 @@ def predict(message, history, system_prompt, temperature, max_new_tokens, top_k,
 # Load model
 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
 quantization_config = BitsAndBytesConfig(
-    load_in_4bit=True,
-    bnb_4bit_compute_dtype=torch.bfloat16
+    load_in_8bit=True,
+    # bnb_4bit_compute_dtype=torch.bfloat16
 )
 tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
 model = AutoModelForCausalLM.from_pretrained(
     MODEL_ID,
     device_map="auto",
-    # quantization_config=quantization_config,
+    quantization_config=quantization_config,
     attn_implementation="flash_attention_2",
 )