Update app.py

app.py

@@ -88,8 +88,9 @@ def load_model_and_tokenizer(doc_text_count):
     # Load the pre-trained model with quantization
     model = AutoModelForCausalLM.from_pretrained(
         model_name,
-        trust_remote_code=True,
-        model_max_length=1.3*round(doc_text_count * 0.3 + 1)
+        torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
+        device_map="auto",
+        trust_remote_code=True
         ,token=HF_TOKEN
     )