delete parameter no_split_module_classes

app.py

@@ -78,7 +78,7 @@ def load_model():
         torch_dtype=dtype,
         trust_remote_code=True,
         # quantization_config=bnb_config,
-        no_split_module_classes=["Block"],
+        # no_split_module_classes=["Block"],
         device_map="auto"
     ).to(device).eval()
     # tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)