update examples

examples/merge_lora/merge.sh

@@ -1,4 +1,5 @@
 #!/bin/bash
+# DO NOT use quantized model or quantization_bit when merging lora weights
 
 CUDA_VISIBLE_DEVICES= python ../../src/export_model.py \
     --model_name_or_path meta-llama/Llama-2-7b-hf \