continue pre-training Example

[keyuchen21]

System Info

Will that be possible to provide a continue pre-training Example for llama?
There are a lot of example/paper on BERT, but I'm wondering if there is a good way to continue pre-training llama?

Example from BERT

Information

• The official example scripts
• My own modified scripts

🐛 Describe the bug

NA

Error logs

NA

Expected behavior

NA