About Merge Adapter Weights of LLMs in StackLLM blog

[manred1997]

Thanks for the stackllm blog, I read the blog and saw a passage like this: "Since we will tune the model again with different objectives, we merge the adapter weights with the original model weights" in your blog. Can you explain in more detail how to merge the weight adapter model? Why are you using this method? and Is it the same as a bagging ensemble technique?