如何提升下游任务上finetune的效果

您好，请问按照给定的config在下游任务上进行Finetune效果不太好可能是什么原因？是否需要引入lora等方式增加可学习参数的数量以提升模型在下游任务上的效果呢？
感谢您的解答！

No files, tests, or entry points are named. Start by locating the downstream fine-tuning configuration and training entry point, then reproduce the reported result with the given config. Done would require a documented diagnosis of the poor effect and an agreed scope for evaluating whether LoRA or another parameter-efficient method is appropriate.

Written by the indexing model from the issue text.