2024/02/12:
Pro&DPO -V1增加DPO模型的合并。DPO论文。
搭配DPO (Direct Preference Optimization) 的Lora使用,可以提升提示词的准确性。
Used with Lora of DPO (Direct Preference Optimization), the accuracy of prompt words can be improved.
Pro&DPO-V1 adds the merging of DPO model. DPO papers.