计划加入DPO训练来缓解模型幻觉问题吗
jiezhangGt opened this issue · comments
Jay Zhang commented
如题,请问有计划引入DPO强化学习,来进一步提升模型效果,降低模型幻觉等问题吗?谢谢
Official repo for "Mini-Gemini: Mining the Potential of Multi-modality Vision Language Models"
jiezhangGt opened this issue · comments
如题,请问有计划引入DPO强化学习,来进一步提升模型效果,降低模型幻觉等问题吗?谢谢