gzt4se

> > 请问 ACC为0 解决了吗？ @yuanyuan1993 > > epoch 太少了 我这边迭代5个epoch以上就有准确率了 您好，可以具体说明一下您是怎么解决的么    

> Yes, you can definitely finetune on labeled datasets using contrastive loss (or combined with the matching loss) to further boost the retrieval performance. We plan to release the finetuning...