FangShancheng / ABINet

Read Like Humans: Autonomous, Bidirectional and Iterative Language Modeling for Scene Text Recognition
Other
437 stars 73 forks source link

语言模型训练 #109

Open wu-yz opened 7 months ago

wu-yz commented 7 months ago

我训练语言模型的时候loss很快就接近0了,但是评估的时候准确率还是不高,loss也挺大的,这是什么原因呢?有没有好心人解答一下呀?