Jingye Li
Jingye Li
理解是正确的，另外把英文单词当做一个字也是没问题的，只要分词正确就行
There existed such a bug, but I thought it was fixed. I haven't encountered it since I updated the code. Did you try the bert version?
len(datasets[0])显示的是训练集的数据量，随后train_loader会将dataset处理成batch，每次训练通过循环来取一个batch训练，取完所有batch后记为一个epoch，因此len(train_loader)显示的是batch的数量，所以len(train_loader)会比len(datasets[0])小。train_loader加载了训练集并将处理成batch，本身就是用于训练的。
Hi, our code works when the _ner_ value is an empty list in the test set, although the f1 will be 0. I have tested it on the resume dataset,...
首先看看模型是否已经拟合，也可以尝试删除scheduler来增大模型的收敛速度；其次小batch size可能导致模型训练不稳定，可以尝试采用更小的len来截断，增大batch size；还可以分析一下模型测试集的输出，unified模型可能预测出了除flat以外其他类型的实体，导致了精准率偏低。
ADR means adverse drug reactions, you can find the detailed description in [Cadec: A corpus of adverse drug event annotations](https://www.sciencedirect.com/science/article/pii/S1532046415000532). The example is only for demostrating data structure, please replace...