【BERT蒸馏】DistilBERT、Distil-LSTM、TinyBERT、FastBERT(论文+代码)
文章目录0.引言1.FastBERT:aSelf-distillingBERTwithAdaptiveInferenceTime1.1摘要1.2动机1.3贡献(适用于文本分类任务)1.4相关工作1.5模型1.5.1模型结构1.5.2训练步骤1.6实验结果2.DistilBERT,adistilledversionofBERT:smaller,faster,cheaperandlighter2.1摘