Baseline Predictions for Stanford Sentiment Treebank (SST-5)

We obtained model predictions for the Stanford Sentiment Treebank benchmark (SST-5) from eight open-source baselines: dictionary-based methods VADER and TextBlob, traditional machine learning methods like logistic regression and support vector machine (SVM), fastText classifier, and deep learning classifiers: BERT and ELMo with Flair and fine-tuned BERT with Hugging Face.我们从八个开源基准模型中获取了斯坦福情感树库基准测试（Stanford Sentiment Treebank benchmark，SST-5）的模型预测结果，包括基于词典的方法VADER与TextBlob、传统机器学习方法（如逻辑回归与支持向量机（support vector machine，SVM））、fastText分类器，以及深度学习分类器：BERT、搭载Flair的ELMo，以及基于Hugging Face的微调BERT。