通过阅读网上的资料代码,进行自我加工,努力实现常用的机器学习算法。实现算法有KNN、Kmeans、EM、Perceptron、决策树、逻辑回归、svm、adaboost、朴素贝叶斯
Source code for paper "Contrastive Out-of-Distribution Detection for Pretrained Transformers", EMNLP 2021
Code for EMNLP 2021 paper "CLIFF: Contrastive Learning for Improving Faithfulness and Factuality in Abstractive Summarization"
PyTorch code for "Prototypical Contrastive Learning of Unsupervised Representations"
Pytorch implementation of CoCon: A Self-Supervised Approach for Controlled Text Generation