1 result found Sort:

yueyu1030
COSINE yueyu1030
25
200
mit
10
[NAACL 2021] This is the code for our paper `Fine-Tuning Pre-trained Language Model with Weak Supervision: A Contrastive-Regularized Self-Training Approach'.
agnews dataset roberta fine-tuning slot-filling roberta-model self-training language-model pseudo-labeling weak-supervision text-classification contrastive-learning learning-with-noisy-labels weakly-supervised-learning sentence-pair-classification
Created 2020-10-15
28 commits to main branch, last one 2 years ago