Fine-Tuning Pre-trained Language Model with Weak Supervision: A Contrastive-Regularized Self-Training Approach
Open paper
Benchmark
Model
Rank
Results
sentiment-analysis-on-imdb
COSINE
#34
Accuracy: 90.54
text-classification-on-yelp-2
COSINE
#5
Accuracy: 95.97%
word-sense-disambiguation-on-words-in-context
COSINE + Transductive Learning
#1
Accuracy: 85.3
Rank counts only results with a code link.