nlu-figurative-detection

File Outline:

python and notebook execs:

• FL_pooler: pooler model with 3 bertweet output and a linear layer
• FL_pooler_with_f1: added f1 eval metric to FL_pooler
• FL_train: original model for multi head VAD regression task
• FLtest: notebook

slurm job files:

• clf_train: train the FL_pooler
• single_reg_train: train the single VAD regression task
• train_regressor: train newclf with vua dataset
• train_regresssor_bl: train newclf_bl with vua dataset

csv files:

• sarcasm_test/train: twiiter sarcasm file with prev and next sentence tags removed
• train/test_with_vua: sarcasm dataset combined with vua metaphor data set for the model

date: 2023-05-01