Monthly Daycon Speaker's Emotional Recognition AI Competition

Algorithm | NLP | Classification | Macro f1 score

 

Private 12위, LGBM 모델, emoberta-large 모델 사용

2022.12.12 14:16 3,341 Views language

개발환경은 코랩을 이용했습니다

파인튜닝을 하지않고 train과 test속 자연어부분의 감정을 예측
해당 데이터를 이용해 머신러닝 데이터셋을 만들었고 파라미터 튜닝후 lgbm모델로 예측했습니다

사용된모델:
tae898/emoberta-large
출처:
https://huggingface.co/tae898/emoberta-large

Code