BERT codemixed base model for hinglish (cased)
Model description
Input for the model: Any codemixed hinglish text
Output for the model: Sentiment. (0 – Negative, 1 – Neutral, 2 – Positive)
I took a bert-base-multilingual-cased model from Huggingface and finetuned it on SAIL 2017 dataset.
Performance of this model on the SAIL 2017 dataset
| metric | score |
|---|---|
| acc | 0.588889 |
| f1 | 0.582678 |
| acc_and_f1 | 0.585783 |
| precision | 0.586516 |
| recall | 0.588889 |
前往AI网址导航
正文完
