implementation of https://arxiv.org/abs/2003.11644 Using PyTorch for arxiv multi label classification.
| Embedding | Dataset | Seq Length | Train Test Split | Micro-F1 Score | epochs |
|---|---|---|---|---|---|
| GloVe | Arxiv | 128 | 34688 / 17086 | 0.73 | 50 |
for above 100 epochs you get results of 0.75.
you can download the dataset from link below: Download