1. MLP
2. CNN
Tiny ImageNet
3. RNN
Word Embedding + RNN+ FC
| Model | Pool | Text Len | Bidirectional | Accuracy |
|---|---|---|---|---|
| lstm | last | 512 | 1 | 0.661 |
| lstm | max | 512 | 1 | 0.660 |
| lstm | attn | 512 | 1 | 0.659 |
| lstm | mean | 512 | 1 | 0.659 |
| lstm | mean | 256 | 1 | 0.653 |
| lstm | last | 256 | 1 | 0.652 |
| lstm | attn | 256 | 1 | 0.652 |
| lstm | max | 256 | 1 | 0.649 |
| gru | last | 256 | 1 | 0.645 |
| lstm | last | 256 | 0 | 0.637 |
| gru | last | 256 | 0 | 0.625 |
| transformer | cls | 512 | 0 | 0.610 |
| transformer | cls | 256 | 0 | 0.596 |
| rnn | last | 256 | 0 | 0.569 |
| rnn | last | 256 | 1 | 0.544 |
4. GNN
