Table 3 Dimension of each instance’s using the four different character embeddings.
From: Improving deep neural network design with new text data representations
Experiment | Embedding | Dimension |
|---|---|---|
Embedding | log-70 | 8 × 182 |
log-256 | 8 × 182 | |
1-of-70 | 70 × 182 | |
1-of-256 | 256 × 182 | |
Padding | log-256 | 8 × 374 |