Language Modeling on PG-19 subword-level
3.94Forward BPTTransformer-23M
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Transformer-23MT=2562026.05 | 3.94 | 3.76 | 3.94 | |
| Transformer-18MT=2562026.05 | 3.95 | 3.77 | 3.94 | |
| SHARPT=42026.05 | 4.26 | 4.26 | 4.26 | |
| GRUT=42026.05 | 4.5 | 4.52 | 4.51 | |
| LSTMT=42026.05 | 4.52 | 4.54 | 4.53 | |
| RNNT=42026.05 | 4.54 | 4.56 | 4.55 |