| Benchmark | Model | Rank | Results |
|---|---|---|---|
| language-modelling-on-enwiki8 | Feedback Transformer | #5 | Bit per Character (BPC): 0.96Number of params: 77M |
| language-modelling-on-penn-treebank-character | Feedback Transformer | #5 | Bit per Character (BPC): 1.160Number of params: 10.7M |
| language-modelling-on-wikitext-103 | Feedback Transformer (8 layers) | #32 | Test perplexity: 18.2Validation perplexity: 17.5… |
| language-modelling-on-wikitext-103 | Feedback Transformer (4 layers) | #46 | Test perplexity: 22.4Validation perplexity: 21.4Number of params: 44M |