| Benchmark | Model | Rank | Results |
|---|---|---|---|
| image-generation-on-imagenet-64x64 | Routing Transformer | #33 | Bits per dim: 3.43 |
| language-modelling-on-enwiki8 | Routing Transformer (12 layers) | #13 | Bit per Character (BPC): 0.99 |
| language-modelling-on-wikitext-103 | Routing Transformer | #11 | Test perplexity: 15.8 |