OWT
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Language Modeling | OWT (test) | Perplexity (PPL)17.5 | 79 | |
| Language Modeling | OWT | Gen. PPL15.06 | 78 | |
| Text Generation | OWT | GPT2 Perplexity5.33 | 41 | |
| Conditional Generation | OWT | Perplexity (PPL)19.99 | 24 | |
| Language Modeling | OWT (val) | PPL17.5 | 23 | |
| Language Modeling | OWT L=1024 (test) | NELBO PPL20.96 | 11 | |
| Generative Perplexity | OWT L=1024 | Generative Perplexity14.1 | 6 | |
| Language Modeling | OWT | PPL11.2 | 6 | |
| Generative Perplexity | OWT L=2048 | Generative Perplexity13.2 | 5 | |
| Generation Latency | OWT L=8192 | Generation Latency (s)54 | 5 | |
| Generation Latency | OWT L=2048 | Sampling Latency (s)13.3 | 5 | |
| Language Modeling | OWT Qwen-2 tokenizer (val) | Validation Perplexity21.9 | 4 | |
| Language Generation | OWT (test) | MAUVE92.94 | 3 | |
| Language Modeling | OWT L=10240 (test) | Gen. PPL23.4 | 2 |
Showing 14 of 14 rows