PG
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Language Modeling | PG-19 | Perplexity5 | 206 | |
| Layout detection | 30-page PG (test) | Precision96.9 | 8 | |
| Long-Context Generation | PG-19 20K context length | Throughput Speedup (micro)4 | 6 | |
| Long-Context Generation | PG-19 10K context length | Throughput Speedup (micro-averaged)2.92 | 6 | |
| Text recognition | 30-page PG (test) | CER (%)1.05 | 4 | |
| Self-attention inverse temperature scaling analysis | PG19 | Tie Percentage8.8 | 2 | |
| Reading order | 30-page PG (test) | Precision98 | 1 | |
| Line detection | 30-page PG (test) | Precision98.3 | 1 |
Showing 8 of 8 rows