List your company

Compression on 300 MB Monolingual Corpora

58.7Train TokensBaseline
49.96452.23254.556.768Dec 14, 2025

Evaluation results

MethodLinks
Baselinen=16K2025.1258.765
BPEn=16K2025.1258.564
FSPn=16K2025.1258.264.1
Baselinen=32K2025.125459.6
BPEn=32K2025.1253.858.8
FSPn=32K2025.1253.758.9
Baselinen=64K2025.1250.655.7
BPEn=64K2025.1250.454.9
FSPn=64K2025.1250.355.1