LTP For Rust

传统机器学习方法(LTP 3)实现的 CWS / POS / NER 算法。

| method | ltp 3.0(c++) | ap(1) | ap(8) | pa | pa-i(0.5) | pa-ii(0.5) | | ------ | ------------ | ----- | ----- | ----- | --------- | ---------- | | cws | 97.83 | 97.93 | 97.67 | 97.90 | 97.90 | 97.93 | | pos | 98.35 | 98.41 | 98.30 | 98.39 | 98.39 | 98.38 | | ner | 94.17 | 94.28 | 93.42 | 94.02 | 94.06 | 93.95 |

features

Benchmark Compare with Jieba

| Algorithm | Time(s) | Speed(KB/s) | | -------------- | ------: | ----------: | | Jieba cut | 35.29 | 982.49 | | LTP legacy(1) | 58.64 | 591.04 | | LTP legacy(2) | 30.64 | 1131.30 | | LTP legacy(4) | 17.14 | 2022.50 | | LTP legacy(8) | 11.39 | 3044.15 | | LTP legacy(16) | 9.88 | 3508.93 |

Benchmark Pipeline

| Algorithm | Time(s) | Speed(KB/s) | | -------------- | ------: | ----------: | | LTP legacy(1) | 139.70 | 248.11 | | LTP legacy(2) | 75.18 | 461.04 | | LTP legacy(4) | 42.96 | 806.76 | | LTP legacy(8) | 29.34 | 1181.39 | | LTP legacy(16) | 26.28 | 1318.84 |