Add Qwen/Qwen3-0.6B to pythainlp.lm

[wannaphong]

I found Qwen/Qwen3-0.6B. It is good enough for small tasks. I think we can add Qwen/Qwen3-0.6B to pythainlp.lm.

Model: https://huggingface.co/Qwen/Qwen3-0.6B

Apache license 2.0

[tassa-yoniso-manasi-karoto]

I did some quick benchmarks out of curiosity. Claude Opus 4 provided the reference tokenization.

Rank   Tokenizer                 Avg F1 Score
---------------------------------------------
1      gemmi25_pro (LLM)         0.867
2      gemmi25_flashlight (LLM)  0.856
3      qwen3_4b (LLM)            0.803
4      newmm (engine)            0.802
5      gpt4.1 (LLM)              0.790
6      longest (engine)          0.784
7      attacut (engine)          0.776
8      deepcut (engine)          0.775
9      icu (engine)              0.746
10     gemma3n_e2b (LLM)         0.167

tassa-yoniso-manasi-karoto@ef2ee83