---

license: cc-by-sa-4.0

base_model: johnbean393/chiboard-1-s1

tags:

• gguf
• chinese
• lfm2
• chiboard
• pinyin
• ime
• sft
• student

language:

• zh

---

chiboard-1-s1-GGUF

GGUF export of johnbean393/chiboard-1-s1, a 350M SFT student initialization for Chiboard Chinese pinyin-to-Hanzi conversion and revision.

Prompt format:

<|startoftext|>{committed_context}<|reserved_6|>{raw_pinyin}<|reserved_7|>{display}<|reserved_8|>{target}<|im_end|>

The source model is trained with completion-only loss on {target}<|im_end|>.

Serve with exactly one <|startoftext|> token. Most runtimes add it automatically, so do not also embed it in the prompt string.

This is S1, the student initialization for the later Chiboard teacher-distillation stage; it is not the shipped final model.

Files

• Chiboard-S1-Q8_0.gguf: Q8_0 GGUF quantization of johnbean393/chiboard-1-s1.

Training

• Base model: LiquidAI/LFM2.5-350M-Base
• Datasets: johnbean393/chiboard-1-sft, johnbean393/chiboard-1-revision-sft
• Source model: johnbean393/chiboard-1-s1
• Hub target: johnbean393/chiboard-1-s1-GGUF
• Mixture: all plain SFT rows once plus all revision SFT rows twice
• Mixture order: concatenated and globally shuffled before one training pass
• Optimizer steps: 18923
• Peak learning rate: 3e-5
• Cosine learning-rate floor: 3e-6

Conversion

• Source revision: ac4e8c3885ef1a0ad5b3f90a1e4c5639ae669edd
• GGUF architecture: lfm2
• GGUF version: 3
• Quantization: Q8_0
• Context length: 128000
• SHA-256: 8de72b7f81c9b891947aac2965ce3e971c732400291a7090e331e77ccc0f369d