发布概述
We release Saidzoda-Tajik-LFM-2.6B, a Tajik adaptation of the 2.6B-parameter hybrid model LFM2.5-2.6B-Base, built for tool calling in Tajik on modest hardware.
训练配方
The recipe has two stages: LoRA continued pretraining (CPT) on 1,485,319 cleaned Tajik web documents for 3 epochs, then one epoch of supervised fine-tuning (SFT) on a mix of machine-translated Tajik tool calls, English tool calls kept as replay, and Tajik instructions.
工具调用评估
On a 174-item tool-calling test set built from BFCL v3 and machine-translated to Tajik, strict full-call accuracy on Tajik prompts rises from 20.69 for the base model to 64.14 (paired difference +43.45 points, 95% interval +35.2 to +51.7); a lenient parser that also accepts the base model's own call formats puts the base at 31.72.
English shows no detectable change (76.55 and 77.24).
Every CPT checkpoint scores zero: CPT removes the native tool-call format and, as the lenient re-scoring shows, part of the skill itself.
非工具调用评估
- Tajik knowledge benchmark accuracy rises from 28.0 to 35.0.
- Belebele Tajik shows no significant change.
- English MMLU falls from 60.8 to 52.5.
模型局限性
The model is not a general Tajik assistant. It almost never declines when no tool fits (6.9% correct on Tajik, against 31.0% for the base under the lenient parser), it is weak at multi-turn calls, and its calls should be checked before they are executed.
可复现性
We describe the data, the training settings and every evaluation in enough detail to repeat them.