metadata
license: apache-2.0
base_model: unsloth/Qwen3.5-4B-Base
tags:
- tajik
- qwen3.5
- continual-pretraining
- cpt
- merged
language:
- tg
Qwen3.5-4B Tajik — CPT epoch-1, merged
Continual-pretraining (CPT) adaptation of unsloth/Qwen3.5-4B-Base to Tajik (тоҷикӣ), epoch 1, with the LoRA
adapter merged into the base weights (BF16). Equivalent to loading Tohirju/qwen35-4b-tajik-cpt-e1 on the base.
- What it is: the base Qwen3.5-4B after 1 epoch of Tajik continual pre-training (~314M Tajik tokens).
- Use: a stronger Tajik base for generation and as the starting point for the next CPT epoch or for SFT.
- Tajik MCQ (Zehnlab, full ~18,970 Q, loglikelihood): ~48.6% avg (base 44.7%). See the eval repo.
- Note: this is a base/CPT model — it completes text, it does not follow chat instructions (that is the SFT stage).
- Merge verified by logit parity against the adapter (max|Δlogit| < 1e0). Sibling adapters:
Tohirju/qwen35-4b-tajik-cpt-e1,-e2,-e3.
Gated: download requires manual approval. Released by Saidzoda Lab / Saidzoda Engineering Company.