qwen3-0.6B-grammar

Qwen3-0.6B fine-tuned into a bilingual (RU/EN) text normalizer: fixes typos, grammar, punctuation, and casual chat style while preserving meaning, tone, and slang.

Usage

Prompt format: user: "Fix the text: <dirty text>" β†’ the model replies with the cleaned text.

License

CC BY-NC-SA 4.0 β€” non-commercial, share-alike (see LICENSE). For commercial licensing, contact the author.

Base model Qwen/Qwen3-0.6B is Apache 2.0 (see LICENSE-APACHE-2.0); this restriction applies to the fine-tuned weights, not to the upstream base model. Third-party attribution is in NOTICE.

Downloads last month
20
Safetensors
Model size
0.6B params
Tensor type
F32
Β·
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Model tree for benzlokzik/qwen3-0.6B-grammar

Finetuned
Qwen/Qwen3-0.6B
Finetuned
(1,492)
this model

Dataset used to train benzlokzik/qwen3-0.6B-grammar