Takenoko12345678/Japanese-SFT-Qwen3.8-27B-10K
Japanese-SFT-Qwen3.8-27B-10K 日本語の質問 10,043件に、Qwen/Qwen3.8-27B（思考なし）が答えた、日本語の指示チューニング（SFT）用データセットです。 小さなモデルの日本語を自然にするための「蒸留」に使う目的で作りました。実際に Spark-X2.5-4B の日本語化に使っています（下の「このデータで学習した結果」を参照）。 A Japanese SFT dataset of 10,043 instructions answered by Qwen3.8-27B (non-thinking). Prompts come from llm-jp/magpie-sft-v1.0. Built to distill natural Japanese writing into small models. See "Known issues" before use: responses were not fact-checked. 概要 項目 内容 件数 10… See the full description on the dataset page: https://huggingface.co/datasets/Takenoko12345678/Japanese-SFT-Qwen3.8-27B-10K.
| Author / organization | Takenoko12345678 |
|---|---|
| License (as declared) | apache-2.0 |
| Downloads | 265 (as reported by Hugging Face) |
| Likes | 3 |
| Created | 25 September 2026 |
| Last updated at source | 27 September 2026 |
Popularity, last 3 days
265 now (+5 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
- Takenoko12345678/Qwen3.5-0.8B-Japanese-SFTModel · used to train
- mradermacher/Qwen3.5-0.8B-Japanese-SFT-v2-GGUFModel · used to train
- Takenoko12345678/Spark-X2.5-4B-JapaneseModel · used to train
- Takenoko12345678/Qwen3.5-0.8B-Japanese-SFT-v2Model · used to train
- mradermacher/Qwen3.5-0.8B-Japanese-SFT-v2-i1-GGUFModel · used to train
More from Takenoko12345678
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.