arthu1/wind-lite-sft
收藏资源简介:
--- license: mit task_categories: - text-generation - question-answering language: - en tags: - wind - north-ai - sft - instruction-tuning pretty_name: Wind Lite SFT size_categories: - 10K<n<100K --- # Wind Lite SFT Custom supervised fine-tuning dataset for **Wind Lite 1.6** by North AI. ## Dataset Summary 20,000 high-quality instruction-response pairs covering identity grounding, math reasoning, coding, general knowledge, and multi-turn conversations. ## Data Composition | Category | Count | Description | |---|---|---| | Math & Reasoning | ~7,000 | Arithmetic, algebra, percentages, unit conversions — with step-by-step working | | Coding | ~4,000 | Python, JavaScript, SQL, systems — with explanations | | Multi-turn Conversations | ~60 | Debugging sessions, concept explanations, building projects | | General Knowledge | ~200 | Science, history, economics, geography | | Wind Identity | ~25 | Model name, creator, capabilities, limitations | | Personality & Style | ~50 | Tone, advice-giving, how to respond | ## Format All examples use the ChatML messages format: ```json { "messages": [ {"role": "user", "content": "What is 15% of 340?"}, {"role": "assistant", "content": "15% × 340 = 0.15 × 340 = **51**"} ] } ``` Multi-turn examples contain more than 2 messages. ## Usage ```python from datasets import load_dataset ds = load_dataset("arthu1/wind-lite-sft", split="train") ``` ## Model This dataset is used to train [Wind Lite 1.6](https://huggingface.co/arthu1/wind-lite-1-6-0326-beta1) by North AI. ## License MIT




