Llama-Nemotron-Post-Training-Dataset nvidia
Llama-Nemotron-Post-Training-Dataset-v1.1 Release Update [4/8/2025]: v1.1: We are releasing an additional 2.2M Math and 500K Code Reasoning Data in support of our release of Llama-3.1-Nemotron-Ultra-253B-v1. 🎉 Data Overview This dataset is a compilation of SFT and RL data that supports improvements of math, code, general reasoning, and instruction following capabilities of the original Llama instruct model, in support of NVIDIA’s release of… See the full description on the dataset page: https://huggingface.co/datasets/nvidia/Llama-Nemotron-Post-Training-Dataset.
- 类型
- dataset
- 许可
- cc-by-4.0
- 下载量
- 6,676
- 点赞
- 697
- 访问
- public
- 文件
- 0
标签
- 指令微调
- 数据
- 数学推理
- 代码数据
- 推理
- 多任务
- 合成数据
- 对话数据
摘要
该数据集是NVIDIA为提升Llama系列指令模型能力而发布的完整后训练数据集，涵盖SFT和RL两大配置，包含约3300万条数据，主要覆盖数学、代码、科学推理、指令遵循和对话等任务。数据由DeepSeek-R1、Qwen、Llama等多个开源模型合成生成，支持128K上下文，可用于提升模型的数学推理、代码生成及通用推理能力。适用于社区继续翻新和评估开源模型，尤其适合训练AI Agent、聊天机器人和RAG系统。
README
--- license: cc-by-4.0 configs: - config_name: SFT data_files: - split: code path: SFT/code/*.jsonl - split: math path: SFT/math/*.jsonl - split: science path: SFT/science/*.jsonl - split: chat path: SFT/chat/*.jsonl - split: safety path: SFT/safety/*.jsonl default: true - config_name: RL data_file…