开源官网:https://llm.hunyuan.tencent.com/
github(开源模型工具包):https://github.com/Tencent/Hunyuan-Large
huggingface(模型下载):https://huggingface.co/tencent/Hunyuan-Large/tree/main
huggingface demo 地址:https://huggingface.co/spaces/tencent/Hunyuan-Large
技术报告:https://arxiv.org/abs/2411.02265
共享专家路由策略
回收路由策略
专家特定学习率适配策略
高质量的合成数据
SFT 训练
RLHF 训练