The initial release of this project focuses on the Bradley-Terry reward modeling and pairwise preference model. Since then, we have included more advanced techniques to construct a preference model.

Liger Kernel is a collection of Triton kernels designed specifically for LLM training. It can effectively increase multi-GPU training throughput by 20% and reduces memory usage by 60%. We have ...

该导热膏基于绝缘不导电纳米分子导热材料，导热系数达 14.4 W/m·K，同时热阻仅 0.025 K·cm 2 /W，拥有优秀热传导能力。 这款导热膏密度为 2.8g/cm 3 ，采用无毒无味无腐蚀性材料，支持 -50℃~+250℃ 的广阔工作温度范围。 银昕 TF02 导热膏质地湿润、黏稠度适中 ...