论文标题
McTensor：具有多组分浮点的高精度深度学习库
MCTensor: A High-Precision Deep Learning Library with Multi-Component Floating-Point
论文作者
论文摘要
在本文中，我们介绍了McTensor，这是一个基于Pytorch的库，用于为DL培训提供通用和高精度算术。 MCTENSOR的使用方式与Pytorch Tensor相同：我们为具有相同的Pytorch接口的MCTENSOR实现了多个基本，矩阵级计算运算符和NN模块。我们的算法获得了高精度计算，并且还受益于重优化的Pytorch浮点算术算术。我们针对一系列任务评估了针对Pytorch天然算术的mctensor算术，其中使用MCTENSOR在Float16中使用MCTENSOR的模型将匹配或优于float32或float64精度的Pytorch模型。
In this paper, we introduce MCTensor, a library based on PyTorch for providing general-purpose and high-precision arithmetic for DL training. MCTensor is used in the same way as PyTorch Tensor: we implement multiple basic, matrix-level computation operators and NN modules for MCTensor with identical PyTorch interface. Our algorithms achieve high precision computation and also benefits from heavily-optimized PyTorch floating-point arithmetic. We evaluate MCTensor arithmetic against PyTorch native arithmetic for a series of tasks, where models using MCTensor in float16 would match or outperform the PyTorch model with float32 or float64 precision.
