A Unified Architecture for Accelerating Distributed DNN Training in Heterogeneous GPU/CPU Clusters 论文阅读

MLSys 标签

2025

04-06

A Unified Architecture for Accelerating Distributed DNN Training in Heterogeneous GPU/CPU Clusters 论文阅读

04-05

Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism 论文阅读

04-04

Horovod: fast and easy distributed deep learning in TensorFlow 论文阅读

03-13

Scaling Distributed Machine Learning with the Parameter Server 阅读笔记