Atlas · GenAI 2026
DeepSpeed
Distributed-training library (ZeRO sharding, offload) for training and fine-tuning large models across GPUs.
toolPeak: 2022Training InfrastructureAI consensus: 0/3
Prerequisites
It is a distributed-training engine.
Recommended reference
ZeRO: Memory Optimizations Toward Training Trillion Parameter Models — arXiv 1910.02054