DeepSpeed
Original author(s) | Microsoft Research |
---|---|
Developer(s) | Microsoft |
Initial release | May 18, 2020 |
Stable release | v0.14.4
/ June 21, 2024 |
Repository | github |
Written in | Python, CUDA, C++ |
Type | Software library |
License | Apache License 2.0 |
Website | deepspeed |
DeepSpeed is an open source deep learning optimization library for PyTorch.[1]
Library
[edit]The library is designed to reduce computing power and memory use and to train large distributed models with better parallelism on existing computer hardware.[2][3] DeepSpeed is optimized for low latency, high throughput training. It includes the Zero Redundancy Optimizer (ZeRO) for training models with 1 trillion or more parameters.[4] Features include mixed precision training, single-GPU, multi-GPU, and multi-node training as well as custom model parallelism. The DeepSpeed source code is licensed under MIT License and available on GitHub.[5]
The team claimed to achieve up to a 6.2x throughput improvement, 2.8x faster convergence, and 4.6x less communication.[6]
See also
[edit]References
[edit]- ^ "Microsoft Updates Windows, Azure Tools with an Eye on The Future". PCMag UK. May 22, 2020.
- ^ Yegulalp, Serdar (February 10, 2020). "Microsoft speeds up PyTorch with DeepSpeed". InfoWorld.
- ^ "Microsoft unveils "fifth most powerful" supercomputer in the world". Neowin. 18 June 2023.
- ^ "Microsoft trains world's largest Transformer language model". February 10, 2020.
- ^ "microsoft/DeepSpeed". July 10, 2020 – via GitHub.
- ^ "DeepSpeed: Accelerating large-scale model inference and training via system optimizations and compression". Microsoft Research. 2021-05-24. Retrieved 2021-06-19.
Further reading
[edit]- Rajbhandari, Samyam; Rasley, Jeff; Ruwase, Olatunji; He, Yuxiong (2019). "ZeRO: Memory Optimization Towards Training A Trillion Parameter Models". arXiv:1910.02054 [cs.LG].