DeepSpeed

DeepSpeed

Publications

News & features

News & features

Overview

DeepSpeed is a deep learning optimization library that makes distributed training easy, efficient, and effective.

10x Larger Models | 10x Faster Training | Minimal Code Change

DeepSpeed can train DL models with over a hundred billion parameters on current generation of GPU clusters, while achieving over 10x in system performance compared to the state-of-art. Early adopters of DeepSpeed have already produced a language model (LM) with over 17B parameters called Turing-NLG, establishing a new SOTA in the LM category.

DeepSpeed is an important part of Microsoft’s new AI at Scale initiative to enable next-generation AI capabilities at scale. Take a deep dive into large scale AI across Microsoft.

People

Why DeepSpeed

Why DeepSpeed?

Training advanced deep learning models is challenging. Beyond model design, model scientists also need to set up the state-of-the-art training techniques such as distributed training, mixed precision, gradient accumulation, and checkpointing. Yet still, scientists may not achieve the desired system performance and convergence rate. Large model sizes are even more challenging: a large model easily runs out of memory with pure data parallelism and it is difficult to use model parallelism. DeepSpeed addresses these challenges to accelerate model development and training.

Distributed, Effective, and Efficient Training with Ease

The DeepSpeed API is a lightweight wrapper on PyTorch. This means that you can use everything you love in PyTorch and without learning a new platform. In addition, DeepSpeed manages all of the boilerplate state-of-the-art training techniques, such as distributed training, mixed precision, gradient accumulation, and checkpoints so that you can focus on your model development. Most importantly, you can leverage the distinctive efficiency and effectiveness benefit of DeepSpeed to boost speed and scale with just a few lines of code changes to your PyTorch models.

Speed

DeepSpeed achieves high performance and fast convergence through a combination of efficiency optimizations on compute/communication/memory/IO and effectiveness optimizations on advanced hyperparameter tuning and optimizers. For example:

  • DeepSpeed trains BERT-large to parity in 44 mins using 1024 V100 GPUs (64 DGX-2 boxes) and in 2.4 hours using 256 GPUs (16 DGX-2 boxes).
  • DeepSpeed trains GPT2 (1.5 billion parameters) 3.75x faster than state-of-art, NVIDIA Megatron on Azure GPUs.

Read the GPT tutorial >

Memory efficiency

DeepSpeed provides memory-efficient data parallelism and enables training models without model parallelism. For example, DeepSpeed can train models with up to 13 billion parameters on NVIDIA V100 GPUs with 32GB of device memory. In comparison, existing frameworks (e.g., PyTorch’s Distributed Data Parallel) run out of memory with 1.4 billion parameter models.

DeepSpeed reduces the training memory footprint through a novel solution called Zero Redundancy Optimizer (ZeRO). Unlike basic data parallelism where memory states are replicated across data-parallel processes, ZeRO partitions model states and gradients to save significant memory. Furthermore, it also reduces activation memory and fragmented memory. The current implementation (ZeRO-2) reduces memory by up to 8x relative to the state-of-art. You can read more about ZeRO in our paper, and in our blog posts related to ZeRO-1.

With this impressive memory reduction, early adopters of DeepSpeed have already produced a language model (LM) with over 17B parameters called Turing-NLG, establishing a new SOTA in the LM category.

Scalability

DeepSpeed supports efficient data parallelism, model parallelism, and their combination. ZeRO boosts the scaling capability and efficiency further.

  • DeepSpeed provides system support to run models up to 170 billion parameters, 10x larger than the state-of-art (8 billion NVIDIA GPT, 11 billion Google T5).
  • DeepSpeed can run large models more efficiently, up to 10x faster for models with various sizes spanning 1.5B to 170B. More specifically, the data parallelism powered by ZeRO is complementary and can be combined with different types of model parallelism. It allows DeepSpeed to fit models using lower degree of model parallelism and higher batch size, offering significant performance gains compared to using model parallelism alone. Read more: ZeRO paper, and GPT tutorial.
DeepSpeed Speedup

The figure depicts system throughput improvements of DeepSpeed (combining ZeRO-powered data parallelism with model parallelism of NVIDIA Megatron-LM) over using Megatron-LM alone.

Fast convergence for effectiveness

DeepSpeed supports advanced hyperparameter tuning and large batch size optimizers such as LAMB. These improve the effectiveness of model training and reduce the number of samples required to convergence to desired accuracy.

Read the Tuning tutorial >

Good Usability

Only a few lines of code changes are needed to enable a PyTorch model to use DeepSpeed and ZeRO. Compared to current model parallelism libraries, DeepSpeed does not require a code redesign or model refactoring. It also does not put limitations on model dimensions (such as number of attention heads, hidden sizes, and others), batch size, or any other training parameters. For models of up to 13 billion parameters, you can use ZeRO-powered data parallelism conveniently without requiring model parallelism, while in contrast, standard data parallelism will run out of memory for models with more than 1.4 billion parameters. In addition, DeepSpeed conveniently supports flexible combination of ZeRO-powered data parallelism with custom model parallelisms, such as tensor slicing of NVIDIA’s Megatron-LM.

Features

DeepSpeed features

Below we provide a brief feature list, see our detailed feature overview for descriptions and usage.

Contributing

Contributing to DeepSpeed

DeepSpeed welcomes your contributions! Please see our contributing guide for more details on formatting, testing, etc.

Contributor License Agreement

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.opensource.microsoft.com.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

Code of Conduct

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact opencode@microsoft.com with any additional questions or comments.