Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userThe Case for Strong Scaling in Deep Learning: Training Large 3D CNNs with Hybrid Parallelism
90
0
0.0
(
0
)
Added by
Yosuke Oyama
Publication date
2020
fields
Informatics Engineering
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
We present scalable hybrid-parallel algorithms for training large-scale 3D convolutional neural networks. Deep learning-based emerging scientific workflows often require model training with large, high-dimensional samples, which can make training much more costly and even infeasible due to excessive memory usage. We solve these challenges by extensively applying hybrid parallelism throughout the end-to-end training pipeline, including both computations and I/O. Our hybrid-parallel algorithm extends the standard data parallelism with spatial parallelism, which partitions a single sample in the spatial domain, realizing strong scaling beyond the mini-batch dimension with a larger aggregated memory capacity. We evaluate our proposed training algorithms with two challenging 3D CNNs, CosmoFlow and 3D U-Net. Our comprehensive performance studies show that good weak and strong scaling can be achieved for both networks using up 2K GPUs. More importantly, we enable training of CosmoFlow with much larger samples than previously possible, realizing an order-of-magnitude improvement in prediction accuracy.
rate research
Read More
A good parallelization strategy can significantly improve the efficiency or reduce the cost for the distributed training of deep neural networks (DNNs). Recently, several methods have been proposed to find efficient parallelization strategies but they all optimize a single objective (e.g., execution time, memory consumption) and produce only one strategy. We propose FT, an efficient algorithm that searches for an optimal set of parallelization strategies to allow the trade-off among different objectives. FT can adapt to different scenarios by minimizing the memory consumption when the number of devices is limited and fully utilize additional resources to reduce the execution time. For popular DNN models (e.g., vision, language), an in-depth analysis is conducted to understand the trade-offs among different objectives and their influence on the parallelization strategies. We also develop a user-friendly system, called TensorOpt, which allows users to run their distributed DNN training jobs without caring the details of parallelization strategies. Experimental results show that FT runs efficiently and provides accurate estimation of runtime costs, and TensorOpt is more flexible in adapting to resource availability compared with existing frameworks.
The recent Natural Language Processing techniques have been refreshing the state-of-the-art performance at an incredible speed. Training huge language models is therefore an imperative demand in both industry and academy. However, huge language models impose challenges to both hardware and software. Graphical processing units (GPUs) are iterated frequently to meet the exploding demand, and a variety of ASICs like TPUs are spawned. However, there is still a tension between the fast growth of the extremely huge models and the fact that Moores law is approaching the end. To this end, many model parallelism techniques are proposed to distribute the model parameters to multiple devices, so as to alleviate the tension on both memory and computation. Our work is the first to introduce a 3-dimensional model parallelism for expediting huge language models. By reaching a perfect load balance, our approach presents smaller memory and communication cost than existing state-of-the-art 1-D and 2-D model parallelism. Our experiments on 64 TACCs V100 GPUs show that our 3-D parallelism outperforms the 1-D and 2-D parallelism with 2.32x and 1.57x speedup, respectively.
Scaling up deep neural networks has been proven effective in improving model quality, while it also brings ever-growing training challenges. This paper presents Whale, an automatic and hardware-aware distributed training framework for giant models. Whale generalizes the expression of parallelism with four primitives, which can define various parallel strategies, as well as flexible hybrid strategies including combination and nesting patterns. It allows users to build models at an arbitrary scale by adding a few annotations and automatically transforms the local model to a distributed implementation. Moreover, Whale is hardware-aware and highly efficient even when training on GPUs of mixed types, which meets the growing demand of heterogeneous training in industrial clusters. Whale sets a milestone for training the largest multimodal pretrained model M6. The success of M6 is achieved by Whales design to decouple algorithm modeling from system implementations, i.e., algorithm developers can focus on model innovation, since it takes only three lines of code to scale the M6 model to trillions of parameters on a cluster of 480 GPUs.
Scale of data and scale of computation infrastructures together enable the current deep learning renaissance. However, training large-scale deep architectures demands both algorithmic improvement and careful system configuration. In this paper, we focus on employing the system approach to speed up large-scale training. Via lessons learned from our routine benchmarking effort, we first identify bottlenecks and overheads that hinter data parallelism. We then devise guidelines that help practitioners to configure an effective system and fine-tune parameters to achieve desired speedup. Specifically, we develop a procedure for setting minibatch size and choosing computation algorithms. We also derive lemmas for determining the quantity of key components such as the number of GPUs and parameter servers. Experiments and examples show that these guidelines help effectively speed up large-scale deep learning training.
Stochastic gradient descent (SGD) is an inherently sequential training algorithm--computing the gradient at batch $i$ depends on the model parameters learned from batch $i-1$. Prior approaches that break this dependence do not honor them (e.g., sum the gradients for each batch, which is not what sequential SGD would do) and thus potentially suffer from poor convergence. This paper introduces a novel method to combine gradients called Adasum (for adaptive sum) that converges faster than prior work. Adasum is easy to implement, almost as efficient as simply summing gradients, and is integrated into the open-source toolkit Horovod. This paper first provides a formal justification for Adasum and then empirically demonstrates Adasum is more accurate than prior gradient accumulation methods. It then introduces a series of case-studies to show Adasum works with multiple frameworks, (TensorFlow and PyTorch), scales multiple optimizers (Momentum-SGD, Adam, and LAMB) to larger batch-sizes while still giving good downstream accuracy. Finally, it proves that Adasum converges. To summarize, Adasum scales Momentum-SGD on the MLPerf Resnet50 benchmark to 64K examples before communication (no MLPerf v0.5 entry converged with more than 16K), the Adam optimizer to 64K examples before communication on BERT-LARGE (prior work showed Adam stopped scaling at 16K), and the LAMB optimizer to 128K before communication on BERT-LARGE (prior work used 64K), all while maintaining downstream accuracy metrics. Finally, if a user does not need to scale, we show LAMB with Adasum on BERT-LARGE converges in 30% fewer steps than the baseline.
suggested questions
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
