Memory-Efficient Pipeline-Parallel DNN Training

• URL: http://arxiv.org/abs/2006.09503v3
• Date: Thu, 22 Jul 2021 17:25:58 GMT
• Title: Memory-Efficient Pipeline-Parallel DNN Training
• Authors: Deepak Narayanan, Amar Phanishayee, Kaiyu Shi, Xie Chen, Matei Zaharia
• Abstract summary: PipeDream-2BW is a system that supports memory-efficient pipeline parallelism. It can accelerate the training of large GPT and BERT language models by up to 20$times$ with similar final model accuracy.
• Score: 27.83107540482083
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Many state-of-the-art ML results have been obtained by scaling up the number of parameters in existing models. However, parameters and activations for such large models often do not fit in the memory of a single accelerator device; this means that it is necessary to distribute training of large models over multiple accelerators. In this work, we propose PipeDream-2BW, a system that supports memory-efficient pipeline parallelism. PipeDream-2BW uses a novel pipelining and weight gradient coalescing strategy, combined with the double buffering of weights, to ensure high throughput, low memory footprint, and weight update semantics similar to data parallelism. In addition, PipeDream-2BW automatically partitions the model over the available hardware resources, while respecting hardware constraints such as memory capacities of accelerators and interconnect topologies. PipeDream-2BW can accelerate the training of large GPT and BERT language models by up to 20$\times$ with similar final model accuracy.

Related papers

• Superpipeline: A Universal Approach for Reducing GPU Memory Usage in Large Models [40.41898661688188]
  This paper introduces Superpipeline, a framework designed to optimize the execution of large AI models on constrained hardware. Superpipeline reduces GPU memory usage by up to 60% in our experiments while maintaining model accuracy and acceptable processing speeds.
  arXiv  Detail & Related papers  (2024-10-11T13:17:05Z)
• SHERL: Synthesizing High Accuracy and Efficient Memory for Resource-Limited Transfer Learning [63.93193829913252]
  We propose an innovative METL strategy called SHERL for resource-limited scenarios. In the early route, intermediate outputs are consolidated via an anti-redundancy operation. In the late route, utilizing minimal late pre-trained layers could alleviate the peak demand on memory overhead.
  arXiv  Detail & Related papers  (2024-07-10T10:22:35Z)
• Pipette: Automatic Fine-grained Large Language Model Training Configurator for Real-World Clusters [5.190794062263327]
  Training large language models (LLMs) is known to be challenging because of the huge computational and memory capacity requirements. We propose Pipette, which is an automatic fine-grained LLM training for real-world clusters.
  arXiv  Detail & Related papers  (2024-05-28T11:59:44Z)
• 2BP: 2-Stage Backpropagation [0.0]
  This paper introduces 2-stage backpropagation (2BP) By splitting the backward propagation step into two separate stages, we can reduce idle compute time. Using 2BP, we were able to achieve a 1.70x increase in throughput compared to traditional methods.
  arXiv  Detail & Related papers  (2024-05-28T11:02:01Z)
• Partitioned Neural Network Training via Synthetic Intermediate Labels [0.0]
  GPU memory constraints have become a notable bottleneck in training such sizable models. This study advocates partitioning the model across GPU and generating synthetic intermediate labels to train individual segments. This approach results in a more efficient training process that minimizes data communication while maintaining model accuracy.
  arXiv  Detail & Related papers  (2024-03-17T13:06:29Z)
• BitDelta: Your Fine-Tune May Only Be Worth One Bit [57.558376557639555]
  Large Language Models (LLMs) are typically trained in two phases: pre-training on large internet-scale datasets, and fine-tuning for downstream tasks. We introduce a simple method, BitDelta, which successfully quantizes this delta down to 1 bit without compromising performance. By enabling the use of a single high-precision base model accompanied by multiple 1-bit deltas, BitDelta dramatically reduces GPU memory requirements by more than 10x.
  arXiv  Detail & Related papers  (2024-02-15T18:50:06Z)
• Pipeline MoE: A Flexible MoE Implementation with Pipeline Parallelism [91.9372563527801]
  Existing MoE models suffer from tremendous inner-node and inter-node communication overhead. We propose a novel MoE architecture called Pipeline MoE (PPMoE) to tackle them. PPMoE builds expert parallel incorporating with tensor parallel and replaces communication-intensive all-to-all dispatching and gathering.
  arXiv  Detail & Related papers  (2023-04-22T14:09:14Z)
• SWARM Parallelism: Training Large Models Can Be Surprisingly Communication-Efficient [69.61083127540776]
  Deep learning applications benefit from using large models with billions of parameters. Training these models is notoriously expensive due to the need for specialized HPC clusters. We consider alternative setups for training large models: using cheap "preemptible" instances or pooling existing resources from multiple regions.
  arXiv  Detail & Related papers  (2023-01-27T18:55:19Z)
• M6-10T: A Sharing-Delinking Paradigm for Efficient Multi-Trillion Parameter Pretraining [55.16088793437898]
  Training extreme-scale models requires enormous amounts of computes and memory footprint. We propose a simple training strategy called "Pseudo-to-Real" for high-memory-footprint-required large models.
  arXiv  Detail & Related papers  (2021-10-08T04:24:51Z)
• Automatic Graph Partitioning for Very Large-scale Deep Learning [4.472135966077758]
  This work proposes RaNNC (Rapid Neural Network Connector) as for automatic hybrid parallelism. RaNNC automatically partitions the model into a set of subcomponents so that each subcomponent fits a device memory. RaNNC successfully trained models five times larger than those Megatron-LM could, and RaNNC's training throughputs were comparable to Megatron-LM's when pre-training the same models.
  arXiv  Detail & Related papers  (2021-03-30T04:26:04Z)
• Scaling Distributed Deep Learning Workloads beyond the Memory Capacity with KARMA [58.040931661693925]
  We propose a strategy that combines redundant recomputing and out-of-core methods. We achieve an average of 1.52x speedup in six different models over the state-of-the-art out-of-core methods. Our data parallel out-of-core solution can outperform complex hybrid model parallelism in training large models, e.g. Megatron-LM and Turning-NLG.
  arXiv  Detail & Related papers  (2020-08-26T07:24:34Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.