Demystifying Workload Imbalances in Large Transformer Model Training over Variable-length Sequences

• URL: http://arxiv.org/abs/2412.07894v1
• Date: Tue, 10 Dec 2024 20:01:53 GMT
• Title: Demystifying Workload Imbalances in Large Transformer Model Training over Variable-length Sequences
• Authors: Haoyang Li, Fangcheng Fu, Sheng Lin, Hao Ge, Xuanyu Wang, Jiawen Niu, Jie Jiang, Bin Cui,
• Abstract summary: We develop Hydraulis, which jointly optimize the parallel strategies and data assignment. Empirical results demonstrate that Hydraulis outperforms existing systems by 1.32-2.66 times.
• Score: 31.232756326457277
• License:
• Abstract: To optimize large Transformer model training, efficient parallel computing and advanced data management are essential. However, current methods often assume a stable and uniform training workload, neglecting imbalances in data sampling and packing that can impede performance. Specifically, data sampling imbalance arises from uneven sequence length distribution of the training data, while data packing imbalance stems from the discrepancy between the linear memory complexity and quadratic time complexity of the attention mechanism. To address these imbalance issues, we develop Hydraulis, which jointly optimizes the parallel strategies and data assignment. For one thing, we introduce large model training with dynamic heterogeneous parallel strategies in response to the sequence length variations within and across training iterations. For another, we devise a two-stage data assignment approach, which strikes a good balance in terms of the training workloads both within and across model replicas. Empirical results demonstrate that Hydraulis outperforms existing systems by 1.32-2.66 times.

Related papers

• FlexSP: Accelerating Large Language Model Training via Flexible Sequence Parallelism [33.23902060961886]
  Existing sequence parallelism methods assume homogeneous sequence lengths (i.e., all input sequences are equal in length) and therefore leverages a single, static scattering strategy for all input sequences. We show that the sequence lengths in LLM training corpora exhibit substantial variability, often following a long-tail distribution. We propose a Heterogeneous-adaptive sequence parallelism method to address this problem.
  arXiv  Detail & Related papers  (2024-12-02T14:16:03Z)
• Efficient Grammatical Error Correction Via Multi-Task Training and Optimized Training Schedule [55.08778142798106]
  We propose auxiliary tasks that exploit the alignment between the original and corrected sentences. We formulate each task as a sequence-to-sequence problem and perform multi-task training. We find that the order of datasets used for training and even individual instances within a dataset may have important effects on the final performance.
  arXiv  Detail & Related papers  (2023-11-20T14:50:12Z)
• AdaMerging: Adaptive Model Merging for Multi-Task Learning [68.75885518081357]
  This paper introduces an innovative technique called Adaptive Model Merging (AdaMerging) It aims to autonomously learn the coefficients for model merging, either in a task-wise or layer-wise manner, without relying on the original training data. Compared to the current state-of-the-art task arithmetic merging scheme, AdaMerging showcases a remarkable 11% improvement in performance.
  arXiv  Detail & Related papers  (2023-10-04T04:26:33Z)
• Alleviating the Effect of Data Imbalance on Adversarial Training [26.36714114672729]
  We study adversarial training on datasets that obey the long-tailed distribution. We propose a new adversarial training framework -- Re-balancing Adversarial Training (REAT)
  arXiv  Detail & Related papers  (2023-07-14T07:01:48Z)
• RWKV: Reinventing RNNs for the Transformer Era [54.716108899349614]
  We propose a novel model architecture that combines the efficient parallelizable training of transformers with the efficient inference of RNNs. We scale our models as large as 14 billion parameters, by far the largest dense RNN ever trained, and find RWKV performs on par with similarly sized Transformers.
  arXiv  Detail & Related papers  (2023-05-22T13:57:41Z)
• Does compressing activations help model parallel training? [64.59298055364336]
  We present the first empirical study on the effectiveness of compression methods for model parallelism. We implement and evaluate three common classes of compression algorithms. We evaluate these methods across more than 160 settings and 8 popular datasets.
  arXiv  Detail & Related papers  (2023-01-06T18:58:09Z)
• Data splitting improves statistical performance in overparametrized regimes [0.0]
  Distributed learning is a common strategy to reduce the overall training time by exploiting multiple computing devices. We show that in this regime, data splitting has a regularizing effect, hence improving statistical performance and computational complexity.
  arXiv  Detail & Related papers  (2021-10-21T08:10:56Z)
• An Accurate and Efficient Large-scale Regression Method through Best Friend Clustering [10.273838113763192]
  We propose a novel and simple data structure capturing the most important information among data samples. We combine the clustering with regression techniques as a parallel library and utilize a hybrid structure of data and model parallelism to make predictions.
  arXiv  Detail & Related papers  (2021-04-22T01:34:29Z)
• Training Transformers for Information Security Tasks: A Case Study on Malicious URL Prediction [3.660098145214466]
  We implement a malicious/benign predictor URL based on a transformer architecture that is trained from scratch. We show that in contrast to conventional natural language processing (NLP) transformers, this model requires a different training approach to work well.
  arXiv  Detail & Related papers  (2020-11-05T18:58:51Z)
• 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)
• Understanding the Effects of Data Parallelism and Sparsity on Neural Network Training [126.49572353148262]
  We study two factors in neural network training: data parallelism and sparsity. Despite their promising benefits, understanding of their effects on neural network training remains elusive.
  arXiv  Detail & Related papers  (2020-03-25T10:49:22Z)

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.