Showing 1–41 of 41 results for author: Patwary, M

LLM Pruning and Distillation in Practice: The Minitron Approach

Authors: Sharath Turuvekere Sreenivas, Saurav Muralidharan, Raviraj Joshi, Marcin Chochowski, Mostofa Patwary, Mohammad Shoeybi, Bryan Catanzaro, Jan Kautz, Pavlo Molchanov

Abstract: We present a comprehensive report on compressing the Llama 3.1 8B and Mistral NeMo 12B models to 4B and 8B parameters, respectively, using pruning and distillation. We explore two distinct pruning strategies: (1) depth pruning and (2) joint hidden/attention/MLP (width) pruning, and evaluate the results on common benchmarks from the LM Evaluation Harness. The models are then aligned with NeMo Align… ▽ More We present a comprehensive report on compressing the Llama 3.1 8B and Mistral NeMo 12B models to 4B and 8B parameters, respectively, using pruning and distillation. We explore two distinct pruning strategies: (1) depth pruning and (2) joint hidden/attention/MLP (width) pruning, and evaluate the results on common benchmarks from the LM Evaluation Harness. The models are then aligned with NeMo Aligner and tested in instruct-tuned versions. This approach produces a compelling 4B model from Llama 3.1 8B and a state-of-the-art Mistral-NeMo-Minitron-8B (MN-Minitron-8B for brevity) model from Mistral NeMo 12B. We found that with no access to the original data, it is beneficial to slightly fine-tune teacher models on the distillation dataset. We open-source our base model weights on Hugging Face with a permissive license. △ Less

Submitted 26 August, 2024; v1 submitted 21 August, 2024; originally announced August 2024.

Comments: v2: Added missing references. Cleaned up runtime performance section

Compact Language Models via Pruning and Knowledge Distillation

Authors: Saurav Muralidharan, Sharath Turuvekere Sreenivas, Raviraj Joshi, Marcin Chochowski, Mostofa Patwary, Mohammad Shoeybi, Bryan Catanzaro, Jan Kautz, Pavlo Molchanov

Abstract: Large language models (LLMs) targeting different deployment scales and sizes are currently produced by training each variant from scratch; this is extremely compute-intensive. In this paper, we investigate if pruning an existing LLM and then re-training it with a fraction (<3%) of the original training data can be a suitable alternative to repeated, full retraining. To this end, we develop a set o… ▽ More Large language models (LLMs) targeting different deployment scales and sizes are currently produced by training each variant from scratch; this is extremely compute-intensive. In this paper, we investigate if pruning an existing LLM and then re-training it with a fraction (<3%) of the original training data can be a suitable alternative to repeated, full retraining. To this end, we develop a set of practical and effective compression best practices for LLMs that combine depth, width, attention and MLP pruning with knowledge distillation-based retraining; we arrive at these best practices through a detailed empirical exploration of pruning strategies for each axis, methods to combine axes, distillation strategies, and search techniques for arriving at optimal compressed architectures. We use this guide to compress the Nemotron-4 family of LLMs by a factor of 2-4x, and compare their performance to similarly-sized models on a variety of language modeling tasks. Deriving 8B and 4B models from an already pretrained 15B model using our approach requires up to 40x fewer training tokens per model compared to training from scratch; this results in compute cost savings of 1.8x for training the full model family (15B, 8B, and 4B). Minitron models exhibit up to a 16% improvement in MMLU scores compared to training from scratch, perform comparably to other community models such as Mistral 7B, Gemma 7B and Llama-3 8B, and outperform state-of-the-art compression techniques from the literature. We have open-sourced Minitron model weights on Huggingface, with corresponding supplementary material including example code available on GitHub. △ Less

Submitted 19 July, 2024; originally announced July 2024.

Reuse, Don't Retrain: A Recipe for Continued Pretraining of Language Models

Authors: Jupinder Parmar, Sanjev Satheesh, Mostofa Patwary, Mohammad Shoeybi, Bryan Catanzaro

Abstract: As language models have scaled both their number of parameters and pretraining dataset sizes, the computational cost for pretraining has become intractable except for the most well-resourced teams. This increasing cost makes it ever more important to be able to reuse a model after it has completed pretraining; allowing for a model's abilities to further improve without needing to train from scratc… ▽ More As language models have scaled both their number of parameters and pretraining dataset sizes, the computational cost for pretraining has become intractable except for the most well-resourced teams. This increasing cost makes it ever more important to be able to reuse a model after it has completed pretraining; allowing for a model's abilities to further improve without needing to train from scratch. In this work, we detail a set of guidelines that cover how to design efficacious data distributions and learning rate schedules for continued pretraining of language models. When applying these findings within a continued pretraining run on top of a well-trained 15B parameter model, we show an improvement of 9\% in average model accuracy compared to the baseline of continued training on the pretraining set. The resulting recipe provides a practical starting point with which to begin developing language models through reuse rather than retraining. △ Less

Submitted 9 July, 2024; originally announced July 2024.

Comments: Preprint. Under review

Data, Data Everywhere: A Guide for Pretraining Dataset Construction

Authors: Jupinder Parmar, Shrimai Prabhumoye, Joseph Jennings, Bo Liu, Aastha Jhunjhunwala, Zhilin Wang, Mostofa Patwary, Mohammad Shoeybi, Bryan Catanzaro

Abstract: The impressive capabilities of recent language models can be largely attributed to the multi-trillion token pretraining datasets that they are trained on. However, model developers fail to disclose their construction methodology which has lead to a lack of open information on how to develop effective pretraining sets. To address this issue, we perform the first systematic study across the entire p… ▽ More The impressive capabilities of recent language models can be largely attributed to the multi-trillion token pretraining datasets that they are trained on. However, model developers fail to disclose their construction methodology which has lead to a lack of open information on how to develop effective pretraining sets. To address this issue, we perform the first systematic study across the entire pipeline of pretraining set construction. First, we run ablations on existing techniques for pretraining set development to identify which methods translate to the largest gains in model accuracy on downstream evaluations. Then, we categorize the most widely used data source, web crawl snapshots, across the attributes of toxicity, quality, type of speech, and domain. Finally, we show how such attribute information can be used to further refine and improve the quality of a pretraining set. These findings constitute an actionable set of steps that practitioners can use to develop high quality pretraining sets. △ Less

Submitted 8 July, 2024; originally announced July 2024.

Comments: Preprint. Under review

Nemotron-4 340B Technical Report

Authors: Nvidia, :, Bo Adler, Niket Agarwal, Ashwath Aithal, Dong H. Anh, Pallab Bhattacharya, Annika Brundyn, Jared Casper, Bryan Catanzaro, Sharon Clay, Jonathan Cohen, Sirshak Das, Ayush Dattagupta, Olivier Delalleau, Leon Derczynski, Yi Dong, Daniel Egert, Ellie Evans, Aleksander Ficek, Denys Fridman, Shaona Ghosh, Boris Ginsburg, Igor Gitman, Tomasz Grzegorzek , et al. (58 additional authors not shown)

Abstract: We release the Nemotron-4 340B model family, including Nemotron-4-340B-Base, Nemotron-4-340B-Instruct, and Nemotron-4-340B-Reward. Our models are open access under the NVIDIA Open Model License Agreement, a permissive model license that allows distribution, modification, and use of the models and its outputs. These models perform competitively to open access models on a wide range of evaluation be… ▽ More We release the Nemotron-4 340B model family, including Nemotron-4-340B-Base, Nemotron-4-340B-Instruct, and Nemotron-4-340B-Reward. Our models are open access under the NVIDIA Open Model License Agreement, a permissive model license that allows distribution, modification, and use of the models and its outputs. These models perform competitively to open access models on a wide range of evaluation benchmarks, and were sized to fit on a single DGX H100 with 8 GPUs when deployed in FP8 precision. We believe that the community can benefit from these models in various research studies and commercial applications, especially for generating synthetic data to train smaller language models. Notably, over 98% of data used in our model alignment process is synthetically generated, showcasing the effectiveness of these models in generating synthetic data. To further support open research and facilitate model development, we are also open-sourcing the synthetic data generation pipeline used in our model alignment process. △ Less

Submitted 6 August, 2024; v1 submitted 17 June, 2024; originally announced June 2024.

Structure Editor for Building Software Models

Authors: Mohammad Nurullah Patwary, Ana Jovanovic, Allison Sullivan

Abstract: Alloy is well known a declarative modeling language. A key strength of Alloy is its scenario finding toolset, the Analyzer, which allows users to explore all valid scenarios that adhere to the model's constraints up to a user-provided scope. Despite the Analyzer, Alloy is still difficult for novice users to learn and use. A recent empirical study of over 93,000 new user models reveals that users h… ▽ More Alloy is well known a declarative modeling language. A key strength of Alloy is its scenario finding toolset, the Analyzer, which allows users to explore all valid scenarios that adhere to the model's constraints up to a user-provided scope. Despite the Analyzer, Alloy is still difficult for novice users to learn and use. A recent empirical study of over 93,000 new user models reveals that users have trouble from the very start: nearly a third of the models novices write fail to compile. We believe that the issue is that Alloy's grammar and type information is passively relayed to the user despite this information outlining a narrow path for how to compose valid formulas. In this paper, we outline a proof-of-concept for a structure editor for Alloy in which user's build their models using block based inputs, rather than free typing, which by design prevents compilation errors. △ Less

Submitted 13 June, 2024; originally announced June 2024.

StarCoder 2 and The Stack v2: The Next Generation

Authors: Anton Lozhkov, Raymond Li, Loubna Ben Allal, Federico Cassano, Joel Lamy-Poirier, Nouamane Tazi, Ao Tang, Dmytro Pykhtar, Jiawei Liu, Yuxiang Wei, Tianyang Liu, Max Tian, Denis Kocetkov, Arthur Zucker, Younes Belkada, Zijian Wang, Qian Liu, Dmitry Abulkhanov, Indraneil Paul, Zhuang Li, Wen-Ding Li, Megan Risdal, Jia Li, Jian Zhu, Terry Yue Zhuo , et al. (41 additional authors not shown)

Abstract: The BigCode project, an open-scientific collaboration focused on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder2. In partnership with Software Heritage (SWH), we build The Stack v2 on top of the digital commons of their source code archive. Alongside the SWH repositories spanning 619 programming languages, we carefully select other high-quality data… ▽ More The BigCode project, an open-scientific collaboration focused on the responsible development of Large Language Models for Code (Code LLMs), introduces StarCoder2. In partnership with Software Heritage (SWH), we build The Stack v2 on top of the digital commons of their source code archive. Alongside the SWH repositories spanning 619 programming languages, we carefully select other high-quality data sources, such as GitHub pull requests, Kaggle notebooks, and code documentation. This results in a training set that is 4x larger than the first StarCoder dataset. We train StarCoder2 models with 3B, 7B, and 15B parameters on 3.3 to 4.3 trillion tokens and thoroughly evaluate them on a comprehensive set of Code LLM benchmarks. We find that our small model, StarCoder2-3B, outperforms other Code LLMs of similar size on most benchmarks, and also outperforms StarCoderBase-15B. Our large model, StarCoder2- 15B, significantly outperforms other models of comparable size. In addition, it matches or outperforms CodeLlama-34B, a model more than twice its size. Although DeepSeekCoder- 33B is the best-performing model at code completion for high-resource languages, we find that StarCoder2-15B outperforms it on math and code reasoning benchmarks, as well as several low-resource languages. We make the model weights available under an OpenRAIL license and ensure full transparency regarding the training data by releasing the SoftWare Heritage persistent IDentifiers (SWHIDs) of the source code data. △ Less

Submitted 29 February, 2024; originally announced February 2024.

Nemotron-4 15B Technical Report

Authors: Jupinder Parmar, Shrimai Prabhumoye, Joseph Jennings, Mostofa Patwary, Sandeep Subramanian, Dan Su, Chen Zhu, Deepak Narayanan, Aastha Jhunjhunwala, Ayush Dattagupta, Vibhu Jawa, Jiwei Liu, Ameya Mahabaleshwarkar, Osvald Nitski, Annika Brundyn, James Maki, Miguel Martinez, Jiaxuan You, John Kamalu, Patrick LeGresley, Denys Fridman, Jared Casper, Ashwath Aithal, Oleksii Kuchaiev, Mohammad Shoeybi , et al. (2 additional authors not shown)

Abstract: We introduce Nemotron-4 15B, a 15-billion-parameter large multilingual language model trained on 8 trillion text tokens. Nemotron-4 15B demonstrates strong performance when assessed on English, multilingual, and coding tasks: it outperforms all existing similarly-sized open models on 4 out of 7 downstream evaluation areas and achieves competitive performance to the leading open models in the remai… ▽ More We introduce Nemotron-4 15B, a 15-billion-parameter large multilingual language model trained on 8 trillion text tokens. Nemotron-4 15B demonstrates strong performance when assessed on English, multilingual, and coding tasks: it outperforms all existing similarly-sized open models on 4 out of 7 downstream evaluation areas and achieves competitive performance to the leading open models in the remaining ones. Specifically, Nemotron-4 15B exhibits the best multilingual capabilities of all similarly-sized models, even outperforming models over four times larger and those explicitly specialized for multilingual tasks. △ Less

Submitted 27 February, 2024; v1 submitted 26 February, 2024; originally announced February 2024.

BadODD: Bangladeshi Autonomous Driving Object Detection Dataset

Authors: Mirza Nihal Baig, Rony Hajong, Mahdi Murshed Patwary, Mohammad Shahidur Rahman, Husne Ara Chowdhury

Abstract: We propose a comprehensive dataset for object detection in diverse driving environments across 9 districts in Bangladesh. The dataset, collected exclusively from smartphone cameras, provided a realistic representation of real-world scenarios, including day and night conditions. Most existing datasets lack suitable classes for autonomous navigation on Bangladeshi roads, making it challenging for re… ▽ More We propose a comprehensive dataset for object detection in diverse driving environments across 9 districts in Bangladesh. The dataset, collected exclusively from smartphone cameras, provided a realistic representation of real-world scenarios, including day and night conditions. Most existing datasets lack suitable classes for autonomous navigation on Bangladeshi roads, making it challenging for researchers to develop models that can handle the intricacies of road scenarios. To address this issue, the authors proposed a new set of classes based on characteristics rather than local vehicle names. The dataset aims to encourage the development of models that can handle the unique challenges of Bangladeshi road scenarios for the effective deployment of autonomous vehicles. The dataset did not consist of any online images to simulate real-world conditions faced by autonomous vehicles. The classification of vehicles is challenging because of the diverse range of vehicles on Bangladeshi roads, including those not found elsewhere in the world. The proposed classification system is scalable and can accommodate future vehicles, making it a valuable resource for researchers in the autonomous vehicle sector. △ Less

Submitted 19 January, 2024; originally announced January 2024.

Comments: 7 pages

Adding Instructions during Pretraining: Effective Way of Controlling Toxicity in Language Models

Authors: Shrimai Prabhumoye, Mostofa Patwary, Mohammad Shoeybi, Bryan Catanzaro

Abstract: Pretrained large language models have become indispensable for solving various natural language processing (NLP) tasks. However, safely deploying them in real world applications is challenging because they generate toxic content. To address this challenge, we propose two novel pretraining data augmentation strategies that significantly reduce model toxicity without compromising its utility. Our tw… ▽ More Pretrained large language models have become indispensable for solving various natural language processing (NLP) tasks. However, safely deploying them in real world applications is challenging because they generate toxic content. To address this challenge, we propose two novel pretraining data augmentation strategies that significantly reduce model toxicity without compromising its utility. Our two strategies are: (1) MEDA: adds raw toxicity score as meta-data to the pretraining samples, and (2) INST: adds instructions to those samples indicating their toxicity. Our results indicate that our best performing strategy (INST) substantially reduces the toxicity probability up to 61% while preserving the accuracy on five benchmark NLP tasks as well as improving AUC scores on four bias detection tasks by 1.3%. We also demonstrate the generalizability of our techniques by scaling the number of training samples and the number of model parameters. △ Less

Submitted 14 February, 2023; originally announced February 2023.

Comments: This paper will be presented at EACL 2023

Evaluating Parameter Efficient Learning for Generation

Authors: Peng Xu, Mostofa Patwary, Shrimai Prabhumoye, Virginia Adams, Ryan J. Prenger, Wei Ping, Nayeon Lee, Mohammad Shoeybi, Bryan Catanzaro

Abstract: Parameter efficient learning methods (PERMs) have recently gained significant attention as they provide an efficient way for pre-trained language models (PLMs) to adapt to a downstream task. However, these conclusions are mostly drawn from in-domain evaluations over the full training set. In this paper, we present comparisons between PERMs and finetuning from three new perspectives: (1) the effect… ▽ More Parameter efficient learning methods (PERMs) have recently gained significant attention as they provide an efficient way for pre-trained language models (PLMs) to adapt to a downstream task. However, these conclusions are mostly drawn from in-domain evaluations over the full training set. In this paper, we present comparisons between PERMs and finetuning from three new perspectives: (1) the effect of sample and model size to in-domain evaluations, (2) generalization to unseen domains and new datasets, and (3) the faithfulness of generations. Our results show that for in-domain settings (a) there is a cross point of sample size for which PERMs will perform better than finetuning when training with fewer samples, and (b) larger PLMs have larger cross points. For cross-domain and cross-dataset cases, we show that (a) Adapter (Houlsby et al., 2019) performs the best amongst all the PERMs studied here, and (b) it outperforms finetuning if the task dataset is below a certain size. We also compare the faithfulness of generations and show that PERMs can achieve better faithfulness score than finetuning, especially for small training set, by as much as 6%. Finally, we apply Adapter to MT-NLG 530b (Smith et al., 2022) and achieve new state-of-the-art results on Xsum (Narayan et al., 2018) for all ROUGE scores (ROUGE-1 49.17, ROUGE-2 27.20, ROUGE-L 40.98). △ Less

Submitted 24 October, 2022; originally announced October 2022.

Comments: Accepted to EMNLP 2022 main conference

Context Generation Improves Open Domain Question Answering

Authors: Dan Su, Mostofa Patwary, Shrimai Prabhumoye, Peng Xu, Ryan Prenger, Mohammad Shoeybi, Pascale Fung, Anima Anandkumar, Bryan Catanzaro

Abstract: Closed-book question answering (QA) requires a model to directly answer an open-domain question without access to any external knowledge. Prior work on closed-book QA either directly finetunes or prompts a pretrained language model (LM) to leverage the stored knowledge. However, they do not fully exploit the parameterized knowledge. To address this issue, we propose a two-stage, closed-book QA fra… ▽ More Closed-book question answering (QA) requires a model to directly answer an open-domain question without access to any external knowledge. Prior work on closed-book QA either directly finetunes or prompts a pretrained language model (LM) to leverage the stored knowledge. However, they do not fully exploit the parameterized knowledge. To address this issue, we propose a two-stage, closed-book QA framework which employs a coarse-to-fine approach to extract relevant knowledge and answer a question. Our approach first generates a related context for a given question by prompting a pretrained LM. We then prompt the same LM for answer prediction using the generated context and the question. Additionally, to eliminate failure caused by context uncertainty, we marginalize over generated contexts. Experimental results on three QA benchmarks show that our method significantly outperforms previous closed-book QA methods (e.g. exact matching 68.6% vs. 55.3%), and is on par with open-book methods that exploit external knowledge sources (e.g. 68.6% vs. 68.0%). Our method is able to better exploit the stored knowledge in pretrained LMs without adding extra learnable parameters or needing finetuning, and paves the way for hybrid models that integrate pretrained LMs with external knowledge. △ Less

Submitted 27 April, 2023; v1 submitted 12 October, 2022; originally announced October 2022.

Comments: 8 pages; Accepted at EACL2023

Factuality Enhanced Language Models for Open-Ended Text Generation

Authors: Nayeon Lee, Wei Ping, Peng Xu, Mostofa Patwary, Pascale Fung, Mohammad Shoeybi, Bryan Catanzaro

Abstract: Pretrained language models (LMs) are susceptible to generate text with nonfactual information. In this work, we measure and improve the factual accuracy of large-scale LMs for open-ended text generation. We design the FactualityPrompts test set and metrics to measure the factuality of LM generations. Based on that, we study the factual accuracy of LMs with parameter sizes ranging from 126M to 530B… ▽ More Pretrained language models (LMs) are susceptible to generate text with nonfactual information. In this work, we measure and improve the factual accuracy of large-scale LMs for open-ended text generation. We design the FactualityPrompts test set and metrics to measure the factuality of LM generations. Based on that, we study the factual accuracy of LMs with parameter sizes ranging from 126M to 530B. Interestingly, we find that larger LMs are more factual than smaller ones, although a previous study suggests that larger LMs can be less truthful in terms of misconceptions. In addition, popular sampling algorithms (e.g., top-p) in open-ended text generation can harm the factuality due to the ''uniform randomness'' introduced at every sampling step. We propose the factual-nucleus sampling algorithm that dynamically adapts the randomness to improve the factuality of generation while maintaining quality. Furthermore, we analyze the inefficiencies of the standard training method in learning correct associations between entities from factual text corpus (e.g., Wikipedia). We propose a factuality-enhanced training method that uses TopicPrefix for better awareness of facts and sentence completion as the training objective, which can vastly reduce the factual errors. We release our code and FactualityPrompts benchmark at: https://github.com/nayeon7lee/FactualityPrompt. △ Less

Submitted 2 March, 2023; v1 submitted 9 June, 2022; originally announced June 2022.

Comments: NeurIPS 2022

Multi-Stage Prompting for Knowledgeable Dialogue Generation

Authors: Zihan Liu, Mostofa Patwary, Ryan Prenger, Shrimai Prabhumoye, Wei Ping, Mohammad Shoeybi, Bryan Catanzaro

Abstract: Existing knowledge-grounded dialogue systems typically use finetuned versions of a pretrained language model (LM) and large-scale knowledge bases. These models typically fail to generalize on topics outside of the knowledge base, and require maintaining separate potentially large checkpoints each time finetuning is needed. In this paper, we aim to address these limitations by leveraging the inhere… ▽ More Existing knowledge-grounded dialogue systems typically use finetuned versions of a pretrained language model (LM) and large-scale knowledge bases. These models typically fail to generalize on topics outside of the knowledge base, and require maintaining separate potentially large checkpoints each time finetuning is needed. In this paper, we aim to address these limitations by leveraging the inherent knowledge stored in the pretrained LM as well as its powerful generation ability. We propose a multi-stage prompting approach to generate knowledgeable responses from a single pretrained LM. We first prompt the LM to generate knowledge based on the dialogue context. Then, we further prompt it to generate responses based on the dialogue context and the previously generated knowledge. Results show that our knowledge generator outperforms the state-of-the-art retrieval-based model by 5.8% when combining knowledge relevance and correctness. In addition, our multi-stage prompting outperforms the finetuning-based dialogue model in terms of response knowledgeability and engagement by up to 10% and 5%, respectively. Furthermore, we scale our model up to 530 billion parameters and show that larger LMs improve the generation correctness score by up to 10%, and response relevance, knowledgeability and engagement by up to 10%. Our code is available at: https://github.com/NVIDIA/Megatron-LM. △ Less

Submitted 16 March, 2022; originally announced March 2022.

Exploring the Limits of Domain-Adaptive Training for Detoxifying Large-Scale Language Models

Authors: Boxin Wang, Wei Ping, Chaowei Xiao, Peng Xu, Mostofa Patwary, Mohammad Shoeybi, Bo Li, Anima Anandkumar, Bryan Catanzaro

Abstract: Pre-trained language models (LMs) are shown to easily generate toxic language. In this work, we systematically explore domain-adaptive training to reduce the toxicity of language models. We conduct this study on three dimensions: training corpus, model size, and parameter efficiency. For the training corpus, we propose to leverage the generative power of LMs and generate nontoxic datasets for doma… ▽ More Pre-trained language models (LMs) are shown to easily generate toxic language. In this work, we systematically explore domain-adaptive training to reduce the toxicity of language models. We conduct this study on three dimensions: training corpus, model size, and parameter efficiency. For the training corpus, we propose to leverage the generative power of LMs and generate nontoxic datasets for domain-adaptive training, which mitigates the exposure bias and is shown to be more data-efficient than using a curated pre-training corpus. We demonstrate that the self-generation method consistently outperforms the existing baselines across various model sizes on both automatic and human evaluations, even when it uses a 1/3 smaller training corpus. We then comprehensively study detoxifying LMs with parameter sizes ranging from 126M up to 530B (3x larger than GPT-3), a scale that has never been studied before. We find that i) large LMs have similar toxicity levels as smaller ones given the same pre-training corpus, and ii) large LMs require more endeavor to detoxify. We also explore parameter-efficient training methods for detoxification. We demonstrate that adding and training adapter-only layers in LMs not only saves a lot of parameters but also achieves a better trade-off between toxicity and perplexity than whole model adaptation for the large-scale models. △ Less

Submitted 21 October, 2022; v1 submitted 8 February, 2022; originally announced February 2022.

Comments: NeurIPS 2022

Using DeepSpeed and Megatron to Train Megatron-Turing NLG 530B, A Large-Scale Generative Language Model

Authors: Shaden Smith, Mostofa Patwary, Brandon Norick, Patrick LeGresley, Samyam Rajbhandari, Jared Casper, Zhun Liu, Shrimai Prabhumoye, George Zerveas, Vijay Korthikanti, Elton Zhang, Rewon Child, Reza Yazdani Aminabadi, Julie Bernauer, Xia Song, Mohammad Shoeybi, Yuxiong He, Michael Houston, Saurabh Tiwary, Bryan Catanzaro

Abstract: Pretrained general-purpose language models can achieve state-of-the-art accuracies in various natural language processing domains by adapting to downstream tasks via zero-shot, few-shot and fine-tuning techniques. Because of their success, the size of these models has increased rapidly, requiring high-performance hardware, software, and algorithmic techniques to enable training such large models.… ▽ More Pretrained general-purpose language models can achieve state-of-the-art accuracies in various natural language processing domains by adapting to downstream tasks via zero-shot, few-shot and fine-tuning techniques. Because of their success, the size of these models has increased rapidly, requiring high-performance hardware, software, and algorithmic techniques to enable training such large models. As the result of a joint effort between Microsoft and NVIDIA, we present details on the training of the largest monolithic transformer based language model, Megatron-Turing NLG 530B (MT-NLG), with 530 billion parameters. In this paper, we first focus on the infrastructure as well as the 3D parallelism methodology used to train this model using DeepSpeed and Megatron. Next, we detail the training process, the design of our training corpus, and our data curation techniques, which we believe is a key ingredient to the success of the model. Finally, we discuss various evaluation results, as well as other interesting observations and new properties exhibited by MT-NLG. We demonstrate that MT-NLG achieves superior zero-, one-, and few-shot learning accuracies on several NLP benchmarks and establishes new state-of-the-art results. We believe that our contributions will help further the development of large-scale training infrastructures, large-scale language models, and natural language generations. △ Less

Submitted 4 February, 2022; v1 submitted 28 January, 2022; originally announced January 2022.

Comments: Shaden Smith and Mostofa Patwary contributed equally

SDP: Scalable Real-time Dynamic Graph Partitioner

Authors: Md Anwarul Kaium Patwary, Saurabh Garg, Sudheer Kumar Battula, Byeong Kang

Abstract: Time-evolving large graph has received attention due to their participation in real-world applications such as social networks and PageRank calculation. It is necessary to partition a large-scale dynamic graph in a streaming manner to overcome the memory bottleneck while partitioning the computational load. Reducing network communication and balancing the load between the partitions are the criter… ▽ More Time-evolving large graph has received attention due to their participation in real-world applications such as social networks and PageRank calculation. It is necessary to partition a large-scale dynamic graph in a streaming manner to overcome the memory bottleneck while partitioning the computational load. Reducing network communication and balancing the load between the partitions are the criteria to achieve effective run-time performance in graph partitioning. Moreover, an optimal resource allocation is needed to utilise the resources while catering the graph streams into the partitions. A number of existing partitioning algorithms (ADP, LogGP and LEOPARD) have been proposed to address the above problem. However, these partitioning methods are incapable of scaling the resources and handling the stream of data in real-time. In this study, we propose a dynamic graph partitioning method called Scalable Dynamic Graph Partitioner (SDP) using the streaming partitioning technique. The SDP contributes a novel vertex assigning method, communication-aware balancing method, and a scaling technique to produce an efficient dynamic graph partitioner. Experiment results show that the proposed method achieves up to 90% reduction of communication cost and 60%-70% balancing the load dynamically, compared with previous algorithms. Moreover, the proposed algorithm significantly reduces the execution time during partitioning. △ Less

Submitted 29 October, 2021; originally announced October 2021.

Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM

Authors: Deepak Narayanan, Mohammad Shoeybi, Jared Casper, Patrick LeGresley, Mostofa Patwary, Vijay Anand Korthikanti, Dmitri Vainbrand, Prethvi Kashinkunti, Julie Bernauer, Bryan Catanzaro, Amar Phanishayee, Matei Zaharia

Abstract: Large language models have led to state-of-the-art accuracies across a range of tasks. However, training these models efficiently is challenging for two reasons: a) GPU memory capacity is limited, making it impossible to fit large models on even a multi-GPU server, and b) the number of compute operations required to train these models can result in unrealistically long training times. Consequently… ▽ More Large language models have led to state-of-the-art accuracies across a range of tasks. However, training these models efficiently is challenging for two reasons: a) GPU memory capacity is limited, making it impossible to fit large models on even a multi-GPU server, and b) the number of compute operations required to train these models can result in unrealistically long training times. Consequently, new methods of model parallelism such as tensor and pipeline parallelism have been proposed. Unfortunately, naive usage of these methods leads to fundamental scaling issues at thousands of GPUs, e.g., due to expensive cross-node communication or devices spending significant time waiting on other devices to make progress. In this paper, we show how different types of parallelism methods (tensor, pipeline, and data parallelism) can be composed to scale to thousands of GPUs and models with trillions of parameters. We survey techniques for pipeline parallelism and propose a novel interleaved pipeline parallelism schedule that can improve throughput by 10+% with memory footprint comparable to existing approaches. We quantitatively study the trade-offs between tensor, pipeline, and data parallelism, and provide intuition as to how to configure distributed training of a large model. Our approach allows us to perform training iterations on a model with 1 trillion parameters at 502 petaFLOP/s on 3072 GPUs with achieved per-GPU throughput of 52% of theoretical peak. Our code is open sourced at https://github.com/nvidia/megatron-lm. △ Less

Submitted 23 August, 2021; v1 submitted 9 April, 2021; originally announced April 2021.

Comments: Accepted to SC 2021

End-to-End Training of Neural Retrievers for Open-Domain Question Answering

Authors: Devendra Singh Sachan, Mostofa Patwary, Mohammad Shoeybi, Neel Kant, Wei Ping, William L Hamilton, Bryan Catanzaro

Abstract: Recent work on training neural retrievers for open-domain question answering (OpenQA) has employed both supervised and unsupervised approaches. However, it remains unclear how unsupervised and supervised methods can be used most effectively for neural retrievers. In this work, we systematically study retriever pre-training. We first propose an approach of unsupervised pre-training with the Inverse… ▽ More Recent work on training neural retrievers for open-domain question answering (OpenQA) has employed both supervised and unsupervised approaches. However, it remains unclear how unsupervised and supervised methods can be used most effectively for neural retrievers. In this work, we systematically study retriever pre-training. We first propose an approach of unsupervised pre-training with the Inverse Cloze Task and masked salient spans, followed by supervised finetuning using question-context pairs. This approach leads to absolute gains of 2+ points over the previous best result in the top-20 retrieval accuracy on Natural Questions and TriviaQA datasets. We also explore two approaches for end-to-end supervised training of the reader and retriever components in OpenQA models. In the first approach, the reader considers each retrieved document separately while in the second approach, the reader considers all the retrieved documents together. Our experiments demonstrate the effectiveness of these approaches as we obtain new state-of-the-art results. On the Natural Questions dataset, we obtain a top-20 retrieval accuracy of 84, an improvement of 5 points over the recent DPR model. In addition, we achieve good results on answer extraction, outperforming recent models like REALM and RAG by 3+ points. We further scale up end-to-end training to large models and show consistent gains in performance over smaller models. △ Less

Submitted 1 June, 2021; v1 submitted 2 January, 2021; originally announced January 2021.

Comments: ACL 2021

Local Knowledge Powered Conversational Agents

Authors: Sashank Santhanam, Wei Ping, Raul Puri, Mohammad Shoeybi, Mostofa Patwary, Bryan Catanzaro

Abstract: State-of-the-art conversational agents have advanced significantly in conjunction with the use of large transformer-based language models. However, even with these advancements, conversational agents still lack the ability to produce responses that are informative and coherent with the local context. In this work, we propose a dialog framework that incorporates both local knowledge as well as user… ▽ More State-of-the-art conversational agents have advanced significantly in conjunction with the use of large transformer-based language models. However, even with these advancements, conversational agents still lack the ability to produce responses that are informative and coherent with the local context. In this work, we propose a dialog framework that incorporates both local knowledge as well as users' past dialogues to generate high quality conversations. We introduce an approach to build a dataset based on Reddit conversations, where outbound URL links are widely available in the conversations and the hyperlinked documents can be naturally included as local external knowledge. Using our framework and dataset, we demonstrate that incorporating local knowledge can largely improve informativeness, coherency and realisticness measures using human evaluations. In particular, our approach consistently outperforms the state-of-the-art conversational model on the Reddit dataset across all three measures. We also find that scaling the size of our models from 117M to 8.3B parameters yields consistent improvement of validation perplexity as well as human evaluated metrics. Our model with 8.3B parameters can generate human-like responses as rated by various human evaluations in a single-turn dialog setting. △ Less

Submitted 20 October, 2020; originally announced October 2020.

BioMegatron: Larger Biomedical Domain Language Model

Authors: Hoo-Chang Shin, Yang Zhang, Evelina Bakhturina, Raul Puri, Mostofa Patwary, Mohammad Shoeybi, Raghav Mani

Abstract: There has been an influx of biomedical domain-specific language models, showing language models pre-trained on biomedical text perform better on biomedical domain benchmarks than those trained on general domain text corpora such as Wikipedia and Books. Yet, most works do not study the factors affecting each domain language application deeply. Additionally, the study of model size on domain-specifi… ▽ More There has been an influx of biomedical domain-specific language models, showing language models pre-trained on biomedical text perform better on biomedical domain benchmarks than those trained on general domain text corpora such as Wikipedia and Books. Yet, most works do not study the factors affecting each domain language application deeply. Additionally, the study of model size on domain-specific models has been mostly missing. We empirically study and evaluate several factors that can affect performance on domain language applications, such as the sub-word vocabulary set, model size, pre-training corpus, and domain transfer. We show consistent improvements on benchmarks with our larger BioMegatron model trained on a larger domain corpus, contributing to our understanding of domain language model applications. We demonstrate noticeable improvements over the previous state-of-the-art (SOTA) on standard biomedical NLP benchmarks of named entity recognition, relation extraction, and question answering. Model checkpoints and code are available at [https://ngc.nvidia.com] and [https://github.com/NVIDIA/NeMo]. △ Less

Submitted 13 October, 2020; v1 submitted 12 October, 2020; originally announced October 2020.

Comments: Accepted for publication at EMNLP 2020

MEGATRON-CNTRL: Controllable Story Generation with External Knowledge Using Large-Scale Language Models

Authors: Peng Xu, Mostofa Patwary, Mohammad Shoeybi, Raul Puri, Pascale Fung, Anima Anandkumar, Bryan Catanzaro

Abstract: Existing pre-trained large language models have shown unparalleled generative capabilities. However, they are not controllable. In this paper, we propose MEGATRON-CNTRL, a novel framework that uses large-scale language models and adds control to text generation by incorporating an external knowledge base. Our framework consists of a keyword predictor, a knowledge retriever, a contextual knowledge… ▽ More Existing pre-trained large language models have shown unparalleled generative capabilities. However, they are not controllable. In this paper, we propose MEGATRON-CNTRL, a novel framework that uses large-scale language models and adds control to text generation by incorporating an external knowledge base. Our framework consists of a keyword predictor, a knowledge retriever, a contextual knowledge ranker, and a conditional text generator. As we do not have access to ground-truth supervision for the knowledge ranker, we make use of weak supervision from sentence embedding. The empirical results show that our model generates more fluent, consistent, and coherent stories with less repetition and higher diversity compared to prior work on the ROC story dataset. We showcase the controllability of our model by replacing the keywords used to generate stories and re-running the generation process. Human evaluation results show that 77.5% of these stories are successfully controlled by the new keywords. Furthermore, by scaling our model from 124 million to 8.3 billion parameters we demonstrate that larger models improve both the quality of generation (from 74.5% to 93.0% for consistency) and controllability (from 77.5% to 91.5%). △ Less

Submitted 2 October, 2020; originally announced October 2020.

Comments: Accepted in EMNLP 2020 main conference

Large Scale Multi-Actor Generative Dialog Modeling

Authors: Alex Boyd, Raul Puri, Mohammad Shoeybi, Mostofa Patwary, Bryan Catanzaro

Abstract: Non-goal oriented dialog agents (i.e. chatbots) aim to produce varying and engaging conversations with a user; however, they typically exhibit either inconsistent personality across conversations or the average personality of all users. This paper addresses these issues by controlling an agent's persona upon generation via conditioning on prior conversations of a target actor. In doing so, we are… ▽ More Non-goal oriented dialog agents (i.e. chatbots) aim to produce varying and engaging conversations with a user; however, they typically exhibit either inconsistent personality across conversations or the average personality of all users. This paper addresses these issues by controlling an agent's persona upon generation via conditioning on prior conversations of a target actor. In doing so, we are able to utilize more abstract patterns within a person's speech and better emulate them in generated responses. This work introduces the Generative Conversation Control model, an augmented and fine-tuned GPT-2 language model that conditions on past reference conversations to probabilistically model multi-turn conversations in the actor's persona. We introduce an accompanying data collection procedure to obtain 10.3M conversations from 6 months worth of Reddit comments. We demonstrate that scaling model sizes from 117M to 8.3B parameters yields an improvement from 23.14 to 13.14 perplexity on 1.7M held out Reddit conversations. Increasing model scale yielded similar improvements in human evaluations that measure preference of model samples to the held out target distribution in terms of realism (31% increased to 37% preference), style matching (37% to 42%), grammar and content quality (29% to 42%), and conversation coherency (32% to 40%). We find that conditionally modeling past conversations improves perplexity by 0.47 in automatic evaluations. Through human trials we identify positive trends between conditional modeling and style matching and outline steps to further improve persona control. △ Less

Submitted 12 May, 2020; originally announced May 2020.

Authentication, Access Control, Privacy, Threats and Trust Management Towards Securing Fog Computing Environments: A Review

Authors: Abdullah Al-Noman Patwary, Anmin Fu, Ranesh Kumar Naha, Sudheer Kumar Battula, Saurabh Garg, Md Anwarul Kaium Patwary, Erfan Aghasian

Abstract: Fog computing is an emerging computing paradigm that has come into consideration for the deployment of IoT applications amongst researchers and technology industries over the last few years. Fog is highly distributed and consists of a wide number of autonomous end devices, which contribute to the processing. However, the variety of devices offered across different users are not audited. Hence, the… ▽ More Fog computing is an emerging computing paradigm that has come into consideration for the deployment of IoT applications amongst researchers and technology industries over the last few years. Fog is highly distributed and consists of a wide number of autonomous end devices, which contribute to the processing. However, the variety of devices offered across different users are not audited. Hence, the security of Fog devices is a major concern in the Fog computing environment. Furthermore, mitigating and preventing those security measures is a research issue. Therefore, to provide the necessary security for Fog devices, we need to understand what the security concerns are with regards to Fog. All aspects of Fog security, which have not been covered by other literature works needs to be identified and need to be aggregate all issues in Fog security. It needs to be noted that computation devices consist of many ordinary users, and are not managed by any central entity or managing body. Therefore, trust and privacy is also a key challenge to gain market adoption for Fog. To provide the required trust and privacy, we need to also focus on authentication, threats and access control mechanisms as well as techniques in Fog computing. In this paper, we perform a survey and propose a taxonomy, which presents an overview of existing security concerns in the context of the Fog computing paradigm. We discuss the Blockchain-based solutions towards a secure Fog computing environment and presented various research challenges and directions for future research. △ Less

Submitted 29 February, 2020; originally announced March 2020.

Comments: 34 pages, 9 figures

Training Question Answering Models From Synthetic Data

Authors: Raul Puri, Ryan Spring, Mostofa Patwary, Mohammad Shoeybi, Bryan Catanzaro

Abstract: Question and answer generation is a data augmentation method that aims to improve question answering (QA) models given the limited amount of human labeled data. However, a considerable gap remains between synthetic and human-generated question-answer pairs. This work aims to narrow this gap by taking advantage of large language models and explores several factors such as model size, quality of pre… ▽ More Question and answer generation is a data augmentation method that aims to improve question answering (QA) models given the limited amount of human labeled data. However, a considerable gap remains between synthetic and human-generated question-answer pairs. This work aims to narrow this gap by taking advantage of large language models and explores several factors such as model size, quality of pretrained models, scale of data synthesized, and algorithmic choices. On the SQuAD1.1 question answering task, we achieve higher accuracy using solely synthetic questions and answers than when using the SQuAD1.1 training set questions alone. Removing access to real Wikipedia data, we synthesize questions and answers from a synthetic corpus generated by an 8.3 billion parameter GPT-2 model. With no access to human supervision and only access to other models, we are able to train state of the art question answering networks on entirely model-generated data that achieve 88.4 Exact Match (EM) and 93.9 F1 score on the SQuAD1.1 dev set. We further apply our methodology to SQuAD2.0 and show a 2.8 absolute gain on EM score compared to prior work using synthetic data. △ Less

Submitted 21 February, 2020; originally announced February 2020.

DisCo: Physics-Based Unsupervised Discovery of Coherent Structures in Spatiotemporal Systems

Authors: Adam Rupe, Nalini Kumar, Vladislav Epifanov, Karthik Kashinath, Oleksandr Pavlyk, Frank Schlimbach, Mostofa Patwary, Sergey Maidanov, Victor Lee, Prabhat, James P. Crutchfield

Abstract: Extracting actionable insight from complex unlabeled scientific data is an open challenge and key to unlocking data-driven discovery in science. Complementary and alternative to supervised machine learning approaches, unsupervised physics-based methods based on behavior-driven theories hold great promise. Due to computational limitations, practical application on real-world domain science problems… ▽ More Extracting actionable insight from complex unlabeled scientific data is an open challenge and key to unlocking data-driven discovery in science. Complementary and alternative to supervised machine learning approaches, unsupervised physics-based methods based on behavior-driven theories hold great promise. Due to computational limitations, practical application on real-world domain science problems has lagged far behind theoretical development. We present our first step towards bridging this divide - DisCo - a high-performance distributed workflow for the behavior-driven local causal state theory. DisCo provides a scalable unsupervised physics-based representation learning method that decomposes spatiotemporal systems into their structurally relevant components, which are captured by the latent local causal state variables. Complex spatiotemporal systems are generally highly structured and organize around a lower-dimensional skeleton of coherent structures, and in several firsts we demonstrate the efficacy of DisCo in capturing such structures from observational and simulated scientific data. To the best of our knowledge, DisCo is also the first application software developed entirely in Python to scale to over 1000 machine nodes, providing good performance along with ensuring domain scientists' productivity. We developed scalable, performant methods optimized for Intel many-core processors that will be upstreamed to open-source Python library packages. Our capstone experiment, using newly developed DisCo workflow and libraries, performs unsupervised spacetime segmentation analysis of CAM5.1 climate simulation data, processing an unprecedented 89.5 TB in 6.6 minutes end-to-end using 1024 Intel Haswell nodes on the Cori supercomputer obtaining 91% weak-scaling and 64% strong-scaling efficiency. △ Less

Submitted 25 September, 2019; originally announced September 2019.

The Potential Short- and Long-Term Disruptions and Transformative Impacts of 5G and Beyond Wireless Networks: Lessons Learnt from the Development of a 5G Testbed Environment

Authors: Mohmammad N. Patwary, Syed Junaid Nawaz, Md. Abdur Rahman, Shree Krishna Sharma, Md Mamunur Rashid, Stuart J. Barnes

Abstract: The anticipated deployment cost of 5G communication networks in the UK is predicted to be in between £30bn- £50bn, whereas the current annual capital expenditure of the mobile network operators (MNOs) is £2.5bn. This prospect has vastly impacted and has become one of the major delaying factors for building the 5G physical infrastructure, whereas other areas of 5G developments are progressing at th… ▽ More The anticipated deployment cost of 5G communication networks in the UK is predicted to be in between £30bn- £50bn, whereas the current annual capital expenditure of the mobile network operators (MNOs) is £2.5bn. This prospect has vastly impacted and has become one of the major delaying factors for building the 5G physical infrastructure, whereas other areas of 5G developments are progressing at their speed. In this paper, an extensive study is conducted to explore the possibilities of reducing the 5G deployment cost and developing business models. This study suggests that the use of existing public infrastructure has a great potential to contribute to a reduction of about 40% to 60% in the anticipated cost. Also, the recent Ofcom initiative of location-based licensing of radio spectrum is reviewed. Our study suggests that simplification of infrastructure and spectrum will encourage the exponential growth of scenario-specific cellular networks and will potentially disrupt the current business models of telecommunication stakeholders -- specifically MNOs and TowerCos. Moreover, the dense network connectivity will encourage extensive data harvesting as a business opportunity and function within small and medium-sized enterprises as well as within large social networks. Consequently, the rise of new infrastructures and spectrum stakeholders is anticipated. This will fuel the development of a 5G data exchange ecosystem where data transactions are deemed to be high-value business commodities. Also, a data economy-oriented business model is proposed. The study found that with the potential commodification of data and data transactions along with the low-cost physical infrastructure and spectrum, the 5G network will introduce significant disruption in the Telco business ecosystem. △ Less

Submitted 31 May, 2020; v1 submitted 23 September, 2019; originally announced September 2019.

Comments: 22 pages, 9 figures, 11 tables

Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism

Authors: Mohammad Shoeybi, Mostofa Patwary, Raul Puri, Patrick LeGresley, Jared Casper, Bryan Catanzaro

Abstract: Recent work in language modeling demonstrates that training large transformer models advances the state of the art in Natural Language Processing applications. However, very large models can be quite difficult to train due to memory constraints. In this work, we present our techniques for training very large transformer models and implement a simple, efficient intra-layer model parallel approach t… ▽ More Recent work in language modeling demonstrates that training large transformer models advances the state of the art in Natural Language Processing applications. However, very large models can be quite difficult to train due to memory constraints. In this work, we present our techniques for training very large transformer models and implement a simple, efficient intra-layer model parallel approach that enables training transformer models with billions of parameters. Our approach does not require a new compiler or library changes, is orthogonal and complimentary to pipeline model parallelism, and can be fully implemented with the insertion of a few communication operations in native PyTorch. We illustrate this approach by converging transformer based models up to 8.3 billion parameters using 512 GPUs. We sustain 15.1 PetaFLOPs across the entire application with 76% scaling efficiency when compared to a strong single GPU baseline that sustains 39 TeraFLOPs, which is 30% of peak FLOPs. To demonstrate that large language models can further advance the state of the art (SOTA), we train an 8.3 billion parameter transformer language model similar to GPT-2 and a 3.9 billion parameter model similar to BERT. We show that careful attention to the placement of layer normalization in BERT-like models is critical to achieving increased performance as the model size grows. Using the GPT-2 model we achieve SOTA results on the WikiText103 (10.8 compared to SOTA perplexity of 15.8) and LAMBADA (66.5% compared to SOTA accuracy of 63.2%) datasets. Our BERT model achieves SOTA results on the RACE dataset (90.9% compared to SOTA accuracy of 89.4%). △ Less

Submitted 13 March, 2020; v1 submitted 17 September, 2019; originally announced September 2019.

Coloring Big Graphs with AlphaGoZero

Authors: Jiayi Huang, Mostofa Patwary, Gregory Diamos

Abstract: We show that recent innovations in deep reinforcement learning can effectively color very large graphs -- a well-known NP-hard problem with clear commercial applications. Because the Monte Carlo Tree Search with Upper Confidence Bound algorithm used in AlphaGoZero can improve the performance of a given heuristic, our approach allows deep neural networks trained using high performance computing (HP… ▽ More We show that recent innovations in deep reinforcement learning can effectively color very large graphs -- a well-known NP-hard problem with clear commercial applications. Because the Monte Carlo Tree Search with Upper Confidence Bound algorithm used in AlphaGoZero can improve the performance of a given heuristic, our approach allows deep neural networks trained using high performance computing (HPC) technologies to transform computation into improved heuristics with zero prior knowledge. Key to our approach is the introduction of a novel deep neural network architecture (FastColorNet) that has access to the full graph context and requires $O(V)$ time and space to color a graph with $V$ vertices, which enables scaling to very large graphs that arise in real applications like parallel computing, compilers, numerical solvers, and design automation, among others. As a result, we are able to learn new state of the art heuristics for graph coloring. △ Less

Submitted 8 November, 2019; v1 submitted 26 February, 2019; originally announced February 2019.

Window-based Streaming Graph Partitioning Algorithm

Authors: Md Anwarul kaium Patwary, Saurabh Garg, Byeong Kang

Abstract: In the recent years, the scale of graph datasets has increased to such a degree that a single machine is not capable of efficiently processing large graphs. Thereby, efficient graph partitioning is necessary for those large graph applications. Traditional graph partitioning generally loads the whole graph data into the memory before performing partitioning; this is not only a time consuming task b… ▽ More In the recent years, the scale of graph datasets has increased to such a degree that a single machine is not capable of efficiently processing large graphs. Thereby, efficient graph partitioning is necessary for those large graph applications. Traditional graph partitioning generally loads the whole graph data into the memory before performing partitioning; this is not only a time consuming task but it also creates memory bottlenecks. These issues of memory limitation and enormous time complexity can be resolved using stream-based graph partitioning. A streaming graph partitioning algorithm reads vertices once and assigns that vertex to a partition accordingly. This is also called an one-pass algorithm. This paper proposes an efficient window-based streaming graph partitioning algorithm called WStream. The WStream algorithm is an edge-cut partitioning algorithm, which distributes a vertex among the partitions. Our results suggest that the WStream algorithm is able to partition large graph data efficiently while keeping the load balanced across different partitions, and communication to a minimum. Evaluation results with real workloads also prove the effectiveness of our proposed algorithm, and it achieves a significant reduction in load imbalance and edge-cut with different ranges of dataset. △ Less

Submitted 4 February, 2019; originally announced February 2019.

Language Modeling at Scale

Authors: Mostofa Patwary, Milind Chabbi, Heewoo Jun, Jiaji Huang, Gregory Diamos, Kenneth Church

Abstract: We show how Zipf's Law can be used to scale up language modeling (LM) to take advantage of more training data and more GPUs. LM plays a key role in many important natural language applications such as speech recognition and machine translation. Scaling up LM is important since it is widely accepted by the community that there is no data like more data. Eventually, we would like to train on terabyt… ▽ More We show how Zipf's Law can be used to scale up language modeling (LM) to take advantage of more training data and more GPUs. LM plays a key role in many important natural language applications such as speech recognition and machine translation. Scaling up LM is important since it is widely accepted by the community that there is no data like more data. Eventually, we would like to train on terabytes (TBs) of text (trillions of words). Modern training methods are far from this goal, because of various bottlenecks, especially memory (within GPUs) and communication (across GPUs). This paper shows how Zipf's Law can address these bottlenecks by grouping parameters for common words and character sequences, because $U \ll N$, where $U$ is the number of unique words (types) and $N$ is the size of the training set (tokens). For a local batch size $K$ with $G$ GPUs and a $D$-dimension embedding matrix, we reduce the original per-GPU memory and communication asymptotic complexity from $Θ(GKD)$ to $Θ(GK + UD)$. Empirically, we find $U \propto (GK)^{0.64}$ on four publicly available large datasets. When we scale up the number of GPUs to 64, a factor of 8, training time speeds up by factors up to 6.7$\times$ (for character LMs) and 6.3$\times$ (for word LMs) with negligible loss of accuracy. Our weak scaling on 192 GPUs on the Tieba dataset shows a 35\% improvement in LM prediction accuracy by training on 93 GB of data (2.5$\times$ larger than publicly available SOTA dataset), but taking only 1.25$\times$ increase in training time, compared to 3 GB of the same dataset running on 6 GPUs. △ Less

Submitted 23 October, 2018; originally announced October 2018.

Deep Learning Scaling is Predictable, Empirically

Authors: Joel Hestness, Sharan Narang, Newsha Ardalani, Gregory Diamos, Heewoo Jun, Hassan Kianinejad, Md. Mostofa Ali Patwary, Yang Yang, Yanqi Zhou

Abstract: Deep learning (DL) creates impactful advances following a virtuous recipe: model architecture search, creating large training data sets, and scaling computation. It is widely believed that growing training sets and models should improve accuracy and result in better products. As DL application domains grow, we would like a deeper understanding of the relationships between training set size, comput… ▽ More Deep learning (DL) creates impactful advances following a virtuous recipe: model architecture search, creating large training data sets, and scaling computation. It is widely believed that growing training sets and models should improve accuracy and result in better products. As DL application domains grow, we would like a deeper understanding of the relationships between training set size, computational scale, and model accuracy improvements to advance the state-of-the-art. This paper presents a large scale empirical characterization of generalization error and model size growth as training sets grow. We introduce a methodology for this measurement and test four machine learning domains: machine translation, language modeling, image processing, and speech recognition. Our empirical results show power-law generalization error scaling across a breadth of factors, resulting in power-law exponents---the "steepness" of the learning curve---yet to be explained by theoretical work. Further, model improvements only shift the error but do not appear to affect the power-law exponent. We also show that model size scales sublinearly with data size. These scaling relationships have significant implications on deep learning research, practice, and systems. They can assist model debugging, setting accuracy targets, and decisions about data set growth. They can also guide computing system design and underscore the importance of continued computational scaling. △ Less

Submitted 1 December, 2017; originally announced December 2017.

Comments: 19 pages, 11 figures

Galactos: Computing the Anisotropic 3-Point Correlation Function for 2 Billion Galaxies

Authors: Brian Friesen, Md. Mostofa Ali Patwary, Brian Austin, Nadathur Satish, Zachary Slepian, Narayanan Sundaram, Deborah Bard, Daniel J Eisenstein, Jack Deslippe, Pradeep Dubey, Prabhat

Abstract: The nature of dark energy and the complete theory of gravity are two central questions currently facing cosmology. A vital tool for addressing them is the 3-point correlation function (3PCF), which probes deviations from a spatially random distribution of galaxies. However, the 3PCF's formidable computational expense has prevented its application to astronomical surveys comprising millions to bill… ▽ More The nature of dark energy and the complete theory of gravity are two central questions currently facing cosmology. A vital tool for addressing them is the 3-point correlation function (3PCF), which probes deviations from a spatially random distribution of galaxies. However, the 3PCF's formidable computational expense has prevented its application to astronomical surveys comprising millions to billions of galaxies. We present Galactos, a high-performance implementation of a novel, O(N^2) algorithm that uses a load-balanced k-d tree and spherical harmonic expansions to compute the anisotropic 3PCF. Our implementation is optimized for the Intel Xeon Phi architecture, exploiting SIMD parallelism, instruction and thread concurrency, and significant L1 and L2 cache reuse, reaching 39% of peak performance on a single node. Galactos scales to the full Cori system, achieving 9.8PF (peak) and 5.06PF (sustained) across 9636 nodes, making the 3PCF easily computable for all galaxies in the observable universe. △ Less

Submitted 31 August, 2017; originally announced September 2017.

Comments: 11 pages, 7 figures, accepted to SuperComputing 2017

Deep Learning at 15PF: Supervised and Semi-Supervised Classification for Scientific Data

Authors: Thorsten Kurth, Jian Zhang, Nadathur Satish, Ioannis Mitliagkas, Evan Racah, Mostofa Ali Patwary, Tareq Malas, Narayanan Sundaram, Wahid Bhimji, Mikhail Smorkalov, Jack Deslippe, Mikhail Shiryaev, Srinivas Sridharan, Prabhat, Pradeep Dubey

Abstract: This paper presents the first, 15-PetaFLOP Deep Learning system for solving scientific pattern classification problems on contemporary HPC architectures. We develop supervised convolutional architectures for discriminating signals in high-energy physics data as well as semi-supervised architectures for localizing and classifying extreme weather in climate data. Our Intelcaffe-based implementation… ▽ More This paper presents the first, 15-PetaFLOP Deep Learning system for solving scientific pattern classification problems on contemporary HPC architectures. We develop supervised convolutional architectures for discriminating signals in high-energy physics data as well as semi-supervised architectures for localizing and classifying extreme weather in climate data. Our Intelcaffe-based implementation obtains $\sim$2TFLOP/s on a single Cori Phase-II Xeon-Phi node. We use a hybrid strategy employing synchronous node-groups, while using asynchronous communication across groups. We use this strategy to scale training of a single model to $\sim$9600 Xeon-Phi nodes; obtaining peak performance of 11.73-15.07 PFLOP/s and sustained performance of 11.41-13.27 PFLOP/s. At scale, our HEP architecture produces state-of-the-art classification accuracy on a dataset with 10M images, exceeding that achieved by selections on high-level physics-motivated features. Our semi-supervised architecture successfully extracts weather patterns in a 15TB climate dataset. Our results demonstrate that Deep Learning can be optimized and scaled effectively on many-core, HPC systems. △ Less

Submitted 17 August, 2017; originally announced August 2017.

Comments: 12 pages, 9 figures

PANDA: Extreme Scale Parallel K-Nearest Neighbor on Distributed Architectures

Authors: Md. Mostofa Ali Patwary, Nadathur Rajagopalan Satish, Narayanan Sundaram, Jialin Liu, Peter Sadowski, Evan Racah, Suren Byna, Craig Tull, Wahid Bhimji, Prabhat, Pradeep Dubey

Abstract: Computing $k$-Nearest Neighbors (KNN) is one of the core kernels used in many machine learning, data mining and scientific computing applications. Although kd-tree based $O(\log n)$ algorithms have been proposed for computing KNN, due to its inherent sequentiality, linear algorithms are being used in practice. This limits the applicability of such methods to millions of data points, with limited s… ▽ More Computing $k$-Nearest Neighbors (KNN) is one of the core kernels used in many machine learning, data mining and scientific computing applications. Although kd-tree based $O(\log n)$ algorithms have been proposed for computing KNN, due to its inherent sequentiality, linear algorithms are being used in practice. This limits the applicability of such methods to millions of data points, with limited scalability for Big Data analytics challenges in the scientific domain. In this paper, we present parallel and highly optimized kd-tree based KNN algorithms (both construction and querying) suitable for distributed architectures. Our algorithm includes novel approaches for pruning search space and improving load balancing and partitioning among nodes and threads. Using TB-sized datasets from three science applications: astrophysics, plasma physics, and particle physics, we show that our implementation can construct kd-tree of 189 billion particles in 48 seconds on utilizing $\sim$50,000 cores. We also demonstrate computation of KNN of 19 billion queries in 12 seconds. We demonstrate almost linear speedup both for shared and distributed memory computers. Our algorithms outperforms earlier implementations by more than order of magnitude; thereby radically improving the applicability of our implementation to state-of-the-art Big Data analytics problems. In addition, we showcase performance and scalability on the recently released Intel Xeon Phi processor showing that our algorithm scales well even on massively parallel architectures. △ Less

Submitted 27 July, 2016; originally announced July 2016.

Comments: 11 pages in PANDA: Extreme Scale Parallel K-Nearest Neighbor on Distributed Architectures, Md. Mostofa Ali Patwary et.al., IEEE International Parallel and Distributed Processing Symposium (IPDPS), 2016

A New Parallel Algorithm for Two-Pass Connected Component Labeling

Authors: Siddharth Gupta, Diana Palsetia, Md. Mostofa Ali Patwary, Ankit Agrawal, Alok Choudhary

Abstract: Connected Component Labeling (CCL) is an important step in pattern recognition and image processing. It assigns labels to the pixels such that adjacent pixels sharing the same features are assigned the same label. Typically, CCL requires several passes over the data. We focus on two-pass technique where each pixel is given a provisional label in the first pass whereas an actual label is assigned i… ▽ More Connected Component Labeling (CCL) is an important step in pattern recognition and image processing. It assigns labels to the pixels such that adjacent pixels sharing the same features are assigned the same label. Typically, CCL requires several passes over the data. We focus on two-pass technique where each pixel is given a provisional label in the first pass whereas an actual label is assigned in the second pass. We present a scalable parallel two-pass CCL algorithm, called PAREMSP, which employs a scan strategy and the best union-find technique called REMSP, which uses REM's algorithm for storing label equivalence information of pixels in a 2-D image. In the first pass, we divide the image among threads and each thread runs the scan phase along with REMSP simultaneously. In the second phase, we assign the final labels to the pixels. As REMSP is easily parallelizable, we use the parallel version of REMSP for merging the pixels on the boundary. Our experiments show the scalability of PAREMSP achieving speedups up to $20.1$ using $24$ cores on shared memory architecture using OpenMP for an image of size $465.20$ MB. We find that our proposed parallel algorithm achieves linear scaling for a large resolution fixed problem size as the number of processing elements are increased. Additionally, the parallel algorithm does not make use of any hardware specific routines, and thus is highly portable. △ Less

Submitted 20 June, 2016; originally announced June 2016.

Comments: Parallel & Distributed Processing Symposium Workshops (IPDPSW), 2014

GraphMat: High performance graph analytics made productive

Authors: Narayanan Sundaram, Nadathur Rajagopalan Satish, Md Mostofa Ali Patwary, Subramanya R Dulloor, Satya Gautam Vadlamudi, Dipankar Das, Pradeep Dubey

Abstract: Given the growing importance of large-scale graph analytics, there is a need to improve the performance of graph analysis frameworks without compromising on productivity. GraphMat is our solution to bridge this gap between a user-friendly graph analytics framework and native, hand-optimized code. GraphMat functions by taking vertex programs and mapping them to high performance sparse matrix operat… ▽ More Given the growing importance of large-scale graph analytics, there is a need to improve the performance of graph analysis frameworks without compromising on productivity. GraphMat is our solution to bridge this gap between a user-friendly graph analytics framework and native, hand-optimized code. GraphMat functions by taking vertex programs and mapping them to high performance sparse matrix operations in the backend. We get the productivity benefits of a vertex programming framework without sacrificing performance. GraphMat is in C++, and we have been able to write a diverse set of graph algorithms in this framework with the same effort compared to other vertex programming frameworks. GraphMat performs 1.2-7X faster than high performance frameworks such as GraphLab, CombBLAS and Galois. It achieves better multicore scalability (13-15X on 24 cores) than other frameworks and is 1.2X off native, hand-optimized code on a variety of different graph algorithms. Since GraphMat performance depends mainly on a few scalable and well-understood sparse matrix operations, GraphMatcan naturally benefit from the trend of increasing parallelism on future hardware. △ Less

Submitted 24 March, 2015; originally announced March 2015.

Fast Algorithms for the Maximum Clique Problem on Massive Graphs with Applications to Overlapping Community Detection

Authors: Bharath Pattabiraman, Md. Mostofa Ali Patwary, Assefaw H. Gebremedhin, Wei-keng Liao, Alok Choudhary

Abstract: The maximum clique problem is a well known NP-Hard problem with applications in data mining, network analysis, information retrieval and many other areas related to the World Wide Web. There exist several algorithms for the problem with acceptable runtimes for certain classes of graphs, but many of them are infeasible for massive graphs. We present a new exact algorithm that employs novel pruning… ▽ More The maximum clique problem is a well known NP-Hard problem with applications in data mining, network analysis, information retrieval and many other areas related to the World Wide Web. There exist several algorithms for the problem with acceptable runtimes for certain classes of graphs, but many of them are infeasible for massive graphs. We present a new exact algorithm that employs novel pruning techniques and is able to find maximum cliques in very large, sparse graphs quickly. Extensive experiments on different kinds of synthetic and real-world graphs show that our new algorithm can be orders of magnitude faster than existing algorithms. We also present a heuristic that runs orders of magnitude faster than the exact algorithm while providing optimal or near-optimal solutions. We illustrate a simple application of the algorithms in developing methods for detection of overlapping communities in networks. △ Less

Submitted 26 November, 2014; originally announced November 2014.

Comments: 28 pages, 7 figures, 10 tables, 2 algorithms. arXiv admin note: substantial text overlap with arXiv:1209.5818

Journal ref: Internet Mathematics 2014, Special Issue (WAW'13)

Parallel Maximum Clique Algorithms with Applications to Network Analysis and Storage

Authors: Ryan A. Rossi, David F. Gleich, Assefaw H. Gebremedhin, Md. Mostofa Ali Patwary

Abstract: We propose a fast, parallel maximum clique algorithm for large sparse graphs that is designed to exploit characteristics of social and information networks. The method exhibits a roughly linear runtime scaling over real-world networks ranging from 1000 to 100 million nodes. In a test on a social network with 1.8 billion edges, the algorithm finds the largest clique in about 20 minutes. Our method… ▽ More We propose a fast, parallel maximum clique algorithm for large sparse graphs that is designed to exploit characteristics of social and information networks. The method exhibits a roughly linear runtime scaling over real-world networks ranging from 1000 to 100 million nodes. In a test on a social network with 1.8 billion edges, the algorithm finds the largest clique in about 20 minutes. Our method employs a branch and bound strategy with novel and aggressive pruning techniques. For instance, we use the core number of a vertex in combination with a good heuristic clique finder to efficiently remove the vast majority of the search space. In addition, we parallelize the exploration of the search tree. During the search, processes immediately communicate changes to upper and lower bounds on the size of maximum clique, which occasionally results in a super-linear speedup because vertices with large search spaces can be pruned by other processes. We apply the algorithm to two problems: to compute temporal strong components and to compress graphs. △ Less

Submitted 25 December, 2013; v1 submitted 25 February, 2013; originally announced February 2013.

Comments: 11 pages

MSC Class: 05C69 ACM Class: G.2.2

What if CLIQUE were fast? Maximum Cliques in Information Networks and Strong Components in Temporal Networks

Authors: Ryan A. Rossi, David F. Gleich, Assefaw H. Gebremedhin, Md. Mostofa Ali Patwary

Abstract: Exact maximum clique finders have progressed to the point where we can investigate cliques in million-node social and information networks, as well as find strongly connected components in temporal networks. We use one such finder to study a large collection of modern networks emanating from biological, social, and technological domains. We show inter-relationships between maximum cliques and seve… ▽ More Exact maximum clique finders have progressed to the point where we can investigate cliques in million-node social and information networks, as well as find strongly connected components in temporal networks. We use one such finder to study a large collection of modern networks emanating from biological, social, and technological domains. We show inter-relationships between maximum cliques and several other common network properties, including network density, maximum core, and number of triangles. In temporal networks, we find that the largest temporal strong components have around 20-30% of the vertices of the entire network. These components represent groups of highly communicative individuals. In addition, we discuss and improve the performance and utility of the maximum clique finder itself. △ Less

Submitted 30 October, 2012; v1 submitted 22 October, 2012; originally announced October 2012.

MSC Class: 05C69; 05C85; 91D30 ACM Class: G.2.2; H.2.8

Fast Algorithms for the Maximum Clique Problem on Massive Sparse Graphs

Authors: Bharath Pattabiraman, Md. Mostofa Ali Patwary, Assefaw H. Gebremedhin, Wei-keng Liao, Alok Choudhary

Abstract: The maximum clique problem is a well known NP-Hard problem with applications in data mining, network analysis, informatics, and many other areas. Although there exist several algorithms with acceptable runtimes for certain classes of graphs, many of them are infeasible for massive graphs. We present a new exact algorithm that employs novel pruning techniques to very quickly find maximum cliques in… ▽ More The maximum clique problem is a well known NP-Hard problem with applications in data mining, network analysis, informatics, and many other areas. Although there exist several algorithms with acceptable runtimes for certain classes of graphs, many of them are infeasible for massive graphs. We present a new exact algorithm that employs novel pruning techniques to very quickly find maximum cliques in large sparse graphs. Extensive experiments on several types of synthetic and real-world graphs show that our new algorithm is up to several orders of magnitude faster than existing algorithms for most instances. We also present a heuristic variant that runs orders of magnitude faster than the exact algorithm, while providing optimal or near-optimal solutions. △ Less

Submitted 14 November, 2012; v1 submitted 25 September, 2012; originally announced September 2012.

Comments: 15 pages (including 2-page appendix), 5 tables, 4 figures

ACM Class: G.2.2