Menghan Yu

ByteDance Seed, Jiaze Chen, Tiantian Fan et al. · bytedance

We introduce Seed1.5-Thinking, capable of reasoning through thinking before responding, resulting in improved performance on a wide range of benchmarks. Seed1.5-Thinking achieves 86.7 on AIME 2024, 55.0 on Codeforces and 77.3 on GPQA, demonstrating excellent reasoning abilities in STEM and coding. Beyond reasoning tasks, the method demonstrates notable generalization across diverse domains. For instance, it surpasses DeepSeek R1 by 8% in win rate on non-reasoning tasks, indicating its broader applicability. Compared to other state-of-the-art reasoning models, Seed1.5-Thinking is a Mixture-of-Experts (MoE) model with a relatively small size, featuring 20B activated and 200B total parameters. As part of our effort to assess generalized reasoning, we develop two internal benchmarks, BeyondAIME and Codeforces, both of which will be publicly released to support future research. Model trial link: https://www.volcengine.com/experience/ark.

Borui Wan, Mingji Han, Yiyao Sheng et al.

Checkpointing to preserve training states is crucial during the development of Large Foundation Models (LFMs), for training resumption upon various failures or changes in GPU resources and parallelism configurations. In addition, saved checkpoints are dispatched to evaluation tasks or transferred across different training stages (e.g., from pre-training to post-training). All these scenarios require resharding distributed checkpoints from one parallelism to another. In production environments, different LFMs are trained with various frameworks and storage backends, depending on model sizes and training scales. A high-performance checkpointing system is needed to enable efficient checkpoint management at scale throughout the lifecycle of LFM development. We introduce ByteCheckpoint, an industrial-grade checkpointing system for large-scale LFM training. ByteCheckpoint features: a parallelism-agnostic checkpoint representation that enables efficient load-time checkpoint resharding; a generic checkpoint saving/loading workflow to accommodate multiple training frameworks and support different storage backends; full-stack optimizations to ensure high I/O efficiency and scalability; a suite of monitoring tools to streamline large-scale performance analysis and bottleneck detection. Compared to existing open-source checkpointing systems [52, 58], ByteCheckpoint significantly reduces runtime checkpoint stalls, achieving an average reduction of 54.20x. For saving and loading times, ByteCheckpoint achieves improvements of up to 9.96x and 8.80x, respectively.

Menghan Yu, Sourabh Kulhare, Courosh Mehanian et al.

Acquiring large quantities of data and annotations is known to be effective for developing high-performing deep learning models, but is difficult and expensive to do in the healthcare context. Adding synthetic training data using generative models offers a low-cost method to deal effectively with the data scarcity challenge, and can also address data imbalance and patient privacy issues. In this study, we propose a comprehensive framework that fits seamlessly into model development workflows for medical image analysis. We demonstrate, with datasets of varying size, (i) the benefits of generative models as a data augmentation method; (ii) how adversarial methods can protect patient privacy via data substitution; (iii) novel performance metrics for these use cases by testing models on real holdout data. We show that training with both synthetic and real data outperforms training with real data alone, and that models trained solely with synthetic data approach their real-only counterparts. Code is available at https://github.com/Global-Health-Labs/US-DCGAN.

Jinkun Lin, Ziheng Jiang, Zuquan Song et al.

Large language model (LLM) training is one of the most demanding distributed computations today, often requiring thousands of GPUs with frequent synchronization across machines. Such a workload pattern makes it susceptible to stragglers, where the training can be stalled by few slow workers. At ByteDance we find stragglers are not trivially always caused by hardware failures, but can arise from multiple complex factors. This work aims to present a comprehensive study on the straggler issues in LLM training, using a five-month trace collected from our ByteDance LLM training cluster. The core methodology is what-if analysis that simulates the scenario without any stragglers and contrasts with the actual case. We use this method to study the following questions: (1) how often do stragglers affect training jobs, and what effect do they have on job performance; (2) do stragglers exhibit temporal or spatial patterns; and (3) what are the potential root causes for stragglers?

Borui Wan, Gaohong Liu, Zuquan Song et al.

The training scale of large language models (LLMs) has reached tens of thousands of GPUs and is still continuously expanding, enabling faster learning of larger models. Accompanying the expansion of the resource scale is the prevalence of failures (CUDA error, NaN values, job hang, etc.), which poses significant challenges to training stability. Any large-scale LLM training infrastructure should strive for minimal training interruption, efficient fault diagnosis, and effective failure tolerance to enable highly efficient continuous training. This paper presents ByteRobust, a large-scale GPU infrastructure management system tailored for robust and stable training of LLMs. It exploits the uniqueness of LLM training process and gives top priorities to detecting and recovering failures in a routine manner. Leveraging parallelisms and characteristics of LLM training, ByteRobust enables high-capacity fault tolerance, prompt fault demarcation, and localization with an effective data-driven approach, comprehensively ensuring continuous and efficient training of LLM tasks. ByteRobust is deployed on a production GPU platform and achieves 97% ETTR for a three-month training job on 9,600 GPUs.

Ziyi Zhang, Ziheng Jiang, Chengquan Jiang et al.

Low-latency decoding for large language models (LLMs) is crucial for applications like chatbots and code assistants, yet generating long outputs remains slow in single-query settings. Prior work on speculative decoding (which combines a small draft model with a larger target model) and tensor parallelism has each accelerated decoding. However, conventional approaches fail to apply both simultaneously due to imbalanced compute requirements (between draft and target models), KV-cache inconsistencies, and communication overheads under small-batch tensor-parallelism. This paper introduces SwiftSpec, a system that targets ultra-low latency for LLM decoding. SwiftSpec redesigns the speculative decoding pipeline in an asynchronous and disaggregated manner, so that each component can be scaled flexibly and remove draft overhead from the critical path. To realize this design, SwiftSpec proposes parallel tree generation, tree-aware KV cache management, and fused, latency-optimized kernels to overcome the challenges listed above. Across 5 model families and 6 datasets, SwiftSpec achieves an average of 1.75x speedup over state-of-the-art speculative decoding systems and, as a highlight, serves Llama3-70B at 348 tokens/s on 8 Nvidia Hopper GPUs, making it the fastest known system for low-latency LLM serving at this scale.

Rui Li, Xiaoyun Zhi, Jinxin Chi et al.

Large Language Models (LLMs) have become a cornerstone of modern AI, driving breakthroughs in natural language processing and expanding into multimodal jobs involving images, audio, and video. As with most computational software, it is important to distinguish between ordinary runtime performance and startup overhead. Prior research has focused on runtime performance: improving training efficiency and stability. This work focuses instead on the increasingly critical issue of startup overhead in training: the delay before training jobs begin execution. Startup overhead is particularly important in large, industrial-scale LLMs, where failures occur more frequently and multiple teams operate in iterative update-debug cycles. In one of our training clusters, more than 3.5% of GPU time is wasted due to startup overhead alone. In this work, we present the first in-depth characterization of LLM training startup overhead based on real production data. We analyze the components of startup cost, quantify its direct impact, and examine how it scales with job size. These insights motivate the design of Bootseer, a system-level optimization framework that addresses three primary startup bottlenecks: (a) container image loading, (b) runtime dependency installation, and (c) model checkpoint resumption. To mitigate these bottlenecks, Bootseer introduces three techniques: (a) hot block record-and-prefetch, (b) dependency snapshotting, and (c) striped HDFS-FUSE. Bootseer has been deployed in a production environment and evaluated on real LLM training workloads, demonstrating a 50% reduction in startup overhead.