Related papers: SkipPipe: Partial and Reordered Pipelining Framewo…

Large language models (LLMs) rely on pretraining on massive and heterogeneous corpora, where training data composition has a decisive impact on training efficiency and downstream generalization under realistic compute and data budget…

The rapid adoption of large language models and multimodal foundation models has made multimodal data preparation pipelines critical AI infrastructure. These pipelines interleave CPU-heavy preprocessing with accelerator-backed (GPU/NPU/TPU)…

Recent advances in large language models (LLMs) transform how machine learning (ML) pipelines are developed and evaluated. LLMs enable a new type of workload, agentic pipeline search, in which autonomous or semi-autonomous agents generate,…

Table Question Answering (TQA) aims to answer natural language questions over structured tables. Large Language Models (LLMs) enable promising solutions to this problem, with operator-centric solutions that generate table manipulation…

Data parallelism has emerged as a necessary technique to accelerate the training of deep neural networks (DNN). In a typical data parallelism approach, the local workers push the latest updates of all the parameters to the parameter server…

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…

Many deep learning applications benefit from using large models with billions of parameters. Training these models is notoriously expensive due to the need for specialized HPC clusters. In this work, we consider alternative setups for…

To alleviate the shortage of computing power faced by clients in training deep neural networks (DNNs) using federated learning (FL), we leverage the edge computing and split learning to propose a model-splitting allowed FL (SFL) framework,…

Scaling long-context capabilities is crucial for Multimodal Large Language Models (MLLMs). However, real-world multimodal datasets are extremely heterogeneous. Existing training frameworks predominantly rely on static parallelism…

In large-scale LLM pre-training systems with 100k+ GPUs, failures become the norm rather than the exception, and restart costs can dominate wall-clock training time. However, existing fault-tolerance mechanisms are largely unprepared for…

DNN learning jobs are common in today's clusters due to the advances in AI driven services such as machine translation and image recognition. The most critical phase of these jobs for model performance and learning cost is the tuning of…

In the machine learning system, the hybrid model parallelism combining tensor parallelism (TP) and pipeline parallelism (PP) has become the dominant solution for distributed training of Large Language Models~(LLMs) and Multimodal LLMs…

Pipeline parallelism is a key technique for scaling large-model training, but modern workloads exhibit runtime variability in computation and communication. Existing pipeline systems typically consume static, profiled, or adaptively…

Training large language models (LLMs) is a computationally intensive task, which is typically conducted in data centers with homogeneous high-performance GPUs. In this paper, we explore an alternative approach by deploying training…

Efficient scheduling is crucial for interactive Large Language Model (LLM) applications, where low request completion time directly impacts user engagement. Size-based scheduling algorithms like Shortest Remaining Process Time (SRPT) aim to…

Pipeline parallelism enables efficient training of Large Language Models (LLMs) on large-scale distributed accelerator clusters. Yet, pipeline bubbles during startup and tear-down reduce the utilization of accelerators. Although efficient…

With the rapid advancement of large language models (LLMs), efficiently serving LLM inference under limited GPU resources has become a critical challenge. Recently, an increasing number of studies have explored applying serverless computing…

In the realm of Large Language Model (LLM) inference, the inherent structure of transformer models coupled with the multi-GPU tensor parallelism strategy leads to a sequential execution of computation and communication. This results in…

Large Language Models (LLMs) have achieved remarkable success through imitation learning on vast text corpora, but this paradigm creates a training-generation gap and limits robust reasoning. Reinforcement learning (RL) offers a more…

End-to-end training with full-depth backpropagation remains the dominant paradigm for optimizing deep neural networks, but its efficiency deteriorates as models grow deeper. Since every block must be executed and differentiated under a…