4 papers
Sandwich: Joint Configuration Search and Hot-Switching for Efficient CPU LLM Serving
Juntao Zhao, Jiuru Li, Chuan Wu
CPUs are critical for LLM serving due to their availability, cost efficiency, and edge applicability. However, efficient CPU serving is hindered by conflicting prefill/decode resou…
Efficient LLM Serving on Hybrid Real-time and Best-effort Requests
Wan Borui, Zhao Juntao, Jiang Chenyu +2
Recent breakthroughs in large Language Models (LLMs) have enabled various generative tasks on a single model. Real-world services (e.g., OpenAI's ChatGPT [27]) powered by an LLM of…
MegaScale-Data: Scaling Dataloader for Multisource Large Foundation Model Training
Juntao Zhao, Qi Lu, Wei Jia +13
Modern frameworks for training large foundation models (LFMs) employ dataloaders in a data-parallel manner, with each loader processing a disjoint subset of training data. When pre…
QSpec: Speculative Decoding with Complementary Quantization Schemes
Juntao Zhao, Wenhao Lu, Sheng Wang +2
Quantization is widely adopted to accelerate inference and reduce memory consumption in large language models (LLMs). While activation-weight joint quantization enables efficient l…