{FlashNeuron}:{SSD-Enabled}{Large-Batch} Training of Very Deep Neural Networks J Bae, J Lee, Y Jin, S Son, S Kim, H Jang, TJ Ham, JW Lee
19th USENIX Conference on File and Storage Technologies (FAST 21), 387-401, 2021
55 2021 : Increasing GPU Utilization during Generative Inference for Higher ThroughputY Jin, CF Wu, D Brooks, GY Wei
Advances in Neural Information Processing Systems 36, 2024
48 2024 Layerweaver: Maximizing resource utilization of neural processing units via layer-wise scheduling YH Oh, S Kim, Y Jin, S Son, J Bae, J Lee, Y Park, DU Kim, TJ Ham, ...
2021 IEEE International Symposium on High-Performance Computer Architecture …, 2021
46 2021 Behemoth: A Flash-centric Training Accelerator for Extreme-scale {DNNs} S Kim, Y Jin, G Sohn, J Bae, TJ Ham, JW Lee
19th USENIX Conference on File and Storage Technologies (FAST 21), 371-385, 2021
28 2021 {ASAP}: Fast Mobile Application Switch via Adaptive Prepaging S Son, SY Lee, Y Jin, J Bae, J Jeong, TJ Ham, JW Lee, H Yoon
2021 USENIX Annual Technical Conference (USENIX ATC 21), 365-380, 2021
21 2021 Architecting a Flash-Based Storage System for Low-Cost Inference of Extreme-Scale DNNs Y Jin, S Kim, TJ Ham, JW Lee
IEEE Transactions on Computers 71 (12), 3153-3164, 2022
4 2022 Apparatus and method with scheduling JW Lee, OH Younghwan, Y Jin, TJ Ham
US Patent App. 17/887,968, 2023
2 2023 Electronic device and method with scheduling SW Lee, OH Younghwan, J Lee, SON Sam, Y Jin, HAM Taejun
US Patent App. 17/195,748, 2022
1 2022 ACCELERATOR SYSTEM FOR TRAINING DEEP NEURAL NETWORK MODEL USING NAND FLASH MEMORY AND OPERATING METHOD THEREOF JW Lee, Y Jin, JH Bae, GA Sohn, TJ Ham
US Patent App. 18/089,141, 2023
2023 SpeedLimit: Neural Architecture Search for Quantized Transformer Models Y Chai, L Bailey, Y Jin, G Ko, M Karle, D Brooks, GY Wei, H Kung
Workshop on Efficient Systems for Foundation Models@ ICML2023, 2023
2023 Layerweaver+: A QoS-Aware Layer-Wise DNN Scheduler for Multi-Tenant Neural Processing Units YH Oh, Y Jin, TJ Ham, JW Lee
IEICE TRANSACTIONS on Information and Systems 105 (2), 427-431, 2022
2022 Bigger&Faster: Two-stage Neural Architecture Search for Quantized Transformer Models. Y Chai, L Bailey, Y Jin, M Karle, GG Ko
CoRR, 2022
2022