Følg
Shih-Yang Liu
Shih-Yang Liu
PhD Student @ HKUST, NVIDIA Research
Verifisert e-postadresse på connect.ust.hk - Startside
Tittel
Sitert av
Sitert av
År
DoRA: Weight-Decomposed Low-Rank Adaptation
SY Liu, CY Wang, H Yin, P Molchanov, YCF Wang, KT Cheng, MH Chen
ICML 2024 (Oral), 2024
3182024
LLM-FP4: 4-Bit Floating-Point Quantized Transformers
SY Liu, Z Liu, X Huang, P Dong, KT Cheng
EMNLP 2023 Main Conference, 2023
562023
Oscillation-free quantization for low-bit vision transformers
SY Liu, Z Liu, KT Cheng
ICML 2023, 21813-21824, 2023
312023
Efficient quantization-aware training with adaptive coreset selection
X Huang, Z Liu, SY Liu, KT Cheng
122023
Ipr: Interaction-level preference ranking for explicit feedback
SY Liu, HH Chen, CM Chen, MF Tsai, CJ Wang
Proceedings of the 45th International ACM SIGIR Conference on Research and …, 2022
62022
Hymba: A hybrid-head architecture for small language models
X Dong, Y Fu, S Diao, W Byeon, Z Chen, AS Mahabaleshwarkar, SY Liu, ...
arXiv preprint arXiv:2411.13676, 2024
52024
RoLoRA: Fine-tuning Rotated Outlier-free LLMs for Effective Weight-Activation Quantization
X Huang, Z Liu, SY Liu, KT Cheng
EMNLP 2024 Findings, 2024
22024
Genetic Quantization-Aware Approximation for Non-Linear Operations in Transformers
P Dong, Y Tan, D Zhang, T Ni, X Liu, Y Liu, P Luo, L Liang, SY Liu, ...
2024 61st ACM/IEEE Design Automation Conference (DAC), 2024
22024
EoRA: Training-free Compensation for Compressed LLM with Eigenspace Low-Rank Approximation
SY Liu, H Yang, CY Wang, NC Fung, H Yin, C Sakr, S Muralidharan, ...
arXiv preprint arXiv:2410.21271, 2024
12024
CMOSE: Comprehensive Multi-Modality Online Student Engagement Dataset with High-Quality Labels
CH Wu, SY Liu, X Huang, X Wang, R Zhang, L Minciullo, WK Yiu, K Kwan, ...
Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern …, 2024
12024
Robust and Efficient Quantization-aware Training via Coreset Selection
X Huang, Z Liu, SY Liu, KT Cheng
Transactions on Machine Learning Research, 2024
2024
Systemet kan ikke utføre handlingen. Prøv på nytt senere.
Artikler 1–11