HMS-Net: Hierarchical Multi-Scale Sparsity-Invariant Network for Sparse Depth Completion von Huang, Zixuan, Fan, Junming, Cheng, Shenggan, Yi, Shuai, Wang, Xiaogang, Li, Hongsheng

Volltext

FTL: A Universal Framework for Training Low-Bit DNNs via Feature Transfer von Du, Kunyuan, Zhang, Ya, Guan, Haibing, Tian, Qi, Wang, Yanfeng, Cheng, Shenggan, Lin, James

Volltext

Hanayo: Harnessing Wave-like Pipeline Parallelism for Enhanced Large Model Training Efficiency von Liu, Ziming, Cheng, Shenggan, Zhou, Haotian, Yang, You

Volltext

ATP: Adaptive Tensor Parallelism for Foundation Models von Cheng, Shenggan, Liu, Ziming, Du, Jiangsu, You, Yang

Volltext bestellen

tcFFT: Accelerating Half-Precision FFT through Tensor Cores von Li, Binrui, Cheng, Shenggan, Lin, James

Volltext bestellen

WallFacer: Harnessing Multi-dimensional Ring Parallelism for Efficient Long Sequence Model Training von Liu, Ziming, Wang, Shaoyu, Cheng, Shenggan, Zhao, Zhongkai, Wang, Kai, Zhao, Xuanlei, Demmel, James, You, Yang

Volltext bestellen

DSP: Dynamic Sequence Parallelism for Multi-Dimensional Transformers von Zhao, Xuanlei, Cheng, Shenggan, Chen, Chang, Zheng, Zangwei, Liu, Ziming, Yang, Zheming, You, Yang

Volltext bestellen

HeteGen: Heterogeneous Parallel Inference for Large Language Models on Resource-Constrained Devices von Zhao, Xuanlei, Jia, Bin, Zhou, Haotian, Liu, Ziming, Cheng, Shenggan, You, Yang

Volltext bestellen

AutoChunk: Automated Activation Chunk for Memory-Efficient Long Sequence Inference von Zhao, Xuanlei, Cheng, Shenggan, Lu, Guangyang, Fang, Jiarui, Zhou, Haotian, Jia, Bin, Liu, Ziming, You, Yang

Volltext bestellen

FastFold: Reducing AlphaFold Training Time from 11 Days to 67 Hours von Cheng, Shenggan, Zhao, Xuanlei, Lu, Guangyang, Fang, Jiarui, Yu, Zhongming, Zheng, Tian, Wu, Ruidong, Zhang, Xiwen, Peng, Jian, You, Yang

Volltext bestellen

ATP: Adaptive Tensor Parallelism for Foundation Models von Cheng, Shenggan, Liu, Ziming, Du, Jiangsu, Yang, You

Volltext

tcFFT: Accelerating Half-Precision FFT through Tensor Cores von Li, Binrui, Cheng, Shenggan, Lin, James

Volltext

WallFacer: Harnessing Multi-dimensional Ring Parallelism for Efficient Long Sequence Model Training von Liu, Ziming, Wang, Shaoyu, Cheng, Shenggan, Zhao, Zhongkai, Wang, Kai, Zhao, Xuanlei, Demmel, James, Yang, You

Volltext

DSP: Dynamic Sequence Parallelism for Multi-Dimensional Transformers von Zhao, Xuanlei, Cheng, Shenggan, Chen, Chang, Zheng, Zangwei, Liu, Ziming, Yang, Zheming, Yang, You

Volltext

AutoChunk: Automated Activation Chunk for Memory-Efficient Long Sequence Inference von Zhao, Xuanlei, Cheng, Shenggan, Lu, Guangyang, Fang, Jiarui, Zhou, Haotian, Jia, Bin, Liu, Ziming, Yang, You

Volltext

CUBE -- Towards an Optimal Scaling of Cosmological N-body Simulations von Cheng, Shenggan, Hao-Ran, Yu, Inman, Derek, Liao, Qiucheng, Wu, Qiaoya, Lin, James

Volltext

HeteGen: Heterogeneous Parallel Inference for Large Language Models on Resource-Constrained Devices von Zhao, Xuanlei, Jia, Bin, Zhou, Haotian, Liu, Ziming, Cheng, Shenggan, Yang, You

Volltext

Task processing method, computing device and computer readable storage medium von LIN WEI, DIAO LANSONG, CHENG SHENGGAN

Volltext bestellen

HMS-Net: Hierarchical Multi-scale Sparsity-invariant Network for Sparse Depth Completion von Huang, Zixuan, Fan, Junming, Cheng, Shenggan, Yi, Shuai, Wang, Xiaogang, Li, Hongsheng

Volltext bestellen

FastFold: Reducing AlphaFold Training Time from 11 Days to 67 Hours von Cheng, Shenggan, Zhao, Xuanlei, Lu, Guangyang, Fang, Jiarui, Yu, Zhongming, Tian, Zheng, Wu, Ruidong, Zhang, Xiwen, Peng, Jian, Yang, You

Volltext