Search

Your search keyword '"Yang, HaiHua"' showing total 641 results
Start Over Author "Yang, HaiHua"
641 results on '"Yang, HaiHua"'

2. Skywork: A More Open Bilingual Foundation Model

Author

Wei, Tianwen, Zhao, Liang, Zhang, Lichang, Zhu, Bo, Wang, Lijie, Yang, Haihua, Li, Biye, Cheng, Cheng, Lü, Weiwei, Hu, Rui, Li, Chenxia, Yang, Liu, Luo, Xilin, Wu, Xuejie, Liu, Lunan, Cheng, Wenjun, Cheng, Peng, Zhang, Jianhao, Zhang, Xiaoyu, Lin, Lei, Wang, Xiaokun, Ma, Yutuan, Dong, Chuanhai, Sun, Yanqi, Chen, Yifu, Peng, Yongyi, Liang, Xiaojuan, Yan, Shuicheng, Fang, Han, and Zhou, Yahui

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence
Abstract

In this technical report, we present Skywork-13B, a family of large language models (LLMs) trained on a corpus of over 3.2 trillion tokens drawn from both English and Chinese texts. This bilingual foundation model is the most extensively trained and openly published LLMs of comparable size to date. We introduce a two-stage training methodology using a segmented corpus, targeting general purpose training and then domain-specific enhancement training, respectively. We show that our model not only excels on popular benchmarks, but also achieves \emph{state of the art} performance in Chinese language modeling on diverse domains. Furthermore, we propose a novel leakage detection method, demonstrating that test data contamination is a pressing issue warranting further investigation by the LLM community. To spur future research, we release Skywork-13B along with checkpoints obtained during intermediate stages of the training process. We are also releasing part of our SkyPile corpus, a collection of over 150 billion tokens of web text, which is the largest high quality open Chinese pre-training corpus to date. We hope Skywork-13B and our open corpus will serve as a valuable open-source resource to democratize access to high-quality LLMs.

Published
2023

3. SkyMath: Technical Report

Author

Yang, Liu, Yang, Haihua, Cheng, Wenjun, Lin, Lei, Li, Chenxia, Chen, Yifu, Liu, Lunan, Pan, Jianfei, Wei, Tianwen, Li, Biye, Zhao, Liang, Wang, Lijie, Zhu, Bo, Li, Guoliang, Wu, Xuejie, Luo, Xilin, and Hu, Rui

Subjects
Computer Science - Computation and Language, Computer Science - Artificial Intelligence
Abstract

Large language models (LLMs) have shown great potential to solve varieties of natural language processing (NLP) tasks, including mathematical reasoning. In this work, we present SkyMath, a large language model for mathematics with 13 billion parameters. By applying self-compare fine-tuning, we have enhanced mathematical reasoning abilities of Skywork-13B-Base remarkably. On GSM8K, SkyMath outperforms all known open-source models of similar size and has established a new SOTA performance.

Published
2023

13. Bailout intracranial angioplasty or stenting following thrombectomy for acute large vessel occlusion in China (ANGEL-REBOOT): a multicentre, open-label, blinded-endpoint, randomised controlled trial

Author

Gao, Feng, Tong, Xu, Jia, Baixue, Wei, Ming, Pan, Yuesong, Yang, Ming, Sun, Dapeng, Nguyen, Thanh N, Ren, Zeguang, Demiraj, Francis, Yao, Xiaoxi, Xu, Chenghua, Yuan, Guangxiong, Wan, Yue, Tang, Jianjun, Wang, Jing, Jiang, Yuanfei, Wang, Chaobin, Luo, Xiang, Yang, Haihua, Shen, Ruile, Wu, Zhilin, Yuan, Zhengzhou, Wan, Dongjun, Hu, Wei, Liu, Yan, Jing, Ping, Wei, Liping, Zheng, Tuanyuan, Wu, Yingchun, Yang, Xinguang, Sun, Yaxuan, Wen, Changming, Chang, Mingze, Yin, Bo, Li, Di, Duan, Jixin, Sun, Dianjing, Guo, Zaiyu, Xu, Guodong, Wang, Guoqing, Wang, Liyu, Wang, Yang, Jia, Weihua, Ma, Gaoting, Huo, Xiaochuan, Mo, Dapeng, Ma, Ning, Liu, Liping, Zhao, Xingquan, Wang, Yilong, Fiehler, Jens, Wang, Yongjun, and Miao, Zhongrong

Published
2024
Full Text
View/download PDF

32. Anlotinib plus oral fluoropyrimidine S-1 in treating patients with refractory or relapsed small-cell lung cancer (SALTER TRIAL): an open-label, multicenter, single-arm, phase II trial

Author

Wang, Wei, primary, Wu, Guixian, additional, Luo, Wujun, additional, Chen, Ziran, additional, Lin, Ling, additional, Zhou, Chao, additional, Yao, Guifei, additional, Chen, Meifang, additional, Wu, Xiaomai, additional, Ye, Junhui, additional, Yang, Haihua, additional, and Lv, Dongqing, additional

Published
2024
Full Text
View/download PDF

38. Efficacy of Aumolertinib (HS-10296) in Patients With Advanced EGFR T790M+ NSCLC: Updated Post-National Medical Products Administration Approval Results From the APOLLO Registrational Trial

Author

Lu, Shun, Wang, Qiming, Zhang, Guojun, Dong, Xiaorong, Yang, Cheng-Ta, Song, Yong, Chang, Gee-Chen, Lu, You, Pan, Hongming, Chiu, Chao-Hua, Wang, Zhehai, Feng, Jifeng, Zhou, Jianying, Xu, Xingxiang, Guo, Renhua, Chen, Jianhua, Yang, Haihua, Chen, Yuan, Yu, Zhuang, Shiah, Her-Shyong, Wang, Chin-Chou, Yang, Nong, Fang, Jian, Wang, Ping, Wang, Kai, Hu, Yanping, He, Jianxing, Wang, Ziping, Shi, Jianhua, Chen, Shaoshui, Wu, Qiong, Sun, Changan, Li, Chuan, Wei, Hongying, Cheng, Ying, Su, Wu-Chou, Hsia, Te-Chun, Cui, Jiuwei, Sun, Yuping, Ou, Sai-Hong Ignatius, Zhu, Viola W., and Chih-Hsin Yang, James

Published
2022
Full Text
View/download PDF

47. Feasibility of patient‐specific quality assurance (PSQA) for real‐time robotic stereotactic body radiotherapy (SBRT) based on tumor motion traces.

Author

Xu, Qianyi, Fan, Jiajin, Vinogradskiy, Yevgeniy, Chawla, Ashish K., Kubicek, Gregory, Yang, Haihua, Huynh, Kiet, LaCouture, Tamara, Grimm, Jimm, and Nie, Wei

Subjects
STEREOTACTIC radiotherapy, MEDICAL dosimetry, SYNCHRONIC order, QUALITY assurance, ANGLES
Abstract

Purpose: To design a patient specific quality assurance (PSQA) process for the CyberKnife Synchrony system and quantify its dosimetric accuracy using a motion platform driven by patient tumor traces with rotation. Methods: The CyberKnife Synchrony system was evaluated using a motion platform (MODUSQA) and a SRS MapCHECK phantom. The platform was programed to move in the superior‐inferior (SI) direction based on tumor traces. The detector array housed by the StereoPhan was placed on the platform. Extra rotational angles in pitch (head down, 4.0° ± 0.15° or 1.2° ± 0.1°) were added to the moving phantom to examine robot capability of angle correction during delivery. A total of 15 Synchrony patients were performed SBRT PSQA on the moving phantom. All the results were benchmarked by the PSQA results based on static phantom. Results: For smaller pitch angles, the mean gamma passing rates were 99.75% ± 0.87%, 98.63% ± 2.05%, and 93.11% ± 5.52%, for 3%/1 mm, 2%/1 mm, and 1%/1 mm, respectively. Large discrepancy in the passing rates was observed for different pitch angles due to limited angle correction by the robot. For larger pitch angles, the corresponding mean passing rates were dropped to 93.00% ± 10.91%, 88.05% ± 14.93%, and 80.38% ± 17.40%. When comparing with the static phantom, no significant statistic difference was observed for smaller pitch angles (p = 0.1 for 3%/1 mm), whereas a larger statistic difference was observed for larger pitch angles (p < 0.02 for all criteria). All the gamma passing rates were improved, if applying shift and rotation correction. Conclusions: The significance of this work is that it is the first study to benchmark PSQA for the CyberKnife Synchrony system using realistically moving phantoms with rotation. With reasonable delivery time, we found it may be feasible to perform PSQA for Synchrony patients with a realistic breathing pattern. [ABSTRACT FROM AUTHOR]

Published
2024
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results