請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/7818
完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.advisor | 洪士灝 | |
dc.contributor.author | Yung-Chien Lin | en |
dc.contributor.author | 林詠謙 | zh_TW |
dc.date.accessioned | 2021-05-19T17:54:43Z | - |
dc.date.available | 2021-09-01 | |
dc.date.available | 2021-05-19T17:54:43Z | - |
dc.date.copyright | 2019-08-27 | |
dc.date.issued | 2019 | |
dc.date.submitted | 2019-08-21 | |
dc.identifier.citation | [1] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, no. 7553, pp. 436–444, 2015.
[2] H. Qi, E. R. Sparks, and A. Talwalkar, “Paleo: A performance model for deep neural networks,” 2016. [Online]. Available: https://openreview.net/pdf?id=SyVVJ85lg [3] https://github.com/TalwalkarLab/paleo [4] C. Coleman, D. Narayanan, D. Kang, T. Zhao, J. Zhang, L. Nardi, P. Bailis, K. Olukotun, C. Re ́, and M. Zaharia, “Dawnbench: An end-to-end deep learning benchmark and competition,” Training, vol. 100, no. 101, p. 102, 2017. [5] https://mlperf.org [6] https://en.wikipedia.org/wiki/Convolutional_neural_network [7] https://en.wikipedia.org/wiki/Long_short-term_memory [8] “Tensorflow,” https://www.tensorflow.org/ [9] Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXivpreprint arXiv:1409.1556. [10] Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. In Advances in neural information processing systems (pp. 1097-1105). [11] Yann LeCun, Léon Bottou, Yoshua Bengio, and Patrick Haffner. (1998).Gradient-based learning applied to document recognition. In Proceedings of the IEEE (pp. 2278–2324). [12] http://www.image-net.org [13] https://github.com/tensorflow/benchmarks [14] Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016). Rethinking the inception architecture for computer vision. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 2818-2826). [15] He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778). | |
dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/7818 | - |
dc.description.abstract | 隨著深度學習的迅速發展,而為了提高執行的效率,運作深度學習的硬體也 是愈發重要,但效能較高的平臺往往也伴隨著高昂的價格,因此本研究的目標便在於讓使用者能夠快速的推算出一個系統的效能,甚至是可以在取得目標硬體之前便可對其效能做簡易的分析。
目前也有許多相關的研究,但其多是使用公式的方式作效能預測,而此方式 往往使用線性的方式做計算,如此便會忽略許多細節。而本研究所用的方式為收集足夠多相關資料,並使用深度學習的方式去學習不同的配置下的執行時間。本研究同時也對不同的神經網路做預測,甚至是未取得的硬體。 | zh_TW |
dc.description.abstract | With the rapid development of deep learning, in order to improve the efficiency of implementation, the hardware for deep learning is becoming more and more important, but the platform with higher performance is often accompanied by high prices. Therefore, the goal of this research is to let users can quickly calculate the performance of a system, and even can easily analyze its performance before getting the target hardware.
There are a lot of related researches at present, but most of them use formulas to make performance predictions, and this method often uses linear methods to do calculations, so many details are ignored. The method used in this study is to collect enough relevant data and use deep learning to learn the computation time under different configurations. This study also predicts different neural networks, even for unavailable hardware. | en |
dc.description.provenance | Made available in DSpace on 2021-05-19T17:54:43Z (GMT). No. of bitstreams: 1 ntu-108-R06922155-1.pdf: 3007297 bytes, checksum: 6146c196822640ff9887937d371e4106 (MD5) Previous issue date: 2019 | en |
dc.description.tableofcontents | 誌謝 i
中文摘要 ii ABSTRACT iii 目錄 iv 圖目錄 vi 表目錄 vii 第一章 研究簡介 1 1.1 研究動機 1 1.2 相關研究 1 1.3 問題 2 1.4 解決方法 3 第二章 研究背景 5 2.1 神經網路 5 2.2 影響神經網路效能之軟體參數 6 2.3 神經網路的計算結構及軟體架構 6 2.4 PALEO 8 第三章 實驗方法 10 3.1 實作方式 10 3.2 為何使用全連接層? 11 3.3 模型架構 12 3.4 資料收集 13 3.5 如何收集合適之資料 14 3.6 資料收集與分析工具之自動化 15 第四章 實驗結果 17 4.1 預測單層神經網路 17 4.1.1 未使用批次標準化之結果 18 4.1.2 批次標準化 18 4.1.3 使用批次標準化之結果 19 4.1.3 較高誤差項之觀察 22 4.1.5 TensorFlow Timeline 24 4.2 完整模型之預測 26 4.3 其它硬體架構之預測 28 4.4 未取得硬體之預測 30 第五章 結論與未來目標 35 參考資料 37 | |
dc.language.iso | zh-TW | |
dc.title | 預測深度學習模型於加速裝置上之執行時間 | zh_TW |
dc.title | Predicting the Computation Time of Deep Learning Model on Accelerators | en |
dc.type | Thesis | |
dc.date.schoolyear | 107-2 | |
dc.description.degree | 碩士 | |
dc.contributor.oralexamcommittee | 施吉昇,涂嘉恒 | |
dc.subject.keyword | 深度學習,效能預測,系統效能,執行時間, | zh_TW |
dc.subject.keyword | Deep learning,Performance prediction,Performance of a system,Computation time, | en |
dc.relation.page | 38 | |
dc.identifier.doi | 10.6342/NTU201903950 | |
dc.rights.note | 同意授權(全球公開) | |
dc.date.accepted | 2019-08-21 | |
dc.contributor.author-college | 電機資訊學院 | zh_TW |
dc.contributor.author-dept | 資訊工程學研究所 | zh_TW |
顯示於系所單位: | 資訊工程學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-108-1.pdf | 2.94 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。