篩選條件
議題
Scisprint, hosted by the sciwork community, is a monthly coding sprint. It would like to facilitate discussions and exchanges among people in the fields of science, numerical computation, and engineering. Participants, regardless of experience level, can gain valuable development insights in this event.
To join the sprint, please bring your laptop, we're welcome everyone to join with us.
講者
Li-Hung Wang
基本上跟明星的名字同音，從姓名看起來就是個有趣的人。
活躍於 sciwork 社群，負責規劃、主導活動，
專業是網頁設計，最近想要學學 Rust ，可以轉向開發嵌入式系統或是 WASM
modmesh: Visualization that balances performance and flexibility for scientific computing
Visualization and performance are crucial in scientific computing. However, GUI development requires highly flexible code, while performance-oriented code needs specific tuning that may sacrifice flexibility. In this talk, we will demonstrate how to use a hybrid architecture (C++ and Python) that balances performance and flexibility in developing visualization applications.
講者
Chun-Hsu Lai
I’m a WiFi firmware engineer but very interested in numerical software.
pydoc-zhtw: Traditional Chinese (zh-tw) translation of the Python Documentation
Our primary mission is to make Python's official documentation accessible to the Traditional Chinese audience by providing accurate and comprehensive translations. Whether you are a seasoned Python developer or a language expert, we welcome individuals who are enthusiastic about Python and passionate about making knowledge accessible to all.
講者
Matt Wang
None
PcapPlusPlus 是 libpcap 的 C++ Wrapper，而 libpcap 則是 tcpdump 和 Wireshark 能夠分析網路封包的關鍵函式庫。本次演講將介紹 PcapPlusPlus 的架構設計以及 libpcap 的原理，並且介紹一些使用此函式庫的案例。
講者
Anchi Liu
Liu works as a software engineer in Mujin, Inc., Tokyo, Japan. He uses @tigercosmos as the name in the open-source world, and his Chinese nickname is "微中子." He holds an M.S. degree in Computer Science from National Yang Ming Chiao Tung University (NYCU) in 2022, Taiwan. He got a B.S. degree in Bio-Industrial Mechatronics Engineering from National Taiwan University (NTU) in 2019.
自 2022 年底起，隨著 AI 技術的快速進展，大型語言模型（LLMs）已成為業界與學術界矚目的焦點。本次演講，我將以最淺顯的方式，引領大家深入了解大型語言模型在實務應用中的關鍵技術與策略，涵蓋 RAG（Retrieval-Augmented Generation）知識管理、LLM 的服務化，以及模型微調（Fine-Tune）這三大領域，在此次演講，我將帶來深入RAG 知識管理和 LLM 服務化，讓大家能夠以最快的速度，了解如何導入 LLM 應用。
首先，我們將探討 RAG 如何藉由結合檢索（Retrieval）與生成（Generation）技術，提升 LLMs 的專業知識理解能力，使模型能有效處理大量資料並提供更精準的解答。
隨後，我們將分析將 LLM 轉化為服務（LLM as a service）的重要性，探討如何將這些強大的模型轉化為便於接入與使用的服務，使更多企業與開發者得以運用這些尖端技術進行創新與問題解決。
透過上面兩個維度的深入剖析，聽眾將以最易理解的方式入門 LLMs 的核心知識，為後續深入探索 LLMs 應用提供堅實的基礎，進而為未來技術創新與應用開拓更廣闊的視野。
講者
劉育維 Simon Liu
大家好，我是 Simon 劉育維，過去曾擔任過電信業的資深工程師 和 在軟體業擔任 MLOps 客戶技術成功工程師，幫助各大知名企業進行機器學習、深度學習、大型語言模型等人工智慧議題進行人工智慧架構規劃的討論，目前在 Medium 上已經公開超過 70 篇技術文章，我希望能夠嘗試使用 AI 做應用，幫助客戶用 AI 解決痛點。
uTensor: an extremely lightweight machine learning inference framework
uTensor is an extremely lightweight machine learning inference framework built on C++11. It simplifies model deployment by seamlessly converting TensorFlow-trained models into efficient C++ files that can be used to infer on the embedding device and integrate with optimized libraries such as CMSIS-NN by ARM with ease.
講者
dboyliao
None
Commitizen-tools allows teams to effectively implement version control that conforms to the specified rules. It uses interactive forms to create commit messages that meet conventional commits rules. It also bumps versions automatically using Semantic Versioning and generates a changelog using Keep a Changelog.
講者
Wei Lee
None
We will have a summarize about everyone's work, let's have fun then!
講者
Li-Hung Wang
基本上跟明星的名字同音，從姓名看起來就是個有趣的人。
活躍於 sciwork 社群，負責規劃、主導活動，
專業是網頁設計，最近想要學學 Rust ，可以轉向開發嵌入式系統或是 WASM