Published By
Created On
5 Nov 2021 14:40:17 UTC
Transaction ID
Cost
Safe for Work
Free
Yes
How PyCon APAC 2014 Web is built (PyCon APAC 2014)
Speaker: Keith Yang
Sitting on Web Chair of PyCon APAC 2014, I'm still new to Django and we're even new to Mezzanine. How we make it happen? It might be related to these keywords from the perspective of a Pyramid user:
- Mezzanine Rich Text Page & html content
- Mezzanine style
- Multihost middleware
- Django user, admin, and doc
- Django Workshop in Taipei.py
- django-bootstrap
- django-celery-email
- django-compress
- AngularJS
- Stylus
- Fabric
- Transifex
- Bitbucket & Github
About the speaker
An artisan and ranger chief of web development.
Sitting on Web Chair in PyCon APAC 2014, he's also an organizer for Taipei.py - Taipei Python User Group, working on PyClab branch, forwarding to CS education quest. He uses Python every day in biideal, a local startup in Taiwan.
...
https://www.youtube.com/watch?v=GbeIjkhTnik
Sitting on Web Chair of PyCon APAC 2014, I'm still new to Django and we're even new to Mezzanine. How we make it happen? It might be related to these keywords from the perspective of a Pyramid user:
- Mezzanine Rich Text Page & html content
- Mezzanine style
- Multihost middleware
- Django user, admin, and doc
- Django Workshop in Taipei.py
- django-bootstrap
- django-celery-email
- django-compress
- AngularJS
- Stylus
- Fabric
- Transifex
- Bitbucket & Github
About the speaker
An artisan and ranger chief of web development.
Sitting on Web Chair in PyCon APAC 2014, he's also an organizer for Taipei.py - Taipei Python User Group, working on PyClab branch, forwarding to CS education quest. He uses Python every day in biideal, a local startup in Taiwan.
...
https://www.youtube.com/watch?v=GbeIjkhTnik
Author
Content Type
Unspecified
video/mp4
Language
English
More from the publisher
Controlling
VIDEO
DATA
data-analysis-in-python-(pycon-apac
Speaker: Wei-Ting Kuo
Python is one of the best language for data analysis. In this presentation, I'll give an introduction for general data analysis tools written in python, including ipython, numpy, pandas, matplotlib, and sci-kit learn. Some advance topic might be included if time permit: NLTK, SimpleCV, statsmodels, numba etc...
...
https://www.youtube.com/watch?v=L43pEDyUN8I
Transaction
Created
1 year ago
Content Type
Language
video/mp4
English
Controlling
VIDEO
實
實現超自動化的驗證碼辨識
PyCon APAC 2022｜一般演講 Talks｜國泰金控 Cathay Financial Holdings / 美光科技 Micron 冠名贊助
✏️ 共筆 Note：https://hackmd.io/@pycontw/rk0o8aQ1s
?? Slido：https://app.sli.do/event/vnvojAmCGJ3BuYiqE1viLY
? 投影片 Slides：https://docs.google.com/presentation/d/1WsdQoRojBKlfOrDNppgbCmHK1DFC5gOh/edit#slide=id.p1
? 語言 Language：中文演講/英文投影片 Chinese talk w. English slides
? 層級 Level：中階 Intermediate
? 分類 Category：應用 Application
? 摘要 Abstract ?
本演講著重於讓聽眾了解自己如何實現自動辨識驗證碼模型的整體流程，我們是如何進行蒐集文字驗證碼的正確答案以確保獲得足夠的資料，並提供講者的訓練模型方式與經驗。結合 RPA (Robotic Process Automation，機器人流程自動化) 與 AI (Artificial Intelligence，人工智能) 兩個領域，我們可以實現網路驗證碼辨識超自動化 (Hyperautomation) 任務。
? 說明 Description ?
1. 為何需要破解驗證碼：(約 1 分鐘)
由於目前所擔任的職務是負責 RPA 相關任務，常常會收到使用者需要到某個網頁上查詢或點擊按鈕完成高工時且重複性高的例行工作任務。但有礙於部分網站在登入的時候常常會有文字驗證碼需要輸入，導致使用上有斷點，使用者必須人工登打。為了提升使用者滿意度體驗，我們嘗試破解驗證碼，結合深度學習訓練 OCR (Optical Character Recognition，光學字元辨識) 模型，實現超自動化任務。
- RPA：機器人流程自動化，取自 Robotic Process Automation 的縮寫，是指透過軟體技術的應用，達到自動協助辦公室工作者完成日常工作的解決方案。
- OCR：光學字元辨識，取自 Optical Character Recognition 的縮寫，是指對文字資料的圖像檔案進行分析辨識處理，取得文字及版面資訊的過程。
- 超自動化：超自動化是傳統業務流程自動化的延伸，超越了單一流程的範圍。透過將 AI 工具與 RPA 相互結合，超級自動化幾乎可為企業使用者實現所有重複性任務的自動化。
2. 千奇百怪的驗證碼：(約 1 分鐘)
本節跟各位聽眾展示各政府機關等等奇型怪狀的文字驗證碼，其中我們看到有些驗證碼相當歪斜，這會導致模型成效辨識不彰的原因之ㄧ。
3. 想像的做法：(約 1 分鐘)
有些網站很有趣，雖然有圖形驗證碼，但其實會附上答案的連結；有些是會提供音檔，而音檔的答案會附在 script 腳本。以上這些是比較少數的案例，但大多的時候，我們是沒有辦法直接取得驗證碼的答案，就必須透過 OCR 模型辨識取得答案。在還沒有自己訓練模型以前，過去會使用鼎鼎大名的 tesseract OCR 或是語音破解。我們也嘗試使用 tesseract ocr (https://github.com/tesseract-ocr/) 模型去跑搜集的驗證碼一萬張成效，平均辨識成功成效僅為 2.71%，因此在想完成自動登入網站的時候，常常會失敗。
4. 如何進行資料蒐集：(約 8 分鐘)
驗證碼搜集的關鍵在於大量包含正確答案的圖片，有越多的資料才有辦法訓練出一個好模型。透過以下方式搜集約 20 萬張驗證碼圖片和正確答案。蒐集到的答案的方式，統一命名為「答案_uuid.txt」避免檔名重複。本節展示這次模型抓了哪些網站的驗證碼及資料張數。此外，若爬到的驗證碼不確定大小寫，不會用此資料。
為蒐集大量的驗證碼圖片，本節將介紹 6 種有趣的方法來抓驗證碼資料或產製的方法。
5. 模型：(約 5 分鐘)
這次使用的 OCR 模型名稱為 ClovaOCR (https://github.com/clovaai/deep-text-recognition-benchmark)，參照 GITHUB README 即可自行訓練深度學習模型。以下附上訓練策略及成效
- 訓練環境使用 GPU
- 訓練集/驗證集 為 8 : 2 → 16萬：4萬
- 訓練 iteration：30 萬
- 最後 Training Accuracy： 0.99
- 訓練時間： 約 41.23 小時
- 比較 tesseract OCR，新的模型 ClovaOCR 則是在驗證集有 93.49% 的成效，每一張辨識時間僅需 0.1713 秒。
最後也與大家說明如何將 ClovaOCR 將模型推斷重構，符合驗證碼單張推斷的辨識方式。
? 講者介紹 About Speaker - 何信賢 ?
擔任 RPA 工程師與全端工程師，喜歡研究爬蟲和開發網站的服務。 歡迎來找我打羽球或是交流寫程式的心得:)
#pycontw #pyconapac2022 #python #clovaocr #tesseract #ocr #rpa
Follow “PyCon Taiwan”
⭐️ Official Website: https://tw.pycon.org
⭐️ Facebook: https://www.facebook.com/pycontw
⭐️ Instagram: https://www.instagram.com/pycontw
⭐️ Twitter: https://twitter.com/PyConTW
⭐️ LinkedIn: https://www.linkedin.com/company/pycontw
⭐️ Blogger: https://pycontw.blogspot.com
...
https://www.youtube.com/watch?v=oD-FT_33yW4
Transaction
Created
1 year ago
Content Type
Language
video/mp4
zh
Controlling
VIDEO
GETTI
getting-started-with-sparse-modeling
Day 2, R2 11:45–12:15
Blackbox problem has been becoming a popular concern when applying machine learning in specific applications, like medical system, where a user is supposed to understand the behavior of the system. Collecting tons of data for training machine learning model is another headache especially when you newly create a system from scratch. In this talk, I introduce the data analysis approach called "Sparse Modeling" that can produce good results, even if the amount of data is small. Event Horizon Telescope project, capturing blackhole image, is one good example of this nature. It's also referred to as explainable since it can tell you which input features have a strong impact to result generated by a machine learning model. With the overview of the method, I'll show concrete code examples for common use cases like image analysis, using a Python library named spm-image.
Slides: https://speakerdeck.com/hacarus/getting-started-with-sparse-modeling-with-spm-image
Speaker: Takashi Someda
After getting his master’s degree in informatics at Graduate School of Kyoto University, he started his job at Sun Microsystems as an engineer.
For about 20 years in the software industry, he has experienced several roles like software developer, technical evangelist, and data scientist.
Now, as CTO of Hacarus, he is responsible for technical direction with strong passion toward building a creative, self-organized team like Pixer.
...
https://www.youtube.com/watch?v=KjPyDhbqzKE
Transaction
Created
1 year ago
Content Type
Language
video/mp4
English
Controlling
VIDEO
S3EP2
s3ep2-官網大翻新！ui-ux
【S3EP2 | 官網大翻新！UI/UX 設計師聊多了都是淚（？】
2021 年，一群熱血的志工啟動了 PyCon TW 官網大改版，將前後端分離，並且導入介面設計的觀念，才誕生了現在精美又現代化的網站。
可以發現網站從 2020 到 2023 年，網站主視覺與易用性上有很多不一樣的突破！
這次我們邀請到網站改版的幕後推手，一起揭開改版過程中，介面設計師與工程師之間相愛相殺（？的心路歷程。
#PyConTW #2023 #python #website #uiux
-
?點擊收聽：https://reurl.cc/6NMWb5
-
PyCon TW 2023 志工報名表單：https://forms.gle/vLz3Xev8tUpnRqacA
Follow “PyCon Taiwan”
⭐️ Official Website: https://tw.pycon.org
⭐️ Facebook: https://www.facebook.com/pycontw
⭐️ Instagram: https://www.instagram.com/pycontw
⭐️ Twitter: https://twitter.com/PyConTW
⭐️ LinkedIn: https://www.linkedin.com/company/pycontw
⭐️ Blogger: https://pycontw.blogspot.com
...
https://www.youtube.com/watch?v=Qp0FZV7RySs
Transaction
Created
1 year ago
Content Type
Language
video/mp4
zh
Controlling
VIDEO
R2 DA
r2-day1-04-教練，我想寫-django-app
Speaker: Tim Hsu
有的時候，我們會把我們重複的 django 程式碼抽出來變成 django 套件。 但是要能夠讓人方便的重複使用，卻有很多技巧。這裏會分析 Django 常見第三方套件的製作方式。
About the speaker
I am tim, love the simplicity of python
組織/公司 Dorm7
頭銜 Engineer
https://tw.pycon.org/2015apac/zh/program/16
...
https://www.youtube.com/watch?v=biFx5BmURwc
Transaction
Created
1 year ago
Content Type
Language
video/mp4
English
Controlling
VIDEO
BUILD
building-an-arena-for-ai-warriors-pei
PyCon APAC 2022｜一般演講 Talks｜國泰金控 Cathay Financial Holdings / 美光科技 Micron 冠名贊助
✏️ 共筆 Note：https://hackmd.io/@pycontw/ryIyv6XJs
?? Slido：https://app.sli.do/event/8GNQNQGS5N4w7C8xhN1afw
? 投影片 Slides：https://drive.google.com/file/d/1--85I7Qyqs5TDh6PTUDlfqyvf0IizieB/view?usp=sharing
? 語言 Language：中文演講/英文投影片 Chinese talk w. English slides
? 層級 Level：中階 Intermediate
? 分類 Category：應用 Application
? 摘要 Abstract ?
關於機器學習線上比賽，kaggle 類型上傳結果評分的比賽已是遍地開花，但機器學習服務除了考量「模型準確度」還需考慮「服務回應的速度與品質」，如果是即時發題、即時 inference 類型的比賽要怎麼進行呢? 在這個演講中，將會分享如何舉辦一個線上 Machine Learning api 競賽、比較現行常見的機器學習競賽方式，同時探討辦一個線上 Machine Learning api 競賽的競賽機制設計、因應機制而生的系統架構設計、實際架構的效能分享。
? 說明 Description ?
使用到的工具：
- 利用 Flask 完成 slack chatbot
- 使用 message queue 完成非同步的 request 發送
- 全系統平台建置於 GCP，並使用 gke, pub/sub 相關工具
? 講者介紹 About Speaker - Pei-Hsuan Hsieh ?
是個喜歡金融領域的機器學習平台工程師，負責建置 ML platform 中 ETL、API 等各項服務的 CI/CD pipeline ，讓服務能自動化上線。
#pycontw #pyconapac2022 #python #flask #kaggle #machinelearning #chatbot
Follow “PyCon Taiwan”
⭐️ Official Website: https://tw.pycon.org
⭐️ Facebook: https://www.facebook.com/pycontw
⭐️ Instagram: https://www.instagram.com/pycontw
⭐️ Twitter: https://twitter.com/PyConTW
⭐️ LinkedIn: https://www.linkedin.com/company/pycontw
⭐️ Blogger: https://pycontw.blogspot.com
...
https://www.youtube.com/watch?v=bhS0xGzBhcc
Transaction
Created
1 year ago
Content Type
Language
video/mp4
Controlling
VIDEO
[PYCO
pycontw-2013-駭客看-django-by-orange
About Orange:
簡介
I am Orange :)
現任大學生中
專長
主要專精 Web Security 以及 Windows 上的弱點研究
經歷
CHROOT Security Group 成員
台灣駭客年會 2009 Wargame 冠軍
行政院國家技服中心舉辦資安技能金盾獎第六屆、第七屆 冠軍
日本資安研討會 AVTOKYO 2011 講師
香港資安研討會 VXConf 2012 講師
台灣 WebConf, PHPConf 講師
部落格
blog.orange.tw
...
https://www.youtube.com/watch?v=iQXt_7ISGHQ
Transaction
Created
1 year ago
Content Type
Language
video/mp4
Controlling
VIDEO
TASTE
tastegram-an-image-assessment-app-built
PyCon APAC 2022｜一般演講 Talks｜國泰金控 Cathay Financial Holdings / 美光科技 Micron 冠名贊助
✏️ 共筆 Note：https://hackmd.io/@pycontw/BJcR8pmki
?? Slido：https://app.sli.do/event/3ovdbifBx8hXaoWD8EgcxX
? 投影片 Slides：https://bit.ly/pycon22
? 語言 Language：英文 English
? 層級 Level：中階 Intermediate
? 分類 Category：機器學習 Machine Learning
? 摘要 Abstract ?
TensorFlow is an end-to-end open source platform for machine learning. It uses a user-friendly API for defining and training neural networks. It provides stable Python APIs and embraces the style and ergonomics of the Python language. The goal of this talk is to help developers get familiar with this end-to-end platform. To do so, we're going to build an Image Assessment application, TasteGram, completely from scratch all using Python.
? 說明 Description ?
For this talk, we're going to take a top-down approach and start from the aesthetic data. Here, I'll talk about some useful ways of how to preprocess the dataset out of Keras Preprocessing released this year.
And then we'll move on to building and training the model, where we'll get to understand a little bit about how to apply the transfer learning to learn an aesthetic evaluation task.
Finally, we'll tying all of these pieces together at the bottom level and talk a bit about deploying the model on edge devices.
This talk should be pretty helpful for developers to get an idea of how to turn theoretical concepts into working implementations with TensorFlow, and they can develop their own end-to-end ML application all using Python.
? 講者介紹 About Speaker - John Lu ?
John is an Android Developer at Seekrtech, currently focusing on developing a gamified To-do list app, WaterDo. He is deeply motivated by challenges and tends to be excited by breaking conventional ways of thinking and doing. With two years of experience in Machine Learning research, he works on combining the latest AI technology and design to build fun and creative applications.
#pycontw #pyconapac2022 #python #tensorflow #machinelearning #neuralnetwork #tastegram #imageassessment
Follow “PyCon Taiwan”
⭐️ Official Website: https://tw.pycon.org
⭐️ Facebook: https://www.facebook.com/pycontw
⭐️ Instagram: https://www.instagram.com/pycontw
⭐️ Twitter: https://twitter.com/PyConTW
⭐️ LinkedIn: https://www.linkedin.com/company/pycontw
⭐️ Blogger: https://pycontw.blogspot.com
...
https://www.youtube.com/watch?v=nZVphm09s6g
Transaction
Created
1 year ago
Content Type
Language
video/mp4
English
Controlling
VIDEO
DATA
data-oriented-programming-with-python
PyCon APAC 2022｜一般演講 Talks｜國泰金控 Cathay Financial Holdings / 美光科技 Micron 冠名贊助
✏️ 共筆 Note：https://hackmd.io/@pycontw/HJhevT7Jj
?? Slido：https://app.sli.do/event/u2UG4gZau2FQGhW1qmrfNj
? 投影片 Slides：https://docs.google.com/presentation/d/1HFF23gMrgV_prjwPxQ20K8_-m-fkmKcS0y-YLJ0YhTg/edit?usp=sharing
? 語言 Language：英文 English
? 層級 Level：中階 Intermediate
? 分類 Category：最佳實踐與慣例 Best Practices & Patterns
? 摘要 Abstract ?
Data-Oriented Programming (DOP) is a programming paradigm that makes the systems we build less complex inspired by Clojure. The cool thing is that DOP is language-agnostic: it is applicable to any programming language. In Python, we commonly taught using OOP for abstraction in our code. Information systems around software constructs that combine code and data (e.g. objects instantiated from classes). But in DOP, it encourages the separation of code from data. Moreover, DOP provides guidelines about how to represent and manipulate data. The essence of DOP is that it treats data as a first-class citizen; it gives developers the ability to manipulate data inside a program with the same simplicity as they manipulate numbers or strings. In this talk, we will explore what, when, and how to use DOP with Python.
? 說明 Description ?
We know OOP is relatively complex and it has some common pains that Object Oriented developers experience when they develop a system. For example, many relationships between classes, unpredictable code behavior, not trivial data serialization, complex class hierarchies.
This Talk is not meant to be read as critical of OOP. Its purpose is to raise awareness of the
tendency towards increased complexity of OOP as a programming paradigm and to motivate you
to discover a different programming paradigm where the system complexity tends to be reduced,
namely Data Oriented programming.
Data-Oriented programming (DOP) is a programming paradigm aimed at simplifying the design
and implementation of software systems where information is at the center: systems such as
front-end or back-end web applications and web services.
DOP is based on 4 fundamental principles.
1. Separate code (behavior) from data.
2. Represent data with generic data structures.
3. Data is immutable.
4. Separate data schema from data representation.
Each one of them we will explore in this talk with the Python examples.
Resources: https://blog.klipse.tech/databook/2020/09/29/do-principles.html
? 講者介紹 About Speaker - Shreya Khurana ?
Head of Engineering at Zero One Group | Tech Mentor | Mathematical Thinker
#pycontw #pyconapac2022 #python #dataoriented #dop
Follow “PyCon Taiwan”
⭐️ Official Website: https://tw.pycon.org
⭐️ Facebook: https://www.facebook.com/pycontw
⭐️ Instagram: https://www.instagram.com/pycontw
⭐️ Twitter: https://twitter.com/PyConTW
⭐️ LinkedIn: https://www.linkedin.com/company/pycontw
⭐️ Blogger: https://pycontw.blogspot.com
...
https://www.youtube.com/watch?v=vC8c2XWedm0
Transaction
Created
1 year ago
Content Type
Language
video/mp4
English