or
or
By clicking below, you agree to our terms of service.
New to HackMD? Sign up
Syntax | Example | Reference | |
---|---|---|---|
# Header | Header | 基本排版 | |
- Unordered List |
|
||
1. Ordered List |
|
||
- [ ] Todo List |
|
||
> Blockquote | Blockquote |
||
**Bold font** | Bold font | ||
*Italics font* | Italics font | ||
~~Strikethrough~~ | |||
19^th^ | 19th | ||
H~2~O | H2O | ||
++Inserted text++ | Inserted text | ||
==Marked text== | Marked text | ||
[link text](https:// "title") | Link | ||
 | Image | ||
`Code` | Code |
在筆記中貼入程式碼 | |
```javascript var i = 0; ``` |
|
||
:smile: | ![]() |
Emoji list | |
{%youtube youtube_id %} | Externals | ||
$L^aT_eX$ | LaTeX | ||
:::info This is a alert area. ::: |
This is a alert area. |
On a scale of 0-10, how likely is it that you would recommend HackMD to your friends, family or business associates?
Please give us some advice and help us improve HackMD.
Syncing
xxxxxxxxxx
馴養了一隻比猴子聰明的 AI 測試猿 - 諶毅軒
歡迎來到 MOPCON 2019 共筆
- The image file may be corrupted
- The server hosting the image is unavailable
- The image path is incorrect
- The image format is not supported
Learn More →共筆入口:https://hackmd.io/@mopcon/2019
手機版請點選上方 按鈕展開議程列表。
會場 wifi-SSID: mopcon-2019
會場 wifi-PASSWD: mopcon-2019
tags:
MOPCON 2019
Agenda
What's monkey
Fuzz testing
Advantage
2016 - 2019
Daily Crash Free Rate
Development
links
how to quick-start
Monkey Test
Limitation
Hard to reproduce
Minor issue
What's APE
APE with Object …
What We Want
TensorFlow object detection API
Setup
Training Inputs
Training Configs
Pre-trained model Zoo
Sample configs
e.g.
Training
Config, saving places …
TensorBoard 會幫你看 training 有沒有異狀
Deploy
Tensorflow Lite
Testing
Tensorflow Lite
透過 Tf Lite 來撈 API 得到結果
Deploy Tensorflow Lite to Android and iOS
How We Test
Demo 功能:
踩雷分享
About Environment
TensorFlow-CPU vs TensorFlow-GPU
Training Inputs
VoTT
VoTT 真的很方便,可以避免工人智慧。可以透過一個基本的 model 來做標記,人工再來微調
https://github.com/microsoft/VoTT
原先標記方法: 手動標記
Inference
Case1:Necessary
Case2:Not-Allowed
優點:
(fast, effective)
缺點:
F1 Measure
About label
中間不知道為什麼所有的 Label 都不見了,一氣之下把所有的 Label 設定成 test_label,結果意外的正確率沒有下降太多 -> model 比我們想像的聰明
About color
因為 image resizer 的關係,不同的裝置解析度會有差異,所以導致偵測不到。
Vision
Page Object
優點:
UI test 常出現的狀況: