monkeydemon/ai-toolbox

Algorithm Engineer Toolbox, for the purpose of quickly iterating new ideas

AI-Toolbox

Introduction

Algorithm Engineer Toolbox, for the purpose of quickly iterating new ideas.

This repository is a flexible toolbox that contains many independent algorithm demos. It is designed to improve the efficiency of the work, choose the algorithm demo you interested in, and slightly modify it to quickly verify your ideas.

At present, the project is still in the initial stage, mainly covering algorithms in the field of image processing and computer vision, later will cover algorithms in other fields such as natural language processing and machine learning. continue to update...

AI-Toolbox visualization

Here is a visualization of some of the algorithms involved in the toolbox.

Quickly browse through the toolbox for the algorithm that you interested in.

All sample images used are for research purposes only and will be deleted immediately if infringement occurs. Have fun~

digital image processing
FFT for image encryption	SIFT for match	selective search	None	None
classification CNN
VGG	GoogLeNet	ResNet	DenseNet
Applications of CNN
Face Recognition	None	None	None	None
preprocess Tools
KeyFrame Extract	Data Augmentation	Data Annotation	Image Spyder	None

Algorithm checklist

Is this AI-Toolbox useful for you?

Here is a list shows which algorithms are included in our toolbox, you can check.

• CNN
  • basic exercise
    • parameter adjustment exercise(keras)
  • classification
    • VGG
    • GoogleNet
    • ResNet(Keras), ResNet(Tensorflow)
    • DenseNet
    • multi label classification
  • object detection
    • RCNN
    • YOLO
  • face recognition
    • face detection(base on Dlab)
    • face recognition(base on FaceNet)
• digital image processing
  • feature point detection
    • SIFT
  • frequency domain method
    • FFT
  • image segmentation
    • selective search
• preprocess tools
  • keyFrames extract tool
  • data augmentation tool
  • data annotation tool

ai-toolbox's Issues

keyframes_extract_diff 运行错误

➜  keyframes_extract_tool python keyframes_extract_diff.py
/usr/bin/python
target video :pikachu.mp4
frame save directory: ./extract_result/
Using Local Maxima
1923 50
Traceback (most recent call last):
  File "keyframes_extract_diff.py", line 195, in <module>
    plt.locator_params(numticks=100)
  File "/home/yopoint/.local/lib/python3.6/site-packages/matplotlib/pyplot.py", line 2744, in locator_params
    return gca().locator_params(axis=axis, tight=tight, **kwargs)
  File "/home/yopoint/.local/lib/python3.6/site-packages/matplotlib/axes/_base.py", line 3006, in locator_params
    self.xaxis.get_major_locator().set_params(**kwargs)
  File "/home/yopoint/.local/lib/python3.6/site-packages/matplotlib/ticker.py", line 2212, in set_params
    f"set_params() got an unexpected keyword argument '{key}'")
TypeError: set_params() got an unexpected keyword argument 'numticks'
➜  keyframes_extract_tool

您好，我再用百度识图爬虫的时候遇到问题

可以上传一份最新版本的baidu_image_search_demo.py吗
报错是这几行
url_upload_textbox = browser.find_element_by_css_selector('#app > div > div.page-banner > div.page-search > div > div > div.graph-search-left > input')
url_upload_textbox.send_keys(image_url)

        # 识图一下
        search_image_button = browser.find_element_by_css_selector('#app > div > div.page-banner > div.page-search > div > div > div.graph-search-center')
        search_image_button.click()

好像百度识图页面换了，> div.page-banner > div.page-search > 这些不太一样，可以麻烦看看吗？

Contribution | Dify Data Training and AI Application Training Platform

I want to share an AI application platform called Dify that can help train text data for you.I hope this tool can be included in this warehouse, because it can also help people with the same needs.

The goal of Dify is to allow developers (even non-developers) to quickly build something useful based on a large language model, and ensure that it is visualized, operable, and improveable.

In other words, you can upload papers in PDF, Markdown, TXT and other text formats and let AI learn the content. After the application is published, it can interact with people based on the uploaded data.

Practice process

I trained a dedicated chatbot and threw the product documentation of Tencent Cloud CODING DevOps into it, and the effect was good. That is to say, GPT can answer all questions related to its own company's products.

Robot experience address: https://udify.app/chat/F2Y4bKEWbuCb1FTC
Source document address: https://coding.net/help
Training platform: https://dify.ai/
My practice document (official account): https://mp.weixin.qq.com/s/WuFZ3FFcEwPNOcIKQQZsIA
Personal WeChat account: AdolfJobs

Variable weights "alpha" needs to be converted correctly !

You need to convert alpha as below to make it work with Tensorflow 2. Assuming alpha input as a list:

alpha = np.array(alpha).reshape(len(alpha), 1)
alpha = tf.convert_to_tensor(alpha, dtype=tf.float32)

Steve

关于视频提取关键帧耗时问题

作者您好，我运行了您的对于视频关键帧提取的代码，对于一个7分钟的短视频，代码需要耗时3分钟才能提取完毕，所以这个是不是有点耗时啊？