Today, I recommend a pretty awesome AI open source project, the current Star 3.4k, but boldly predict that this project will definitely be popular, and the number of Stars in the future should reach 10k or even 20k!
If you are in a hurry, you can go to GitHub to see the source code directly
Portal: https://github.com/PaddlePaddle/PaddleHub
Don’t say anything, take everyone to look at some project Readme screenshots
1. Look at the introduction first
First of all, "No need for deep learning background, no data and training process", "Share the dividends of the artificial intelligence era" and "All models are open source to download, offline can be run"
I can only say that this repo is definitely reaching out for the gospel of the party!
Continue to see the features section:
The number of models covered, CV, NLP, Audio, Video, four categories are fully covered, and the coverage is abundant. How to use: One-click prediction, one-click service, ten lines of code migration , and the technical threshold is sufficiently reduced.
Let's continue to look down for recent updates
In terms of quantity , the total number has now exceeded 180. Basically, 10-20 updates are maintained every month, and the update speed should be faster and faster. This already feels very cool.
Look at it again in terms of quality , (high-energy early warning of moving pictures ahead!)
2. Text recognition
Contains ultra-lightweight Chinese and English OCR models, high-precision Chinese and English, multilingual German, French, Japanese, Korean OCR recognition.
3. Face detection
Including face detection, mask face detection, multiple algorithms are optional.
4. Image editing
4 times the super-score effect, multiple super-score algorithms are optional.
Coloring black and white pictures can be used to repair old photos,
Image super resolution, black and white picture coloring
5. Target detection
Including pedestrian detection, vehicle detection, and more industrial-grade ultra-large-scale pre-training models are optional.
6. Key point detection
Including single person, multi-person body key point detection, face key point detection, hand key point detection.
7. Image segmentation
Contains excellent portrait cutout models and ACE2P human body analysis world champion models.
8. Image animation
Many cartoonists including Hayao Miyazaki and Makoto Shinkai are migrating their styles, and multiple algorithms are available.
9. Image classification
Including animal classification, dish classification, wild animal product classification, multiple algorithms are available.
10. Lexical analysis
Excellent Chinese word segmentation, part-of-speech tagging and named entity recognition model.
11. Text generation
Including AI poem writing, AI couplets, AI love words, AI hidden poems, multiple algorithms are available.
12. Syntactic analysis
Leading Chinese syntactic analysis model.
13. Sentiment analysis
Support Chinese comment sentiment analysis.
14. Text review
Contains the review of Chinese pornographic text, and multiple algorithms are available.
15. Speech synthesis
TTS speech synthesis algorithm, multiple algorithms are optional
输入:Life was like a box of chocolates, you never know what you're gonna get.
The synthesis effect is as follows:
16.Video classification
Contains short video classification, supports 3000+ tag types, can output TOP-K tags, and multiple algorithms are optional.
Example: Input a short video of swimming, the algorithm can output the result of "swimming".
The pre-trained models used in the above animations are all open source, offline downloadable, and free to use!
How should it be used? Don't worry, let's look down
I don’t need to say anything else. After you click star on github, let’s experience it for yourself.
https://github.com/PaddlePaddle/PaddleHub
In addition, the official WeChat group is also intimately given in the repo. If you have any questions, you can join the group and answer questions more efficiently.
If you fail to scan the code, please add WeChat: 15704308458 , and the operating school will invite you to join the group.
For more related content, please refer to the following.
Official website address: https://www.paddlepaddle.org.cn/hub
GitHub: https://github.com/PaddlePaddle/PaddleHub
Gitee: https://gitee.com/paddlepaddle/PaddleHub