MattKovtun / Tensorflowjsclientsideprediction
An example of a client side prediction made with TensorFlowJS
Programming Languages
javascript 184084 projects - #8 most used programming language
Projects that are alternatives of or similar to Tensorflowjsclientsideprediction
Pylogging🏉 Python Logging Library
Stars: ✭ 9 (-73.53%)
Mutual labels: flask
BibiAn e-commerce fullstack solution for Flask 出口电商全栈解决方案
Stars: ✭ 914 (+2588.24%)
Mutual labels: flask
Flask HttpauthSimple extension that provides Basic, Digest and Token HTTP authentication for Flask routes
Stars: ✭ 951 (+2697.06%)
Mutual labels: flask
Keras Flask Deploy Webapp😺 Pretty & simple image classifier app template. Deploy your own trained model or pre-trained model (VGG, ResNet, Densenet) to a web app using Flask in 10 minutes.
Stars: ✭ 856 (+2417.65%)
Mutual labels: flask
JaxbinMathJax and TeX pastebin
Stars: ✭ 15 (-55.88%)
Mutual labels: flask
Peepluspython+vue3前后端分离项目
Stars: ✭ 28 (-17.65%)
Mutual labels: flask
Flask BonesAn example of a large scale Flask application using blueprints and extensions.
Stars: ✭ 849 (+2397.06%)
Mutual labels: flask
Ludolatin🏛️ An open-source tool for learning Latin
Stars: ✭ 20 (-41.18%)
Mutual labels: flask
Corpus ChristiChurch management suite, open source, fully internationalized
Stars: ✭ 29 (-14.71%)
Mutual labels: flask
OsThis is the software framework for an OPSORO robot, to be used in conjunction with Ono2 or OPSORO grid and OPSOROHAT.
Stars: ✭ 11 (-67.65%)
Mutual labels: flask
NeurodoA web tool for the analysis of cognitive dysfunction in patient journaling during Schizophrenia clinical drug trials
Stars: ✭ 14 (-58.82%)
Mutual labels: flask
Ecache👏👏 Integrate cache(redis) [flask etc.] with SQLAlchemy.
Stars: ✭ 28 (-17.65%)
Mutual labels: flask
Flask ScaffoldPrototype Database driven Web apps in Angular 6, Bootstrap 4 and REST API's with Flask (Python 3 framework)
Stars: ✭ 853 (+2408.82%)
Mutual labels: flask
ZayzafounA simple web blog script written in Python & Flask
Stars: ✭ 30 (-11.76%)
Mutual labels: flask
Social ListenerPython project used to collect tweets and social-network data from Social's API
Stars: ✭ 9 (-73.53%)
Mutual labels: flask
Flask Boto3Flask extension that ties boto3 connectors to the application context
Stars: ✭ 27 (-20.59%)
Mutual labels: flask
Botvid 19Messenger Bot that scrapes for COVID-19 data and periodically updates subscribers via Facebook Messages. Created using Python/Flask, MYSQL, HTML, Heroku
Stars: ✭ 34 (+0%)
Mutual labels: flask
AppThe SimpleLogin back-end
Stars: ✭ 958 (+2717.65%)
Mutual labels: flask
P2p一个基于 python 的 flask 框架的资讯网站， http://119.29.100.53:8086/
Stars: ✭ 28 (-17.65%)
Mutual labels: flask
This example meant to show how to do a client side prediction for trained model
Reason
Imagine you created a model which does some great stuff and helps people. You put this model on a web, and daily usage is around 1000 queries, not a lot. The simple server can handle it, but one day this model was discovered by the public, and you started receiving 100k queries daily, the same server would probably die. So now you can either scale server and add more and more memory, or you can try rewriting prediction to the client side. In case you choose the second option here is a tutorial for you.
Components
Backend: Flask (I know that TFJS supports node now but for the sake of suitable preprocessing it's in python)
Preprocessing: cv2, numpy, any python library you want
Frontend: tensorflowjs (I have a script in head from cdn, made for python developers in order not to download a module)
Model
You can download my model or train a new one. Don't forget to convert it to an appropriate format for TFJS.
Usage
Simply run app.py, open link in browser
Upload files and press predict, if you select around 100-300 files, you can observe how browser instantly eats your CPU. :)
For more details please read:
https://medium.com/@mattkovtun/client-side-prediction-with-tensorflow-js-e143ed53235b
Enjoy!
Note that the project description data, including the texts, logos, images, and/or trademarks,
for each open source project belongs to its rightful owner.
If you wish to add or remove any projects, please contact us at
[email protected].
