PFMonitor -- for deep learning engineer's freedom
The PFMonitor project is aimed to free your eyes on the computer screen while you are training your neural networks.
It designed for engineers,students and professors to CHECK your deep learning training process just by PHONE, not sticking to the front of PCs.
This monitor contains two main features:
- Cross-platform. you can show the notifications on ios, android, linux, ubuntu, windows.
- Free of use
Run the monitor script on the PCs or servers which are training some models after setting the correct parameters,
and the monitor will check your process's status and the file list under your folder.
If there are some changes, monitor will send notifications to the local computer and remote phone
Getting Started
Installing
- system: win10/Ubuntu/...
- python:3+
- Clone or download the repository
- Commands to install requirements via pip wheels
pip install -r requirements.txt - Install pushbullet on PC and phone, then sign in or sign up.
- Create the access token at the pushbullet.This access token is a api key for socket connection.
Usage
What the only thing you need to do is setting the correct parameters:
if __name__=='__main__':
monitor=Monitor(
api_key='***********',#access token
model_path='***********',folder which you wanna monitor
pid=666# int number not string
)
monitor.monitor()# start monitoring.......
Demo
Author
- Chris Li - all work
License
This project is licensed under the MIT License - see the LICENSE file for details