mvladk / Train_Mask-RCNN-for-object-detection-in_In_60_Lines-of-Code

Train mask RCNN for object detection in pytorch in 60 lines of code

Train_Mask-RCNN-for-object-detection-in_In_60_Lines-of-Code

Toturial for this code can be seen here: https://medium.com/@sagieppel/train-mask-rcnn-net-for-object-detection-in-60-lines-of-code-9b6bbff292c3

Dataset

This code is for using training mask RCNN with pytorch and the LabPics Version 2 datast.

Requirement

This demand pytorch 1.1 and opencv. Runnig it with newere pytorch version might cause some isseus. Pytorch installation instructions are available at:

https://pytorch.org/get-started/locally/

OpenCV can be installed using:

pip install opencv-python

Requirements

Setting enviroment with conda

1. Install Anaconda
2. Create a virtual environment with the required dependencies (Pytorch, torchvision and OpenCV): conda env create -f environment.yml
3. Activate the virtual environment: conda activate maskRcnn

Hardware

For using the trained net, no specific hardware is needed, but the net will run much faster on GPU.

For effectively training the net a good GPU is needed.

Train model

Code with pre trained model weights can be download from here: https://icedrive.net/s/v6xR9fbh6F24zXSW8bBt6aji6F6x