Skip to content

Latest commit

 

History

History
49 lines (29 loc) · 1.1 KB

README.md

File metadata and controls

49 lines (29 loc) · 1.1 KB
Raw

RGMP PyTorch

This is forked from the official demo code for the paper. PDF

Added training script with TensorBoard support.

Test Environment

  • Ubuntu
  • python 3.6
  • Pytorch 0.3.1
    • installed with CUDA.

How to Run Inference

  1. Download DAVIS-2017.
  2. Edit path for DAVIS_ROOT in run.py.
DAVIS_ROOT = '<Your DAVIS path>'
  1. Download weights.pth and place it the same folde as run.py.
  2. To run single-object video object segmentation on DAVIS-2016 validation.
python run.py
  1. To run multi-object video object segmentation on DAVIS-2017 validation.
python run.py -MO
  1. Results will be saved in ./results/SO or ./results/MO.

How to train a model

python3 train.py

TensorBoard Support

Install TensorBoardX to view loss, IoU and generated masks in real-time during training.