kuanhungchen / VOC2007-Multilabel-TODO.pytorch

(HOMEWORK)

VOC2007-Multilabel-TODO.pytorch

Image with multilabel of diningtable, person and sofa

Requirements

• Python 3.6
• torch 0.4.1
• torchvision 0.2.1
• tqdm

Setup

1. Download PASCAL VOC 2007 dataset

   • Training / Validation (5011 images)
   • Test (4952 images)

2. Extract to data folder, now your folder structure should be like:

   VOC2007-Multilabel-TODO.pytorch
       - data
           - VOCdevkit
               - VOC2007
                   - Annotations
                       - 000001.xml
                       - 000002.xml
                       ...
                   - ImageSets
                       - Main
                           ...
                           test.txt
                           ...
                           trainval.txt
                           ...
                   - JPEGImages
                       - 000001.jpg
                       - 000002.jpg
                       ...
           - ...
   

Usage

1. Train

   $ python train.py -d=./data -c=./checkpoints
   

2. Evaluate

   $ python eval.py ./checkpoint/model-100.pth -d=./data