yolov1_training_test.txt

1. yolov1 source code download in LMS system

2. downloading VOCdatasets in yolov1-resnetgit directory
  >> wget http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtrainval_06-Nov-2007.tar
  >> wget http://host.robots.ox.ac.uk/pascal/VOC/voc2007/VOCtest_06-Nov-2007.tar
  >> wget http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar

3. Extracting all tars into one director named 'VOCdevkit'
  >> tar -xvf VOCtrainval_06-Nov-2007.tar
  >> tar -xvf VOCtest_06-Nov-2007.tar
  >> tar -xvf VOCtrainval_11-May-2012.tar

4. It should have this basic structure
 - ./VOCdevkit/VOC2007   
 - ./VOCdevkit/VOC2012

5. Generating the train and test list
 >> python3 ./tools/convert_voc.py --dir_path ./Dataset 

	├── Dataset 
	    ├── Images
	        ├── 0001.jpg
	        ├── 0002.jpg
	    ├── Labels
	        ├── 0001.txt
	        ├── 0002.txt
	    ├── train.txt
	    ├── test.txt

6.  Traing
   (1)  before training,
      a.  modifying root directory 
           - root = '../Dataset/' --> root = '../Datasets/'
           -  with open('../Dataset/train.txt') as f --> with open('./Dataset/train.txt') as f
           -  with open('../Dataset/test.txt') as f: --> with open('./Dataset/test.txt') as f:
      b.  modifying proper hyperparametes 
         - for example batch_size = 8
      c. if you have an exist pretrained weight in ./weight folder, wright down as following.
         - python3 main.py --pre_weights yolov1_0010.pth 
   
   >>python3 main.py

7. Evaluation
   >> python3 eval.py
   - In evaluation.py, im_show=False change to True to see the results.

8. Detection
   (1)  To show the result image
        >> python3 detect.py --image assets/person.jpg
   (2) To save result image
        >> python3 detect.py --image assets/person.jpg --save_img
9. Weights
   - Run the download.sh file in weights folder or download from this link