update docs

model/utils/creator_tool.py

@@ -53,7 +53,7 @@ class ProposalTargetCreator(object):
         sampled as foregrounds.
 
         Offsets and scales of bounding boxes are calculated using
-        :func:`chainercv.links.model.faster_rcnn.bbox2loc`.
+        :func:`model.utils.bbox_tools.bbox2loc`.
         Also, types of input arrays and output arrays are same.
 
         Here are notations.
@@ -141,7 +141,7 @@ class AnchorTargetCreator(object):
 
     Offsets and scales to match anchors to the ground truth are
     calculated using the encoding scheme of
-    :func:`~chainercv.links.model.faster_rcnn.bbox2loc`.
+    :func:`model.utils.bbox_tools.bbox2loc`.
 
     .. [#] Shaoqing Ren, Kaiming He, Ross Girshick, Jian Sun. \
     Faster R-CNN: Towards Real-Time Object Detection with \

utils/config.py

@@ -8,8 +8,8 @@ from pprint import pprint
 class Config:
     # data
     voc_data_dir = '/home/cy/.chainer/dataset/pfnet/chainercv/voc/VOCdevkit/VOC2007/'
-    min_size = 600
-    max_size = 1000
+    min_size = 600  # image resize
+    max_size = 1000 # image resize
     num_workers = 8
     test_num_workers = 8
 
@@ -18,7 +18,8 @@ class Config:
     roi_sigma = 1.
 
     # param for optimizer
-    weight_decay = 0.0005  # 0.0005 in origin paper but 0.0001 in tf-faster-rcnn
+    # 0.0005 in origin paper but 0.0001 in tf-faster-rcnn
+    weight_decay = 0.0005
     lr_decay = 0.1  # 1e-3 -> 1e-4
     lr = 1e-3
 
@@ -33,20 +34,20 @@ class Config:
     pretrained_model = 'vgg16'
 
     # training
-    epoch = 100
+    epoch = 14
 
 
-    use_adam = False
-    use_chainer = False
-    use_drop = False
+    use_adam = False # Use Adam optimizer
+    use_chainer = False # try match everything as chainer
+    use_drop = False # use dropout in RoIHead
     # debug
     debug_file = '/tmp/debugf'
 
     test_num = 10000
     # model
-    load_path = None  # '/mnt/3/rpn.pth'
+    load_path = None
 
-    caffe_pretrain = False
+    caffe_pretrain = False # use caffe pretrained model instead of torchvision
     caffe_pretrain_path = 'checkpoints/vgg16-caffe.pth'
 
     def _parse(self, kwargs):