本文轉載自：
http://blog.csdn.net/sinat_30071459/article/details/53202977
說明：
本文假設你已經做好數據集，格式和VOC2007一致，並且Linux系統已經配置好caffe所需環境（博客裏教程很多），下面是訓練的一些修改。
py-R-FCN源碼下載地址：
https://github.com/Orpine/py-R-FCN
也有Matlab版本：
https://github.com/daijifeng001/R-FCN
本文用到的是Python版本。
本文主要參考https://github.com/Orpine/py-R-FCN。
準備工作：
（1）配置caffe環境(網上找教程)
（2）安裝cython, python-OpenCV, easydict
-
pip install cython
-
pip install easydict
-
apt-get install python-opencv
然後，我們就可以開始配置R-FCN了。
1.下載py-R-FCN
-
git clone https://github.com/Orpine/py-R-FCN.git
下面稱你的py-R-FCN路徑爲RFCN_ROOT.
2.下載caffe
注意，該caffe版本是微軟版本
-
cd $RFCN_ROOT
-
git clone https://github.com/Microsoft/caffe.git
如果一切正常的話，python代碼會自動添加環境變量 $RFCN_ROOT/caffe/python，否則，你需要自己添加環境變量。
3.Build Cython
4.Build caffe和pycaffe
-
cd $RFCN_ROOT/caffe
-
cp Makefile.config.example Makefile.config
然後修改Makefile.config。caffe必須支持python層，所以WITH_PYTHON_LAYER := 1是必須的。其他配置可參考：
Makefile.config
接着：
-
cd $RFCN_ROOT/caffe
-
make -j8 && make pycaffe
如果沒有出錯，則：
5.測試Demo
經過上面的工作，我們可以測試一下是否可以正常運行。
然後將模型放在$RFCN_ROOT/data。看起來是這樣的：
$RFCN_ROOT/data/rfcn_models/resnet50_rfcn_final.caffemodel
$RFCN_ROOT/data/rfcn_models/resnet101_rfcn_final.caffemodel
運行：
-
cd $RFCN_ROOT
-
./tools/demo_rfcn.py --net ResNet-50
![]()
6.用我們的數據集訓練
（1）拷貝數據集
假設我們已經做好數據集了，格式是和VOC2007一致，將你的數據集
拷貝到$RFCN_ROOT/data下。看起來是這樣的：
$VOCdevkit0712/ # development kit
$VOCdevkit/VOCcode/ # VOC utility code
$VOCdevkit/VOC0712 # image sets, annotations, etc.
# ... and several other directories ...
如果你的文件夾名字不是VOCdevkit0712和VOC0712，修改成0712就行了。
（作者是用VOC2007和VOC2012訓練的，所以文件夾名字帶0712。也可以修改代碼，但是那樣比較麻煩一些，修改文件夾比較簡單）
（2）下載預訓練模型
然後將caffemodel放在$RFCN_ROOT/data/imagenet_models (data下沒有該文件夾就新建一個)
（3）修改模型網絡
打開$RFCN_ROOT/models/pascal_voc/ResNet-50/rfcn_end2end (以end2end爲例)
注意：下面的cls_num指的是你數據集的類別數+1（背景）。比如我有15類，+1類背景，cls_num=16.
<1>修改class-aware/train_ohem.prototxt
-
layer {
-
name: 'input-data'
-
type: 'Python'
-
top: 'data'
-
top: 'im_info'
-
top: 'gt_boxes'
-
python_param {
-
module: 'roi_data_layer.layer'
-
layer: 'RoIDataLayer'
-
param_str: "'num_classes': 16" #cls_num
-
}
-
}
-
layer {
-
name: 'roi-data'
-
type: 'Python'
-
bottom: 'rpn_rois'
-
bottom: 'gt_boxes'
-
top: 'rois'
-
top: 'labels'
-
top: 'bbox_targets'
-
top: 'bbox_inside_weights'
-
top: 'bbox_outside_weights'
-
python_param {
-
module: 'rpn.proposal_target_layer'
-
layer: 'ProposalTargetLayer'
-
param_str: "'num_classes': 16" #cls_num
-
}
-
}
-
layer {
-
bottom: "conv_new_1"
-
top: "rfcn_cls"
-
name: "rfcn_cls"
-
type: "Convolution"
-
convolution_param {
-
num_output: 784 #cls_num*(score_maps_size^2)
-
kernel_size: 1
-
pad: 0
-
weight_filler {
-
type: "gaussian"
-
std: 0.01
-
}
-
bias_filler {
-
type: "constant"
-
value: 0
-
}
-
}
-
param {
-
lr_mult: 1.0
-
}
-
param {
-
lr_mult: 2.0
-
}
-
}
-
layer {
-
bottom: "conv_new_1"
-
top: "rfcn_bbox"
-
name: "rfcn_bbox"
-
type: "Convolution"
-
convolution_param {
-
num_output: 3136 #4*cls_num*(score_maps_size^2)
-
kernel_size: 1
-
pad: 0
-
weight_filler {
-
type: "gaussian"
-
std: 0.01
-
}
-
bias_filler {
-
type: "constant"
-
value: 0
-
}
-
}
-
param {
-
lr_mult: 1.0
-
}
-
param {
-
lr_mult: 2.0
-
}
-
}
-
layer {
-
bottom: "rfcn_cls"
-
bottom: "rois"
-
top: "psroipooled_cls_rois"
-
name: "psroipooled_cls_rois"
-
type: "PSROIPooling"
-
psroi_pooling_param {
-
spatial_scale: 0.0625
-
output_dim: 16 #cls_num
-
group_size: 7
-
}
-
}
-
layer {
-
bottom: "rfcn_bbox"
-
bottom: "rois"
-
top: "psroipooled_loc_rois"
-
name: "psroipooled_loc_rois"
-
type: "PSROIPooling"
-
psroi_pooling_param {
-
spatial_scale: 0.0625
-
output_dim: 64 #4*cls_num
-
group_size: 7
-
}
-
}
<2>修改class-aware/test.prototxt
-
layer {
-
bottom: "conv_new_1"
-
top: "rfcn_cls"
-
name: "rfcn_cls"
-
type: "Convolution"
-
convolution_param {
-
num_output: 784 #cls_num*(score_maps_size^2)
-
kernel_size: 1
-
pad: 0
-
weight_filler {
-
type: "gaussian"
-
std: 0.01
-
}
-
bias_filler {
-
type: "constant"
-
value: 0
-
}
-
}
-
param {
-
lr_mult: 1.0
-
}
-
param {
-
lr_mult: 2.0
-
}
-
}
-
layer {
-
bottom: "conv_new_1"
-
top: "rfcn_bbox"
-
name: "rfcn_bbox"
-
type: "Convolution"
-
convolution_param {
-
num_output: 3136 #4*cls_num*(score_maps_size^2)
-
kernel_size: 1
-
pad: 0
-
weight_filler {
-
type: "gaussian"
-
std: 0.01
-
}
-
bias_filler {
-
type: "constant"
-
value: 0
-
}
-
}
-
param {
-
lr_mult: 1.0
-
}
-
param {
-
lr_mult: 2.0
-
}
-
}
-
layer {
-
bottom: "rfcn_cls"
-
bottom: "rois"
-
top: "psroipooled_cls_rois"
-
name: "psroipooled_cls_rois"
-
type: "PSROIPooling"
-
psroi_pooling_param {
-
spatial_scale: 0.0625
-
output_dim: 16 #cls_num
-
group_size: 7
-
}
-
}
-
layer {
-
bottom: "rfcn_bbox"
-
bottom: "rois"
-
top: "psroipooled_loc_rois"
-
name: "psroipooled_loc_rois"
-
type: "PSROIPooling"
-
psroi_pooling_param {
-
spatial_scale: 0.0625
-
output_dim: 64 #4*cls_num
-
group_size: 7
-
}
-
}
-
layer {
-
name: "cls_prob_reshape"
-
type: "Reshape"
-
bottom: "cls_prob_pre"
-
top: "cls_prob"
-
reshape_param {
-
shape {
-
dim: -1
-
dim: 16 #cls_num
-
}
-
}
-
}
-
layer {
-
name: "bbox_pred_reshape"
-
type: "Reshape"
-
bottom: "bbox_pred_pre"
-
top: "bbox_pred"
-
reshape_param {
-
shape {
-
dim: -1
-
dim: 64 #4*cls_num
-
}
-
}
-
}
<3>修改train_agnostic.prototxt
-
layer {
-
name: 'input-data'
-
type: 'Python'
-
top: 'data'
-
top: 'im_info'
-
top: 'gt_boxes'
-
python_param {
-
module: 'roi_data_layer.layer'
-
layer: 'RoIDataLayer'
-
param_str: "'num_classes': 16" #cls_num
-
}
-
}
-
layer {
-
bottom: "conv_new_1"
-
top: "rfcn_cls"
-
name: "rfcn_cls"
-
type: "Convolution"
-
convolution_param {
-
num_output: 784 #cls_num*(score_maps_size^2) ###
-
kernel_size: 1
-
pad: 0
-
weight_filler {
-
type: "gaussian"
-
std: 0.01
-
}
-
bias_filler {
-
type: "constant"
-
value: 0
-
}
-
}
-
param {
-
lr_mult: 1.0
-
}
-
param {
-
lr_mult: 2.0
-
}
-
}
-
layer {
-
bottom: "rfcn_cls"
-
bottom: "rois"
-
top: "psroipooled_cls_rois"
-
name: "psroipooled_cls_rois"
-
type: "PSROIPooling"
-
psroi_pooling_param {
-
spatial_scale: 0.0625
-
output_dim: 16 #cls_num ###
-
group_size: 7
-
}
-
}
<4>修改train_agnostic_ohem.prototxt
-
layer {
-
name: 'input-data'
-
type: 'Python'
-
top: 'data'
-
top: 'im_info'
-
top: 'gt_boxes'
-
python_param {
-
module: 'roi_data_layer.layer'
-
layer: 'RoIDataLayer'
-
param_str: "'num_classes': 16" #cls_num ###
-
}
-
}
-
layer {
-
bottom: "conv_new_1"
-
top: "rfcn_cls"
-
name: "rfcn_cls"
-
type: "Convolution"
-
convolution_param {
-
num_output: 784 #cls_num*(score_maps_size^2) ###
-
kernel_size: 1
-
pad: 0
-
weight_filler {
-
type: "gaussian"
-
std: 0.01
-
}
-
bias_filler {
-
type: "constant"
-
value: 0
-
}
-
}
-
param {
-
lr_mult: 1.0
-
}
-
param {
-
lr_mult: 2.0
-
}
-
}
-
layer {
-
bottom: "rfcn_cls"
-
bottom: "rois"
-
top: "psroipooled_cls_rois"
-
name: "psroipooled_cls_rois"
-
type: "PSROIPooling"
-
psroi_pooling_param {
-
spatial_scale: 0.0625
-
output_dim: 16 #cls_num ###
-
group_size: 7
-
}
-
}
<5>修改test_agnostic.prototxt
-
layer {
-
bottom: "conv_new_1"
-
top: "rfcn_cls"
-
name: "rfcn_cls"
-
type: "Convolution"
-
convolution_param {
-
num_output: 784 #cls_num*(score_maps_size^2) ###
-
kernel_size: 1
-
pad: 0
-
weight_filler {
-
type: "gaussian"
-
std: 0.01
-
}
-
bias_filler {
-
type: "constant"
-
value: 0
-
}
-
}
-
param {
-
lr_mult: 1.0
-
}
-
param {
-
lr_mult: 2.0
-
}
-
}
-
layer {
-
bottom: "rfcn_cls"
-
bottom: "rois"
-
top: "psroipooled_cls_rois"
-
name: "psroipooled_cls_rois"
-
type: "PSROIPooling"
-
psroi_pooling_param {
-
spatial_scale: 0.0625
-
output_dim: 16 #cls_num ###
-
group_size: 7
-
}
-
}
-
layer {
-
name: "cls_prob_reshape"
-
type: "Reshape"
-
bottom: "cls_prob_pre"
-
top: "cls_prob"
-
reshape_param {
-
shape {
-
dim: -1
-
dim: 16 #cls_num ###
-
}
-
}
-
}
(4)修改代碼
<1>$RFCN/lib/datasets/pascal_voc.py
-
class pascal_voc(imdb):
-
def __init__(self, image_set, year, devkit_path=None):
-
imdb.__init__(self, 'voc_' + year + '_' + image_set)
-
self._year = year
-
self._image_set = image_set
-
self._devkit_path = self._get_default_path() if devkit_path is None \
-
else devkit_path
-
self._data_path = os.path.join(self._devkit_path, 'VOC' + self._year)
-
self._classes = ('__background__', # always index 0
-
'你的標籤1','你的標籤2',你的標籤3','你的標籤4'
-
)
改成你的數據集標籤。
<2>$RFCN_ROOT/lib/datasets/imdb.py
主要是assert (boxes[:, 2] >= boxes[:, 0]).all()可能出現AssertionError，具體解決辦法參考：
PS：
上面將有無ohem的prototxt都改了，但是這裏訓練用的是ohem。
另外，默認的迭代次數很大，可以修改$RFCN\experiments\scripts\rfcn_end2end_ohem.sh:
-
case $DATASET in
-
pascal_voc)
-
TRAIN_IMDB="voc_0712_trainval"
-
TEST_IMDB="voc_0712_test"
-
PT_DIR="pascal_voc"
-
ITERS=110000
修改ITERS爲你想要的迭代次數即可。
（5）開始訓練
-
cd $RFCN_ROOT
-
./experiments/scripts/rfcn_end2end_ohem.sh 0 ResNet-50 pascal_voc
正常的話，就開始迭代了：
![]()
$RFCN_ROOT/experiments/scripts裏還有一些其他的訓練方法，也可以測試一下（經過上面的修改，無ohem的end2end訓練也改好了，其他訓練方法修改的過程差不多）。
（6）結果
將訓練得到的模型($RFCN_ROOT/output/rfcn_end2end_ohem/voc_0712_trainval裏最後的caffemodel)拷貝到$RFCN_ROOT/data/rfcn_models下，然後打開$RFCN_ROOT/tools/demo_rfcn.py，將CLASSES修改成你的標籤，NETS修改成你的model，im_names修改成你的測試圖片(放在data/demo下),最後：
-
cd $RFCN_ROOT
-
./tools/demo_rfcn.py --net ResNet-50
![]()
