Small changes.

* commit '962ecb0403eb44af10ebb685675ad9caf04ac9d9':
  clean up redundent function.
  Fix compatibility for python3
  clean up bbox.
  clean up nms.
  fix heading space.
  fix link.
  Update README.md
  fix visualization.

# Conflicts:
#	README.md
#	experiments/scripts/train_faster_rcnn.sh
#	lib/model/test.py
#	lib/setup.py
#	lib/utils/bbox.pyx
#	lib/utils/visualization.py

Author: ruotianluo

README.md

@@ -70,9 +70,11 @@ Additional features not mentioned in the [report](https://arxiv.org/pdf/1702.021
 
   | GPU model  | Architecture |
   | ------------- | ------------- |
-  | TitanX (Maxwell/Pascal)  | sm_52  |
-  | Grid K520 (AWS g2.2xlarge)  | sm_30  |
-  | Tesla K80 (AWS p2.xlarge)   | sm_37  |
+  | TitanX (Maxwell/Pascal) | sm_52 |
+  | GTX 960M | sm_50 |
+  | GTX 1080 (Ti) | sm_61 |
+  | Grid K520 (AWS g2.2xlarge) | sm_30 |
+  | Tesla K80 (AWS p2.xlarge) | sm_37 |
 
   **Note**: You are welcome to contribute the settings on your end if you have made the code work properly on other GPUs.
 

experiments/scripts/train_faster_rcnn.sh

@@ -51,35 +51,35 @@ echo Logging output to "$LOG"
 
 set +x
 if [[ ! -z  ${EXTRA_ARGS_SLUG}  ]]; then
-    NET_FINAL=output/${NET}/${TRAIN_IMDB}/${EXTRA_ARGS_SLUG}/${NET}_faster_rcnn_iter_${ITERS}.pth
+  NET_FINAL=output/${NET}/${TRAIN_IMDB}/${EXTRA_ARGS_SLUG}/${NET}_faster_rcnn_iter_${ITERS}.ckpt
 else
-    NET_FINAL=output/${NET}/${TRAIN_IMDB}/default/${NET}_faster_rcnn_iter_${ITERS}.pth
+  NET_FINAL=output/${NET}/${TRAIN_IMDB}/default/${NET}_faster_rcnn_iter_${ITERS}.ckpt
 fi
 set -x
 
 if [ ! -f ${NET_FINAL}.index ]; then
-    if [[ ! -z  ${EXTRA_ARGS_SLUG}  ]]; then
-        CUDA_VISIBLE_DEVICES=${GPU_ID} time python ./tools/trainval_net.py \
-            --weight data/imagenet_weights/${NET}.pth \
-            --imdb ${TRAIN_IMDB} \
-            --imdbval ${TEST_IMDB} \
-            --iters ${ITERS} \
-            --cfg experiments/cfgs/${NET}.yml \
-            --tag ${EXTRA_ARGS_SLUG} \
-            --net ${NET} \
-            --set ANCHOR_SCALES ${ANCHORS} ANCHOR_RATIOS ${RATIOS} \
-            TRAIN.STEPSIZE ${STEPSIZE} ${EXTRA_ARGS}
-    else
-        CUDA_VISIBLE_DEVICES=${GPU_ID} time python ./tools/trainval_net.py \
-            --weight data/imagenet_weights/${NET}.pth \
-            --imdb ${TRAIN_IMDB} \
-            --imdbval ${TEST_IMDB} \
-            --iters ${ITERS} \
-            --cfg experiments/cfgs/${NET}.yml \
-            --net ${NET} \
-            --set ANCHOR_SCALES ${ANCHORS} ANCHOR_RATIOS ${RATIOS} \
-            TRAIN.STEPSIZE ${STEPSIZE} ${EXTRA_ARGS}
-    fi
+  if [[ ! -z  ${EXTRA_ARGS_SLUG}  ]]; then
+    CUDA_VISIBLE_DEVICES=${GPU_ID} time python ./tools/trainval_net.py \
+      --weight data/imagenet_weights/${NET}.ckpt \
+      --imdb ${TRAIN_IMDB} \
+      --imdbval ${TEST_IMDB} \
+      --iters ${ITERS} \
+      --cfg experiments/cfgs/${NET}.yml \
+      --tag ${EXTRA_ARGS_SLUG} \
+      --net ${NET} \
+      --set ANCHOR_SCALES ${ANCHORS} ANCHOR_RATIOS ${RATIOS} \
+      TRAIN.STEPSIZE ${STEPSIZE} ${EXTRA_ARGS}
+  else
+    CUDA_VISIBLE_DEVICES=${GPU_ID} time python ./tools/trainval_net.py \
+      --weight data/imagenet_weights/${NET}.ckpt \
+      --imdb ${TRAIN_IMDB} \
+      --imdbval ${TEST_IMDB} \
+      --iters ${ITERS} \
+      --cfg experiments/cfgs/${NET}.yml \
+      --net ${NET} \
+      --set ANCHOR_SCALES ${ANCHORS} ANCHOR_RATIOS ${RATIOS} \
+      TRAIN.STEPSIZE ${STEPSIZE} ${EXTRA_ARGS}
+  fi
 fi
 
 ./experiments/scripts/test_faster_rcnn.sh $@

lib/datasets/factory.py

@@ -22,6 +22,11 @@
     name = 'voc_{}_{}'.format(year, split)
     __sets[name] = (lambda split=split, year=year: pascal_voc(split, year))
 
+for year in ['2007', '2012']:
+  for split in ['train', 'val', 'trainval', 'test']:
+    name = 'voc_{}_{}_diff'.format(year, split)
+    __sets[name] = (lambda split=split, year=year: pascal_voc(split, year, use_diff=True))
+
 # Set up coco_2014_<split>
 for year in ['2014']:
   for split in ['train', 'val', 'minival', 'valminusminival', 'trainval']:

lib/datasets/pascal_voc.py

@@ -23,12 +23,14 @@
 
 
 class pascal_voc(imdb):
-  def __init__(self, image_set, year, devkit_path=None):
-    imdb.__init__(self, 'voc_' + year + '_' + image_set)
+  def __init__(self, image_set, year, use_diff=False):
+    name = 'voc_' + year + '_' + image_set
+    if use_diff:
+      name += '_diff'
+    imdb.__init__(self, name)
     self._year = year
     self._image_set = image_set
-    self._devkit_path = self._get_default_path() if devkit_path is None \
-      else devkit_path
+    self._devkit_path = self._get_default_path()
     self._data_path = os.path.join(self._devkit_path, 'VOC' + self._year)
     self._classes = ('__background__',  # always index 0
                      'aeroplane', 'bicycle', 'bird', 'boat',
@@ -47,7 +49,7 @@ def __init__(self, image_set, year, devkit_path=None):
     # PASCAL specific config options
     self.config = {'cleanup': True,
                    'use_salt': True,
-                   'use_diff': False,
+                   'use_diff': use_diff,
                    'matlab_eval': False,
                    'rpn_file': None}
 
@@ -240,7 +242,7 @@ def _do_python_eval(self, output_dir='output'):
       filename = self._get_voc_results_file_template().format(cls)
       rec, prec, ap = voc_eval(
         filename, annopath, imagesetfile, cls, cachedir, ovthresh=0.5,
-        use_07_metric=use_07_metric)
+        use_07_metric=use_07_metric, use_diff=self.config['use_diff'])
       aps += [ap]
       print(('AP for {} = {:.4f}'.format(cls, ap)))
       with open(os.path.join(output_dir, cls + '_pr.pkl'), 'wb') as f:

lib/datasets/voc_eval.py

@@ -72,7 +72,8 @@ def voc_eval(detpath,
              classname,
              cachedir,
              ovthresh=0.5,
-             use_07_metric=False):
+             use_07_metric=False,
+             use_diff=False):
   """rec, prec, ap = voc_eval(detpath,
                               annopath,
                               imagesetfile,
@@ -133,7 +134,10 @@ def voc_eval(detpath,
   for imagename in imagenames:
     R = [obj for obj in recs[imagename] if obj['name'] == classname]
     bbox = np.array([x['bbox'] for x in R])
-    difficult = np.array([x['difficult'] for x in R]).astype(np.bool)
+    if use_diff:
+      difficult = np.array([False for x in R]).astype(np.bool)
+    else:
+      difficult = np.array([x['difficult'] for x in R]).astype(np.bool)
     det = [False] * len(R)
     npos = npos + sum(~difficult)
     class_recs[imagename] = {'bbox': bbox,

lib/nets/mobilenet_v1.py

@@ -249,37 +249,6 @@ def _head_to_tail(self, pool5, is_training, reuse=False):
       fc7 = tf.reduce_mean(fc7, axis=[1, 2])
     return fc7
 
-  def _build_network(self, is_training=True):
-    # select initializers
-    if cfg.TRAIN.TRUNCATED:
-      initializer = tf.truncated_normal_initializer(mean=0.0, stddev=0.01)
-      initializer_bbox = tf.truncated_normal_initializer(mean=0.0, stddev=0.001)
-    else:
-      initializer = tf.random_normal_initializer(mean=0.0, stddev=0.01)
-      initializer_bbox = tf.random_normal_initializer(mean=0.0, stddev=0.001)
-    
-    net_conv = self._image_to_head(is_training)
-    with tf.variable_scope(self._scope, self._scope):
-      # build the anchors for the image
-      self._anchor_component()
-      # region proposal network
-      rois = self._region_proposal(net_conv, is_training, initializer)
-      # region of interest pooling
-      if cfg.POOLING_MODE == 'crop':
-        pool5 = self._crop_pool_layer(net_conv, rois, "pool5")
-      else:
-        raise NotImplementedError
-
-    fc7 = self._head_to_tail(pool5, is_training)
-    with tf.variable_scope(self._scope, self._scope):
-      # region classification
-      cls_prob, bbox_pred = self._region_classification(fc7, is_training, 
-                                                        initializer, initializer_bbox)
-      
-    self._score_summaries.update(self._predictions)
-
-    return rois, cls_prob, bbox_pred
-
   def get_variables_to_restore(self, variables, var_keep_dic):
     variables_to_restore = []