shuffle where data are stored

data/.gitignore

+selective_search_data*
+imagenet_models*
+fast_rcnn_models*
+VOCdevkit*
+cache

data/README.md

+This directory holds:
+- Precomputed object proposals
+- Caffe models pretrained on ImageNet
+- Fast R-CNN models
+- Symlinks to datasets
+
+To download precomputed selective search proposals, run:
+
+```
+./tools/scripts/fetch_selective_search_data.sh
+```
+
+This script will populate `data/selective_search_data`.
+
+To download Caffe models pretrained on ImageNet, run:
+
+```
+./tools/scripts/fetch_imagenet_models.sh
+```
+
+This script will populate `data/imagenet_models`.
+
+To download Fast R-CNN models, run:
+
+```
+./tools/scripts/fetch_fast_rcnn_models.sh
+```
+
+This script will populate `data/fast_rcnn_models`.
+
+In order to train and test with PASCAL VOC, you will need to establish symlinks.
+From the `data` directory (`cd data`):
+
+```
+# For VOC 2007
+ln -s /your/path/to/VOC2007/VOCdevkit VOCdevkit2007
+
+# For VOC 2012
+ln -s /your/path/to/VOC2012/VOCdevkit VOCdevkit2012
+```
+
+Since you'll likely be experimenting with multiple installs of Fast R-CNN in
+parallel, you'll probably want to keep all of this data in a shared place and
+use symlinks. On my system I create the following symlinks inside `data`:
+
+```
+# data/cache holds various outputs created by the datasets package
+ln -s /data/fast_rcnn_shared/cache
+
+# move the imagenet_models to shared location and symlink to them
+ln -s /data/fast_rcnn_shared/imagenet_models
+
+# move the selective search data to a shared location and symlink to them
+ln -s /data/fast_rcnn_shared/selective_search_data
+
+ln -s /data/VOC2007/VOCdevkit VOCdevkit2007
+ln -s /data/VOC2012/VOCdevkit VOCdevkit2012
+```

datasets/imdb.py

@@ -62,7 +62,7 @@ class imdb(object):
     def cache_path(self):
         return os.path.abspath(os.path.join(
             os.path.dirname(__file__),
-            'cache'))
+            '..', 'data', 'cache'))
 
     @property
     def num_images(self):

datasets/pascal_voc.py

@@ -81,7 +81,7 @@ class pascal_voc(datasets.imdb):
         """
         path = os.path.abspath(os.path.join(
                     os.path.dirname(__file__),
-                    '..', 'datasets', 'VOCdevkit' + self._year))
+                    '..', 'data', 'VOCdevkit' + self._year))
         return path
 
     def gt_roidb(self):
@@ -142,10 +142,11 @@ class pascal_voc(datasets.imdb):
         return a
 
     def _load_selective_search_roidb(self, gt_roidb):
-        filename = os.path.join(self.cache_path, 'selective_search_data',
-                                self.name + '.mat')
+        filename = os.path.abspath(os.path.join(self.cache_path, '..',
+                                                'selective_search_data',
+                                                self.name + '.mat'))
         assert os.path.exists(filename), \
-                'Selective search data not found at: {}'.format(filename)
+               'Selective search data not found at: {}'.format(filename)
         raw_data = sio.loadmat(filename)
 
         num_images = raw_data['boxes'].ravel().shape[0]

tools/scripts/fetch_imagenet_models.sh

+#!/bin/bash
+
+DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )/../../data" && pwd )"
+cd $DIR
+
+FILE=imagenet_models.tgz
+URL=https://www.dropbox.com/s/22it3se7e4zi2mi/imagenet_models.tgz?dl=0
+CHECKSUM=66dbfdf04297e1e68b49f168a2ccc59d
+
+if [ -f $FILE ]; then
+  echo "File already exists. Checking md5..."
+  os=`uname -s`
+  if [ "$os" = "Linux" ]; then
+    checksum=`md5sum $FILE | awk '{ print $1 }'`
+  elif [ "$os" = "Darwin" ]; then
+    checksum=`cat $FILE | md5`
+  fi
+  if [ "$checksum" = "$CHECKSUM" ]; then
+    echo "Checksum is correct. No need to download."
+    exit 0
+  else
+    echo "Checksum is incorrect. Need to download again."
+  fi
+fi
+
+echo "Downloading pretrained ImageNet models (1G)..."
+
+wget $URL -O $FILE
+
+echo "Unzipping..."
+
+tar zxvf $FILE
+
+echo "Done. Please run this command again to verify that checksum = $CHECKSUM."

tools/scripts/fetch_selective_search_data.sh

+#!/bin/bash
+
+DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )/../../data" && pwd )"
+cd $DIR
+
+FILE=selective_search_data.tgz
+URL=https://www.dropbox.com/s/g1z7iolrtxdo56c/selective_search_data.tgz?dl=0
+CHECKSUM=7cc85568609e1ac645f102c37eb376c3
+
+if [ -f $FILE ]; then
+  echo "File already exists. Checking md5..."
+  os=`uname -s`
+  if [ "$os" = "Linux" ]; then
+    checksum=`md5sum $FILE | awk '{ print $1 }'`
+  elif [ "$os" = "Darwin" ]; then
+    checksum=`cat $FILE | md5`
+  fi
+  if [ "$checksum" = "$CHECKSUM" ]; then
+    echo "Checksum is correct. No need to download."
+    exit 0
+  else
+    echo "Checksum is incorrect. Need to download again."
+  fi
+fi
+
+echo "Downloading precomputed selective search boxes (1.8G)..."
+
+wget $URL -O $FILE
+
+echo "Unzipping..."
+
+tar zxvf $FILE
+
+echo "Done. Please run this command again to verify that checksum = $CHECKSUM."