add resnet101

experiments/cfgs/res101.yml

@@ -0,0 +1,14 @@
+EXP_DIR: res101
+TRAIN:
+  HAS_RPN: True
+  IMS_PER_BATCH: 1
+  BBOX_NORMALIZE_TARGETS_PRECOMPUTED: True
+  RPN_POSITIVE_OVERLAP: 0.7
+  RPN_BATCHSIZE: 256
+  PROPOSAL_METHOD: gt
+  BG_THRESH_LO: 0.0
+  DISPLAY: 20
+  BATCH_SIZE: 256
+TEST:
+  HAS_RPN: True
+POOLING_MODE: crop

experiments/scripts/test_faster_rcnn.sh

@@ -0,0 +1,69 @@
+#!/bin/bash
+
+set -x
+set -e
+
+export PYTHONUNBUFFERED="True"
+
+GPU_ID=$1
+DATASET=$2
+NET=$3
+
+array=( $@ )
+len=${#array[@]}
+EXTRA_ARGS=${array[@]:3:$len}
+EXTRA_ARGS_SLUG=${EXTRA_ARGS// /_}
+
+case ${DATASET} in
+  pascal_voc)
+    TRAIN_IMDB="voc_2007_trainval"
+    TEST_IMDB="voc_2007_test"
+    ITERS=70000
+    ;;
+  pascal_voc_0712)
+    TRAIN_IMDB="voc_2007_trainval+voc_2012_trainval"
+    TEST_IMDB="voc_2007_test"
+    ITERS=25000
+    ;;
+  coco)
+    TRAIN_IMDB="coco_2014_train+coco_2014_valminusminival"
+    TEST_IMDB="coco_2014_minival"
+    ITERS=490000
+    ;;
+  *)
+    echo "No dataset given"
+    exit
+    ;;
+esac
+
+LOG="experiments/logs/test_${NET}_${TRAIN_IMDB}_${EXTRA_ARGS_SLUG}.txt.`date +'%Y-%m-%d_%H-%M-%S'`"
+exec &> >(tee -a "$LOG")
+echo Logging output to "$LOG"
+
+set +x
+if [[ ! -z  ${EXTRA_ARGS_SLUG}  ]]; then
+  NET_FINAL=output/${NET}/${TRAIN_IMDB}/${EXTRA_ARGS_SLUG}/${NET}_faster_rcnn_iter_${ITERS}.ckpt
+else
+  NET_FINAL=output/${NET}/${TRAIN_IMDB}/default/${NET}_faster_rcnn_iter_${ITERS}.ckpt
+fi
+set -x
+
+if [[ ! -z  ${EXTRA_ARGS_SLUG}  ]]; then
+  CUDA_VISIBLE_DEVICES=${GPU_ID} time python ./tools/test_net.py \
+    --imdb ${TEST_IMDB} \
+    --weight data/imagenet_weights/${NET}.weights \
+    --model ${NET_FINAL} \
+    --cfg experiments/cfgs/${NET}.yml \
+    --tag ${EXTRA_ARGS_SLUG} \
+    --net ${NET} \
+    --set ${EXTRA_ARGS}
+else
+  CUDA_VISIBLE_DEVICES=${GPU_ID} time python ./tools/test_net.py \
+    --imdb ${TEST_IMDB} \
+    --weight data/imagenet_weights/${NET}.weights \
+    --model ${NET_FINAL} \
+    --cfg experiments/cfgs/${NET}.yml \
+    --net ${NET} \
+    --set ${EXTRA_ARGS}
+fi
+

experiments/scripts/train_faster_rcnn.sh

@@ -0,0 +1,77 @@
+#!/bin/bash
+
+set -x
+set -e
+
+export PYTHONUNBUFFERED="True"
+
+GPU_ID=$1
+DATASET=$2
+NET=$3
+
+array=( $@ )
+len=${#array[@]}
+EXTRA_ARGS=${array[@]:3:$len}
+EXTRA_ARGS_SLUG=${EXTRA_ARGS// /_}
+
+case ${DATASET} in
+  pascal_voc)
+    TRAIN_IMDB="voc_2007_trainval"
+    TEST_IMDB="voc_2007_test"
+    STEPSIZE=50000
+    ITERS=70000
+    ;;
+  pascal_voc_0712)
+    TRAIN_IMDB="voc_2007_trainval+voc_2012_trainval"
+    TEST_IMDB="voc_2007_test"
+    STEPSIZE=50000
+    ITERS=70000
+    ;;
+  coco)
+    TRAIN_IMDB="coco_2014_train+coco_2014_valminusminival"
+    TEST_IMDB="coco_2014_minival"
+    STEPSIZE=350000
+    ITERS=490000
+    ;;
+  *)
+    echo "No dataset given"
+    exit
+    ;;
+esac
+
+LOG="experiments/logs/${NET}_${TRAIN_IMDB}_${EXTRA_ARGS_SLUG}_${NET}.txt.`date +'%Y-%m-%d_%H-%M-%S'`"
+exec &> >(tee -a "$LOG")
+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}.ckpt
+else
+    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}.ckpt \
+            --imdb ${TRAIN_IMDB} \
+            --imdbval ${TEST_IMDB} \
+            --iters ${ITERS} \
+            --cfg experiments/cfgs/${NET}.yml \
+            --tag ${EXTRA_ARGS_SLUG} \
+            --net ${NET} \
+            --set 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 TRAIN.STEPSIZE ${STEPSIZE} ${EXTRA_ARGS}
+    fi
+fi
+
+./experiments/scripts/test_faster_rcnn.sh $@

experiments/scripts/vgg16.sh

@@ -47,7 +47,7 @@ set -x
 if [ ! -f ${NET_FINAL}.index ]; then
     if [[ ! -z  ${EXTRA_ARGS_SLUG}  ]]; then
         CUDA_VISIBLE_DEVICES=${GPU_ID} time python ./tools/trainval_vgg16_net.py \
-            --weight data/imagenet_weights/vgg16.weights \
+            --weight data/imagenet_weights/vgg_16.ckpt \
             --imdb ${TRAIN_IMDB} \
             --imdbval ${TEST_IMDB} \
             --iters ${ITERS} \
@@ -56,7 +56,7 @@ if [ ! -f ${NET_FINAL}.index ]; then
             --set TRAIN.STEPSIZE ${STEPSIZE} ${EXTRA_ARGS}
     else
         CUDA_VISIBLE_DEVICES=${GPU_ID} time python ./tools/trainval_vgg16_net.py \
-            --weight data/imagenet_weights/vgg16.weights \
+            --weight data/imagenet_weights/vgg_16.ckpt \
             --imdb ${TRAIN_IMDB} \
             --imdbval ${TEST_IMDB} \
             --iters ${ITERS} \

lib/datasets/pascal_voc.py

@@ -103,7 +103,6 @@ def gt_roidb(self):
     cache_file = os.path.join(self.cache_path, self.name + '_gt_roidb.pkl')
     if os.path.exists(cache_file):
       with open(cache_file, 'rb') as fid:
-        roidb = pickle.load(fid)
         try:
           roidb = pickle.load(fid)
         except:

lib/layer_utils/proposal_layer.py

@@ -25,7 +25,6 @@ def proposal_layer(rpn_cls_prob, rpn_bbox_pred, im_info, cfg_key, _feat_stride,
   scales = np.array(anchor_scales)
   num_anchors = scales.shape[0] * 3
   im_info = im_info[0]
-
   # Get the scores and bounding boxes
   scores = rpn_cls_prob[:, :, :, num_anchors:]
   rpn_bbox_pred = rpn_bbox_pred.reshape((-1, 4))

lib/layer_utils/snippets.py

@@ -3,6 +3,9 @@
 # Licensed under The MIT License [see LICENSE for details]
 # Written by Xinlei Chen
 # --------------------------------------------------------
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
 
 import numpy as np
 import numpy.random as npr

lib/model/bbox_transform.py

@@ -37,7 +37,6 @@ def bbox_transform_inv(boxes, deltas):
     return np.zeros((0, deltas.shape[1]), dtype=deltas.dtype)
 
   boxes = boxes.astype(deltas.dtype, copy=False)
-
   widths = boxes[:, 2] - boxes[:, 0] + 1.0
   heights = boxes[:, 3] - boxes[:, 1] + 1.0
   ctr_x = boxes[:, 0] + 0.5 * widths
@@ -47,7 +46,7 @@ def bbox_transform_inv(boxes, deltas):
   dy = deltas[:, 1::4]
   dw = deltas[:, 2::4]
   dh = deltas[:, 3::4]
-
+  
   pred_ctr_x = dx * widths[:, np.newaxis] + ctr_x[:, np.newaxis]
   pred_ctr_y = dy * heights[:, np.newaxis] + ctr_y[:, np.newaxis]
   pred_w = np.exp(dw) * widths[:, np.newaxis]

lib/model/config.py

@@ -97,7 +97,7 @@
 
 # solver.prototxt specifies the snapshot path prefix, this adds an optional
 # infix to yield the path: <prefix>[_<infix>]_iters_XYZ.caffemodel
-__C.TRAIN.SNAPSHOT_PREFIX = 'vgg16_faster_rcnn'
+__C.TRAIN.SNAPSHOT_PREFIX = 'res101_faster_rcnn'
 # __C.TRAIN.SNAPSHOT_INFIX = ''
 
 # Use a prefetch thread in roi_data_layer.layer

lib/model/test.py

@@ -95,7 +95,9 @@ def im_detect(sess, net, im):
   _, scores, bbox_pred, rois = net.test_image(sess, blobs['data'], blobs['im_info'])
 
   boxes = rois[:, 1:5] / im_scales[0]
-
+  # print(scores.shape, bbox_pred.shape, rois.shape, boxes.shape)
+  scores = np.reshape(scores, [scores.shape[0], -1])
+  bbox_pred = np.reshape(bbox_pred, [bbox_pred.shape[0], -1])
   if cfg.TEST.BBOX_REG:
     # Apply bounding-box regression deltas
     box_deltas = bbox_pred

lib/model/train_val.py

@@ -22,6 +22,7 @@
 import time
 
 import tensorflow as tf
+from tensorflow.python import pywrap_tensorflow
 
 
 class SolverWrapper(object):
@@ -79,6 +80,17 @@ def snapshot(self, sess, iter):
 
     return filename, nfilename
 
+  def get_variables_in_checkpoint_file(self, file_name):
+    try:
+      reader = pywrap_tensorflow.NewCheckpointReader(file_name)
+      var_to_shape_map = reader.get_variable_to_shape_map()
+      return var_to_shape_map 
+    except Exception as e:  # pylint: disable=broad-except
+      print(str(e))
+      if "corrupted compressed block contents" in str(e):
+        print("It's likely that your checkpoint file has been compressed "
+              "with SNAPPY.")
+
   def train_model(self, sess, max_iters):
     # Build data layers for both training and validation set
     self.data_layer = RoIDataLayer(self.roidb, self.imdb.num_classes)
@@ -99,11 +111,11 @@ def train_model(self, sess, max_iters):
                                             tag='default', anchor_scales=anchors)
       # Define the loss
       loss = layers['total_loss']
-
       # Set learning rate and momentum
       lr = tf.Variable(cfg.TRAIN.LEARNING_RATE, trainable=False)
       momentum = cfg.TRAIN.MOMENTUM
       self.optimizer = tf.train.MomentumOptimizer(lr, momentum)
+
       # Compute the gradients wrt the loss
       gvs = self.optimizer.compute_gradients(loss)
       # Double the gradient of the bias if set
@@ -148,10 +160,17 @@ def train_model(self, sess, max_iters):
       # Fresh train directly from VGG weights
       print('Loading initial model weights from {:s}'.format(self.pretrained_model))
       variables = tf.global_variables()
+
       # Only initialize the variables that were not initialized when the graph was built
-      for vbs in self.net._initialized:
-        variables.remove(vbs)
       sess.run(tf.variables_initializer(variables, name='init'))
+      var_keep_dic = self.get_variables_in_checkpoint_file(self.pretrained_model)
+      variables_to_restore = []
+      # print(var_keep_dic)
+      for v in variables:
+          if v.name.split(':')[0] in var_keep_dic:
+              variables_to_restore.append(v)
+      restorer = tf.train.Saver(variables_to_restore)
+      restorer.restore(sess, self.pretrained_model)
       print('Loaded.')
       sess.run(tf.assign(lr, cfg.TRAIN.LEARNING_RATE))
       last_snapshot_iter = 0