Release a COCO Faster R-CNN model

Author: rbgirshick

.gitignore

@@ -1,3 +1,5 @@
 *.pyc
 .ipynb_checkpoints
 lib/build
+lib/pycocotools/_mask.c
+lib/pycocotools/_mask.so

lib/pycocotools/coco.py

@@ -106,8 +106,9 @@ def createIndex(self):
             for cat in self.dataset['categories']:
                 cats[cat['id']] = cat
             catToImgs = {cat['id']: [] for cat in self.dataset['categories']}
-            for ann in self.dataset['annotations']:
-                catToImgs[ann['category_id']] += [ann['image_id']]
+            if 'annotations' in self.dataset:
+                for ann in self.dataset['annotations']:
+                    catToImgs[ann['category_id']] += [ann['image_id']]
 
         print 'index created!'
 

models/README.md

@@ -1,7 +1,50 @@
-Prototxt files that define models and solvers.
+## Model Zoo
 
-Three models are defined, with some variations of each to support experiments
-in the paper.
- - Caffenet (model **S**)
- - VGG_CNN_M_1024 (model **M**)
- - VGG16 (model **L**)
+### COCO Faster R-CNN VGG-16 trained using end-to-end
+
+Model URL: www.cs.berkeley.edu/~rbg/faster-rcnn-data/coco_vgg16_faster_rcnn_final.caffemodel
+
+Training command:
+```
+tools/train_net.py \
+    --gpu 0 \
+    --solver ./models/coco/VGG16/faster_rcnn_end2end/solver.prototxt \
+    --weights data/imagenet_models/VGG16.v2.caffemodel \
+    --imdb coco_2014_train+coco_2014_valminusminival \
+    --iters 490000 \
+    --cfg ./experiments/cfgs/faster_rcnn_end2end.yml
+```
+
+`py-faster-rcnn` commit: 68eec95
+
+test-dev2015 results
+```
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.242
+ Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.453
+ Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.235
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.077
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.264
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.371
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.238
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.340
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.346
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.120
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.385
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.544
+```
+
+test-standard2015 results
+```
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.242
+ Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.453
+ Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.234
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.072
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.264
+ Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.369
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.238
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.341
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.347
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.115
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.389
+ Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.544
+```