Pretrained model for FasterRCNN/ShuffleNet

examples/FasterRCNN/README.md

@@ -37,6 +37,11 @@ To predict on an image (and show output in a window):
 ./train.py --predict input.jpg
 ```
 
+To evaluate the performance (pretrained models can be downloaded in [model zoo](https://drive.google.com/open?id=1J0xuDAuyOWiuJRm2LfGoz5PUv9_dKuxq):
+```
+./train.py --evaluate output.json --load /path/to/model
+```
+
 ## Results
 
 Mean Average Precision @IoU=0.50:0.95:

examples/ShuffleNet/README.md

@@ -21,3 +21,8 @@ Train (takes 24 hours on 8 Maxwell TitanX):
 ```bash
 ./shufflenet.py --data /path/to/ilsvrc/
 ```
+
+Eval the [pretrained model](https://drive.google.com/open?id=1Q1C2BCkbOK2HfhUB0Yfn_W_F68bqnA6y):
+```
+./shufflenet.py --eval --data /path/to/ilsvrc --load /path/to/model
+```

examples/ShuffleNet/shufflenet.py

@@ -21,7 +21,8 @@ from tensorpack.tfutils.scope_utils import under_name_scope
 from tensorpack.utils.gpu import get_nr_gpu
 
 from imagenet_utils import (
-    fbresnet_augmentor, get_imagenet_dataflow, ImageNetModel, GoogleNetResize)
+    fbresnet_augmentor, get_imagenet_dataflow,
+    ImageNetModel, GoogleNetResize, eval_on_ILSVRC12)
 
 TOTAL_BATCH_SIZE = 256
 
@@ -178,6 +179,7 @@ if __name__ == '__main__':
     parser.add_argument('--gpu', help='comma separated list of GPU(s) to use.')
     parser.add_argument('--data', help='ILSVRC dataset dir')
     parser.add_argument('--load', help='load model')
+    parser.add_argument('--eval', action='store_true')
     parser.add_argument('--flops', action='store_true', help='print flops and exit')
     args = parser.parse_args()
 
@@ -186,7 +188,11 @@ if __name__ == '__main__':
 
     model = Model()
 
-    if args.flops:
+    if args.eval:
+        batch = 128    # something that can run on one gpu
+        ds = get_data('val', batch)
+        eval_on_ILSVRC12(model, get_model_loader(args.load), ds)
+    elif args.flops:
         # manually build the graph with batch=1
         input_desc = [
             InputDesc(tf.float32, [1, 224, 224, 3], 'input'),