misc.

README.md

 # tensorpack
 Neural Network Toolbox on TensorFlow
 
-In development. API might change a bit. See [examples](https://github.com/ppwwyyxx/tensorpack/tree/master/examples) to learn.
+In development. API might change a bit.
+See [examples](https://github.com/ppwwyyxx/tensorpack/tree/master/examples) to learn.
 
 ## Features:
-+ Scoped abstraction of common models.
-+ Use `Dataflow` to define data preprocessing in pure Python.
-+ Callbacks systems to control training.
-+ Training and testing are described together. Just need to follow the conventions to setup stuffs.
-+ Write summary easier for tensorboard.
+
+Focused on modularity:
+
++ Models has Scoped abstraction of common models.
++ Dataflow defines data preprocessing in pure Python.
++ Callbacks systems to control training behavior.

examples/load-alexnet.py

@@ -21,6 +21,12 @@ from tensorpack.callbacks import *
 from tensorpack.dataflow import *
 from tensorpack.dataflow.dataset import ILSVRCMeta
 
+"""
+Usage:
+    python2 -m tensorpack.utils.loadcaffe PATH/TO/CAFFE/{deploy.prototxt,bvlc_alexnet.caffemodel} alexnet.npy
+    ./load-alexnet.py --load alexnet.npy --input cat.png
+"""
+
 BATCH_SIZE = 10
 MIN_AFTER_DEQUEUE = 500
 CAPACITY = MIN_AFTER_DEQUEUE + 3 * BATCH_SIZE

examples/load-vgg16.py

@@ -23,8 +23,8 @@ from tensorpack.dataflow.dataset import ILSVRCMeta
 
 """
 Usage:
-    python2 -m tensorpack.utils.loadcaffe PATH/TO/models/VGG/{VGG_ILSVRC_16_layers_deploy.prototxt,VGG_ILSVRC_16_layers.caffemodel} vgg16.npy
-    ./load_vgg16.py --load vgg16.npy --input cat.png
+    python2 -m tensorpack.utils.loadcaffe PATH/TO/VGG/{VGG_ILSVRC_16_layers_deploy.prototxt,VGG_ILSVRC_16_layers.caffemodel} vgg16.npy
+    ./load-vgg16.py --load vgg16.npy --input cat.png
 """
 
 class Model(ModelDesc):

tensorpack/dataflow/dataset/atari.py

@@ -168,10 +168,8 @@ class AtariPlayer(RLEnvironment):
 
     def get_stat(self):
         try:
-            print self.stats
             return {'avg_score': np.mean(self.stats['score']),
-                    'max_score': float(np.max(self.stats['score']))
-                    }
+                    'max_score': float(np.max(self.stats['score'])) }
         except ValueError:
             return {}