update demo

examples/ConvolutionalPoseMachines/README.md

 # Convolutional Pose Machines
 
-A script to load and run pretrained CPM model released by Shih-En. The original code in caffe is [here](https://github.com/shihenw/convolutional-pose-machines-release).
+A script to load and run pre-trained CPM model released by Shih-En. The original code in caffe is [here](https://github.com/shihenw/convolutional-pose-machines-release).
 
 ## Usage:
+
+Prepare the model:
 ```
-# download the relased caffe model:
+# download the released caffe model:
 wget http://pearl.vasc.ri.cmu.edu/caffe_model_github/model/_trained_MPI/pose_iter_320000.caffemodel
 wget https://github.com/shihenw/convolutional-pose-machines-release/raw/master/model/_trained_MPI/pose_deploy_resize.prototxt
 # convert the model to a dict:
 python -m tensorpack.utils.loadcaffe pose_deploy_resize.prototxt pose_iter_320000.caffemodel CPM-original.npy
-# run it on a test image, and produce output.jpg
+```
+
+Or you can download the converted model from [model zoo](https://drive.google.com/open?id=0B9IPQTvr2BBkRU8zM2w2ZGh3eU0)
+
+Run it on an image, and produce `output.jpg`:
+```
 python load-cpm.py --load CPM-original.npy --input test.jpg
 ```
+Input image will get resized to 368x368. Note that this CPM comes without person detection, so the
+person has to be in the center of the image (and not too small).
+
+![demo](demo.jpg)
+
+For a real-time CPM application in tensorpack, check out [Stereo Pose Machines](https://github.com/ppwwyyxx/Stereo-Pose-Machines).

examples/ConvolutionalPoseMachines/load-cpm.py

@@ -4,18 +4,14 @@
 # Author: Yuxin Wu <ppwwyyxxc@gmail.com>
 
 import cv2
-import tqdm
 import tensorflow as tf
 import numpy as np
-import os, argparse
-import matplotlib.pyplot as plt
+import argparse
 
 from tensorpack import *
 from tensorpack.utils.argtools import memoized
-from tensorpack.tfutils.symbolic_functions import *
-from tensorpack.tfutils.summary import *
-from tensorpack.dataflow.dataset import ILSVRCMeta
 
+import matplotlib.pyplot as plt
 _CM = plt.get_cmap('jet')
 def colorize(img, heatmap):
     """ img: bgr, [0,255]
@@ -105,6 +101,7 @@ def run_test(model_path, img_file):
     ))
 
     im = cv2.imread(img_file, cv2.IMREAD_COLOR).astype('float32')
+    im = cv2.resize(im, (368,368))
     out = predict_func([[im]])[0][0]
     hm = out[:,:,:14].sum(axis=2)
     viz = colorize(im, hm)

examples/README.md

@@ -12,7 +12,8 @@ Training examples with __reproducible__ and meaningful performance.
 + [InceptionV3 with 74% accuracy (similar to the official code)](Inception/inceptionv3.py)
 + [Fully-convolutional Network for Holistically-Nested Edge Detection(HED)](HED)
 + [Spatial Transformer Networks on MNIST addition](SpatialTransformer)
-+ Load a pretrained [AlexNet](load-alexnet.py) or [VGG16](load-vgg16.py) model.
++ Load a pre-trained [AlexNet](load-alexnet.py) or [VGG16](load-vgg16.py) model.
++ Load a pre-trained [Convolutional Pose Machines](ConvolutionalPoseMachines/).
 
 ## Reinforcement Learning:
 + [Deep Q-Network(DQN) variants on Atari games](Atari2600)