训练过程中，当训练完毕

1. mon_sess.should_stop()=True时报错

   training and sparisifying

   while not mon_sess.should_stop():
   mon_sess.run(train_op)
   2.错误定位在cifar10_train.py中112号处会发生错cifprune_op = dt.prune(tf.get_collection('is prune'))
   3.错误描述
   2019-10-11 17:37:10.286919: W tensorflow/core/framework/op_kernel.cc:1490] Unknown: KeyError: b'g2b2c0/BatchNorm/gamma'
   Traceback (most recent call last):

File "/home/ubuntu/anaconda3/lib/python3.6/site-packages/tensorflow/python/ops/script_ops.py", line 209, in call
ret = func(*args)

File "/home/ubuntu/Downloads/batchnorm_prune-master/src/python/cifar10/batchnorm_prune/Transformer.py", line 111, in
collection.append(tf.py_func(lambda a, b: self.flow.get_channel_pruned(a, b),

File "/home/ubuntu/Downloads/batchnorm_prune-master/src/python/cifar10/batchnorm_prune/Graph.py", line 155, in get_channel_pruned
print(self.edgename_map[varname])

KeyError: b'g2b2c0/BatchNorm/gamma'

本人环境：
python3.6
tensorflow1.14
请作者看看，谢谢

1. Need to remove dependency on explicitly computing the pruning weights. (Current is hard-coding)

2. Need to properly handle input.op.name in compressed mode. (Current is hard-coding)