fix the bug about the softmaxlayer

[fengshikun]

When I execute the code below:

# Import the converted model's class
from mynet import MyNet

# Create an instance, passing in the input data
net = MyNet({'data':my_input_data})

It will cause the tensorflow error:

assert all(arg.dtype in dtype_hierarchy for arg in [start, limit, delta])
AssertionError

Change the kaffe/tensorflow/network.py line 215: return tf.nn.softmax(input, name) to return tf.nn.softmax(input, name=name) fix this bug

[chiggum]

tf.nn.softmax(logits=input, name=name)