You can do [n.name for n in tf.get_default_graph().as_graph_def().node]. Also, if you are prototyping in an IPython notebook, ...
確定! 回上一頁