这个问题最初发布在 tflearn github repo 上，但我在那里没有运气： https ://github.com/tflearn/tflearn/issues/682
我正在尝试从自动编码器中保存代表中间层的编码器模型。使用 MNIST 示例，当我运行此处找到的脚本时：
https://github.com/tflearn/tflearn/blob/master/examples/images/autoencoder.py
然后尝试使用保存 encoding_model
encoding_model = tflearn.DNN(encoder, session=model.session)
encoding_model.save('encoder.tfl')
我收到以下错误消息：
回溯（最后一次调用）：文件“”，第 1 行，在文件“/usr/local/lib/python2.7/dist-packages/tflearn/models/dnn.py”，第 260 行，保存 self.trainer .save(model_file) 文件“/usr/local/lib/python2.7/dist-packages/tflearn/helpers/trainer.py”，第 376 行，保存 self.saver.save(self.session, model_file, global_step= global_step) 文件“/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/saver.py”，第 1363 行，保存 {self.saver_def.filename_tensor_name: checkpoint_file}) 文件“/usr/ local/lib/python2.7/dist-packages/tensorflow/python/client/session.py”，第 767 行，运行 run_metadata_ptr) 文件“/usr/local/lib/python2.7/dist-packages/tensorflow/python /client/session.py"，第 965 行，在 _run feed_dict_string、options、run_metadata) 文件中"/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py”，第 1015 行，在 _do_run 目标列表，选项，run_metadata）文件“/usr/local/lib/python2.7/ dist-packages/tensorflow/python/client/session.py”，第 1035 行，在 _do_call raise type(e)(node_def, op, message) tensorflow.python.framework.errors_impl.FailedPreconditionError: Attempting to use uninitialized value Global_Step_1 [[节点：Global_Step_1/_96 = _SendT=DT_FLOAT, client_terminated=false, recv_device="/job:localhost/replica:0/task:0/cpu:0", send_device="/job:localhost/replica:0/task:0 /gpu:0", send_device_incarnation=1, tensor_name="edge_31_Global_Step_1", _device="/job:localhost/replica:0/task:0/gpu:0"]]在 _do_run target_list, options, run_metadata) 文件“/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py”，第 1035 行，在 _do_call raise type(e)(node_def, op , 消息) tensorflow.python.framework.errors_impl.FailedPreconditionError: Attempting to use uninitialized value Global_Step_1 [[Node: Global_Step_1/_96 = _SendT=DT_FLOAT, client_terminated=false, recv_device="/job:localhost/replica:0/task:0 /cpu:0", send_device="/job:localhost/replica:0/task:0/gpu:0", send_device_incarnation=1, tensor_name="edge_31_Global_Step_1", _device="/job:localhost/replica:0/task :0/gpu:0"]]在 _do_run target_list, options, run_metadata) 文件“/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py”，第 1035 行，在 _do_call raise type(e)(node_def, op , 消息) tensorflow.python.framework.errors_impl.FailedPreconditionError: Attempting to use uninitialized value Global_Step_1 [[Node: Global_Step_1/_96 = _SendT=DT_FLOAT, client_terminated=false, recv_device="/job:localhost/replica:0/task:0 /cpu:0", send_device="/job:localhost/replica:0/task:0/gpu:0", send_device_incarnation=1, tensor_name="edge_31_Global_Step_1", _device="/job:localhost/replica:0/task :0/gpu:0"]]消息) tensorflow.python.framework.errors_impl.FailedPreconditionError: Attempting to use uninitialized value Global_Step_1 [[Node: Global_Step_1/_96 = _SendT=DT_FLOAT, client_terminated=false, recv_device="/job:localhost/replica:0/task:0/ cpu:0", send_device="/job:localhost/replica:0/task:0/gpu:0", send_device_incarnation=1, tensor_name="edge_31_Global_Step_1", _device="/job:localhost/replica:0/task: 0/gpu:0"]]消息) tensorflow.python.framework.errors_impl.FailedPreconditionError: Attempting to use uninitialized value Global_Step_1 [[Node: Global_Step_1/_96 = _SendT=DT_FLOAT, client_terminated=false, recv_device="/job:localhost/replica:0/task:0/ cpu:0", send_device="/job:localhost/replica:0/task:0/gpu:0", send_device_incarnation=1, tensor_name="edge_31_Global_Step_1", _device="/job:localhost/replica:0/task: 0/gpu:0"]]_device="/job:localhost/replica:0/task:0/gpu:0"]]_device="/job:localhost/replica:0/task:0/gpu:0"]]
我认为 ADAM 优化器变量未初始化。保存这样的模型的正确方法是什么？