这两个问题的同一问题(1 / 2)。当我在Lucid之后检索要使用tutorial显示的TF模型时,会发生这种情况。我的完整堆栈跟踪是:
/usr/local/lib/python3.6/dist-packages/h5py/的初始化的.py:36: FutureWarning:转换issubdtype的第二个参数 不推荐
float
到np.floating
。将来,它将被对待 为np.float64 == np.dtype(float).type
。来自._conv导入 register_converters为_register_converters2018年5月8日 01:22:34.477907:我 tensorflow / stream_executor / cuda / cuda_gpu_executor.cc:898]成功 从SysFS读取的NUMA节点具有负值(-1),但必须存在 至少有一个NUMA节点,因此返回NUMA节点零
2018年5月8日 01:22:34.478221:我 tensorflow / core / common_runtime / gpu / gpu_device.cc:1344]找到设备0 具有属性:名称:特斯拉K80专业:3个未成年人:7 memoryClockRate(GHz):0.8235 pciBusID:0000:00:04.0 totalMemory: 11.17GiB freeMemory:362.12MiB
2018-05-08 01:22:34.478266:I tensorflow / core / common_runtime / gpu / gpu_device.cc:1423]添加可见 gpu设备:0 2018-05-08 01:22:34.871564:我 tensorflow / core / common_runtime / gpu / gpu_device.cc:911]设备 使用强度1边缘矩阵互连StreamExecutor:2018-05-08 01:22:34.871635:我 tensorflow / core / common_runtime / gpu / gpu_device.cc:917] 0 2018-05-08 01:22:34.871680:我 tensorflow / core / common_runtime / gpu / gpu_device.cc:930] 0:N 2018-05-08 01:22:34.871877:我 tensorflow / core / common_runtime / gpu / gpu_device.cc:1041]已创建 TensorFlow设备(/ job:localhost / replica:0 / task:0 / device:GPU:0 with 101 MB内存) - >物理GPU(设备:0,名称:特斯拉K80,pci总线 id:0000:00:04.0,计算能力:3.7)
追踪(最近的电话 最后):
文件 " /usr/local/lib/python3.6/dist-packages/tensorflow/python/tools/freeze_graph.py" ;, 380号线 app.run(main = main,argv = [sys.argv [0]] + unparsed)
File" /usr/local/lib/python3.6/dist-packages/tensorflow/python/platform/app.py", 第126行,在运行中 _sys.exit(main(argv))
File" /usr/local/lib/python3.6/dist-packages/tensorflow/python/tools/freeze_graph.py", 第274行,主要部分 FLAGS.saved_model_tags,checkpoint_version)
File" /usr/local/lib/python3.6/dist-packages/tensorflow/python/tools/freeze_graph.py", 第256行,在freeze_graph中 checkpoint_version = checkpoint_version)
File" /usr/local/lib/python3.6/dist-packages/tensorflow/python/tools/freeze_graph.py", 第130行,在freeze_graph_with_def_protos中 var_list = var_list,write_version = checkpoint_version)
File" /usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py", 第1311行, init self.build()File" /usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py", 1320行,在构建中 self._build(self._filename,build_save = True,build_restore = True)
文件 " /usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py" ;, 第1357行,在_build中 build_save = build_save,build_restore = build_restore)
File" /usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py", 第787行,在_build_internal中 saveables = self._ValidateAndSliceInputs(names_to_saveables)
文件 " /usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py" ;, 第688行,在_ValidateAndSliceInputs中 变量) TypeError:names_to_saveables必须是字符串映射到Tensors / Variables的字符串名称。不是变量: Tensor(" conv2d_1 / bias:0",shape =(16,),dtype = float32)
我的网络图定义如下:
input_1
conv2d_1 / kernel
conv2d_1 / kernel / read
conv2d_1 /偏置
conv2d_1 / bias / read
conv2d_1 / convolution
conv2d_1 / BiasAdd
conv2d_1 / RELU
max_pooling2d_1 / MaxPool
conv2d_2 / kernel
conv2d_2 /内核/读
conv2d_2 /偏置
conv2d_2 / bias / read
conv2d_2 / convolution
conv2d_2 / BiasAdd
conv2d_2 / RELU
max_pooling2d_2 / MaxPool
conv2d_3 /内核
conv2d_3 / kernel / read
conv2d_3 / bias
conv2d_3 /偏压/读
conv2d_3 / convolution
conv2d_3 / BiasAdd
conv2d_3 / RELU
max_pooling2d_3 / MaxPool
conv2d_4 / kernel
conv2d_4 /内核/读
conv2d_4 / bias
conv2d_4 / bias / read
conv2d_4 / convolution
conv2d_4 / BiasAdd
conv2d_4 / Relu
up_sampling2d_1 /形状
up_sampling2d_1 / strided_slice /堆
up_sampling2d_1 / strided_slice / stack_1
up_sampling2d_1 / strided_slice / stack_2
up_sampling2d_1 / strided_slice
up_sampling2d_1 / Const
up_sampling2d_1 / MUL
up_sampling2d_1 / ResizeNearestNeighbor
conv2d_5 /内核
conv2d_5 / kernel / read
conv2d_5 / bias
conv2d_5 /偏压/读
conv2d_5 / convolution
conv2d_5 / BiasAdd
conv2d_5 / RELU up_sampling2d_2 / Shape up_sampling2d_2 / strided_slice / stack
up_sampling2d_2 / strided_slice / stack_1
up_sampling2d_2 / strided_slice / stack_2 up_sampling2d_2 / strided_slice
up_sampling2d_2 / Const up_sampling2d_2 / mul
up_sampling2d_2 / ResizeNearestNeighbor conv2d_6 / kernel
conv2d_6 / kernel / read conv2d_6 / bias conv2d_6 / bias / read
conv2d_6 / convolution conv2d_6 / BiasAdd conv2d_6 / Relu
up_sampling2d_3 / Shape up_sampling2d_3 / strided_slice / stack
up_sampling2d_3 / strided_slice / stack_1
up_sampling2d_3 / strided_slice / stack_2 up_sampling2d_3 / strided_slice
up_sampling2d_3 / Const up_sampling2d_3 / mul
up_sampling2d_3 / ResizeNearestNeighbor conv2d_7 / kernel
conv2d_7 / kernel / read conv2d_7 / bias conv2d_7 / bias / read
conv2d_7 / convolution conv2d_7 / BiasAdd conv2d_7 / Relu output_node0