【发布时间】:2019-04-27 20:21:05
【问题描述】:
当我在容器 tensorflow/tensorflow:latest-gpu 中运行 tensorflow 图像训练作业时,它不起作用。
错误信息:
Cannot assign a device for operation InceptionV3/InceptionV3/Conv2d_1a_3x3/Conv2D: Operation was explicitly assigned to /device:GPU:0 but available devices are [ /job:localhost/replica:0/task:0/device:CPU:0, /job:localhost/replica:0/task:0/device:XLA_CPU:0, /job:localhost/replica:0/task:0/device:XLA_GPU:0 ]. Make sure the device specification refers to a valid device.
[[node InceptionV3/InceptionV3/Conv2d_1a_3x3/Conv2D (defined at /usr/local/lib/python2.7/dist-packages/tensorflow/contrib/layers/python/layers/layers.py:1057) = Conv2D[T=DT_FLOAT, data_format="NHWC", dilations=[1, 1, 1, 1], padding="VALID", strides=[1, 2, 2, 1], use_cudnn_on_gpu=true, _device="/device:GPU:0"](fifo_queue_Dequeue, InceptionV3/Conv2d_1a_3x3/weights/read)]]
GPU 信息:
英伟达-smi
2018 年 11 月 26 日星期一 07:48:59
+-------------------------------------------------- --------------------------------------------+
| NVIDIA-SMI 410.72 驱动程序版本:410.72 CUDA 版本:10.0 |
|-------------------------------+----------------- -----+-----------+
| GPU名称持久化-M|总线 ID Disp.A |挥发性的 Uncorr。纠错 |
| Fan Temp Perf Pwr:Usage/Cap|内存使用 | GPU-Util Compute M. |
|=================================+================= =====+========================|
| 0 GeForce GT 630 关闭 | 00000000:01:00.0 不适用 |不适用 |
| 25% 47C P0 不适用 / 不适用 | 0MiB / 1998MiB |不适用 默认 |
+-------------------------------+------------------ -----+-----------+
+--------------------------------------------- -----------------------------------------------+ |进程:GPU 内存 | | GPU PID 类型 进程名称 用法 | |=================================================== =============================| | 0 不支持 | +-------------------------------------------------- ----------------------------------------+
【问题讨论】:
-
你在加载之前训练过的模型吗?
-
这是我的步骤 mkdir download_data_flower mkdir train_output python download_and_convert_data.py --dataset_name=flowers --dataset_dir=download_data_flower python train_image_classifier.py --batch_size=64 --model_name=inception_v3 --dataset_name=flowers - -dataset_split_name=train --dataset_dir=download_data_flower --train_dir=train_output
标签: tensorflow gpu