Tensorflow无法构建image_retraining:label_image

Question

我已经用AVX和SSE编译了TensorFlow (从git存储库克隆)。Hello脚本可以工作，并且不会打印TensorFlow支持二进制指令的警告。

稍后，我创建了包含图像的目录，并运行了重新训练器：

$ bazel-bin/tensorflow/examples/image_retraining/retrain --image_dir ~/mydata

但是，当我试图构建label_image时，我会得到以下错误：

bazel build --config=opt tensorflow/examples/image_retraining:label_image
ERROR: Skipping 'tensorflow/examples/image_retraining:label_image': no such target '//tensorflow/examples/image_retraining:label_image': target 'label_image' not declared in package 'tensorflow/examples/image_retraining' defined by /home/user/myretraining/tensorflow/tensorflow/examples/image_retraining/BUILD
WARNING: Target pattern parsing failed.
ERROR: no such target '//tensorflow/examples/image_retraining:label_image': target 'label_image' not declared in package 'tensorflow/examples/image_retraining' defined by /home/user/myretraining/tensorflow/tensorflow/examples/image_retraining/BUILD
INFO: Elapsed time: 0.118s
FAILED: Build did NOT complete successfully (0 packages loaded)

(/home/user/myretraining is the source dir root)

我做错了什么？如何编译label_image？

Answer 1

您的label_image.py不在此目录中，tensorflow/examples/image，在我的示例中，是在tensorflow/examples/label_label_image.py文件夹中。所以试试

bazel build tensorflow/examples/label_image:label_image

希望能帮上忙。