diff options
author | Vijay Vasudevan <vrv@google.com> | 2017-04-26 19:02:10 -0800 |
---|---|---|
committer | TensorFlower Gardener <gardener@tensorflow.org> | 2017-04-26 20:31:03 -0700 |
commit | 95ca363c6caf5d152eef594c7fc3703a5d490070 (patch) | |
tree | e275d9b938d8c162d2a63e2677c80d590260c567 /LICENSE | |
parent | be43153b21fba4280f0f2a016242614a21d115f3 (diff) |
Make error message when explicitly adding an invalid device more clear.
Before, you would get something like:
`InvalidArgumentError: Cannot assign a device to node 'save/RestoreV2_45': Could not satisfy explicit device specification '/job:ps/task:1/device:CPU:0' because no devices matching that specification are registered in this process; available devices: /job:localhost/replica:0/task:0/cpu:0 [[Node: save/RestoreV2_45 = RestoreV2[dtypes=[DT_FLOAT], _device="/job:ps/task:1/device:CPU:0"](save/Const, save/RestoreV2_45/tensor_names, save/RestoreV2_45/shape_and_slices)]] Caused by op u'save/RestoreV2_45', defined`
And now this reads:
`InvalidArgumentError: Cannot assign a device for operation 'save/RestoreV2_45':
Operation was explicitly assigned to '/job:ps/task:1/device:CPU:0'
but available devices are [ /job:localhost/replica:0/task:0/cpu:0 ]. Make
sure the device specification refers to a valid device.
This drops the additional debug_info because this condition is tightly
scoped to lack of devices, and so node information is unnecessary.
Change: 154379360
Diffstat (limited to 'LICENSE')
0 files changed, 0 insertions, 0 deletions