Models: Errors when reading the DELF pre-trained model

[zhpcui]

System information
Linux Ubuntu 16.04
TensorFlow 1.4.0 (GPU support) installed using Anaconda
CUDA: 8.0
cuDNN: 6.0
Python: 2.7
GPU: NVIDIA GeForce GTX 1080 Ti

Describe the problem
I successfully installed DELF by following the illustration (https://github.com/tensorflow/models/blob/master/research/delf/INSTALL_INSTRUCTIONS.md).
However when I tried the given example with pre-trained model (delf_v1_20171026), I got some errors as the following:
...
W tensorflow/core/framework/op_kernel.cc:1192] Data loss: Unable to read file (parameters/delf_v1_20171026/model/variables/variables.index). Perhaps the file is corrupt or was produced by a newer version of TensorFlow with format changes (failed to seek to header entry): corrupted compressed block contents
...

[zhpcui]

I have solved the problem by switching TF 1.4.0 to TF 1.3.0.

[andrefaraujo]

thanks for your continued investigation, @zhpcui! Sorry that you were having problems with this. I will update the delf repo to mention this issue.

[andrefaraujo]

I discussed with @vrv and this seems exactly the same problem as #2648. My understanding is that there were some problems in the TF 1.4.0 release, which are being worked on right now and will be fixed very soon.

[zhpcui]

Thank you very much, Andre.

[angerson]

This appears to be resolved (or on the track to resolution), so I'm closing this issue to keep the tracker focused. Feel free to reopen it if needed.