About v2 training — egmentation fault(core dumped)

[jn-z]

Hi,
Thank you for your last help. Maybe this problem needs your help again this time.
When I run python training.py, an errror message showed up:
Traceback (most recent call last):
...
Tensor("tower_0/netFlow1/dense5/LeakyRelu:0", shape=(16, 4608), dtype=float32, device=/device:GPU:0)
Tensor("tower_1/netFlow1/dense5/LeakyRelu:0", shape=(16, 4608), dtype=float32, device=/device:GPU:1)
nothing to restore. no checkpoint found.
train evolution "0_flow1" ...
start iteration 0
2017-12-19 22:36:48.508170: I tensorflow/stream_executor/dso_loader.cc:139] successfully opened CUDA library libcupti.so.8.0 locally
Segmentation fault(core dumped)

Any ideas?
Thanks,
JUNZHANG

[ialhashim]

This could happen when your dataset is not in the right folder. E.g. ../../datasets/traindata/sun3d_train_0.1m_to_0.2m.h5