Fast-rcnn training error: Check failed: data_ *** Check failure stack trace: *** Aborted (core dumped) when training a VGG16 network

[Truthuleave]

When I trained a vgg16 network using ./tools/train_net.py --gpu 0 --solver models/VGG16/solver.prototxt --weights data/imagenet_models/VGG16.v2.caffemodel, I got an error almost at the end.
Error is as follows:

I0920 10:43:24.275528 14792 solver.cpp:60] Solver scaffolding done.
Loading pretrained model weights from data/imagenet_models/VGG16.v2.caffemodel
[libprotobuf WARNING google/protobuf/io/coded_stream.cc:537] Reading dangerously large protocol message. If the message turns out to be larger than 2147483647 bytes, parsing will be halted for security reasons. To increase the limit (or to disable these warnings), see CodedInputStream::SetTotalBytesLimit() in google/protobuf/io/coded_stream.h.
[libprotobuf WARNING google/protobuf/io/coded_stream.cc:78] The total number of bytes read was 553432430
I0920 10:43:24.639192 14792 net.cpp:816] Ignoring source layer pool5
I0920 10:43:24.724637 14792 net.cpp:816] Ignoring source layer fc8
I0920 10:43:24.724663 14792 net.cpp:816] Ignoring source layer prob
Solving...
F0920 10:43:25.510537 14792 blob.cpp:95] Check failed: data_
*** Check failure stack trace: ***
Aborted (core dumped)

The snapshot:

Anyone has encountered this before,please?

[Truthuleave]

GTX960M 4G memory + cuda 8.0+cudnn 7.1.4 for cuda 8