error - CUDA driver version is insufficient for CUDA runtime version

Hello,

I have installed CUDA - 8.0 , CUDNN-5.0 using JetPack 3.1 on my Ubuntu 16.04 x86_64 PC.
I am trying to run a tensorflow object detection algorithm on this environment.
When I run the below command on the terminal

$lspci | grep VGA we get this:
01:00.0 VGA compatible controller: NVIDIA Corporation GK107 [GeForce GT 640 OEM] (rev a1)

When I run the algorithm given in this link :
https://github.com/smallcorgi/Faster-RCNN_TF

I am facing this error:
python demo.py --model /home/techm/Faster-RCNN_TF/VGGnet_fast_rcnn_iter_70000.ckpt

W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn’t compiled to use SSE3 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn’t compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn’t compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn’t compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
Tensor(“Placeholder:0”, shape=(?, ?, ?, 3), dtype=float32)
Tensor(“conv5_3/conv5_3:0”, shape=(?, ?, ?, 512), dtype=float32)
Tensor(“rpn_conv/3x3/rpn_conv/3x3:0”, shape=(?, ?, ?, 512), dtype=float32)
Tensor(“rpn_cls_score/rpn_cls_score:0”, shape=(?, ?, ?, 18), dtype=float32)
Tensor(“rpn_cls_prob:0”, shape=(?, ?, ?, ?), dtype=float32)
Tensor(“rpn_cls_prob_reshape:0”, shape=(?, ?, ?, 18), dtype=float32)
Tensor(“rpn_bbox_pred/rpn_bbox_pred:0”, shape=(?, ?, ?, 36), dtype=float32)
Tensor(“Placeholder_1:0”, shape=(?, 3), dtype=float32)
Tensor(“conv5_3/conv5_3:0”, shape=(?, ?, ?, 512), dtype=float32)
Tensor(“rois:0”, shape=(?, 5), dtype=float32)
[<tf.Tensor ‘conv5_3/conv5_3:0’ shape=(?, ?, ?, 512) dtype=float32>, <tf.Tensor ‘rois:0’ shape=(?, 5) dtype=float32>]
Tensor(“fc7/fc7:0”, shape=(?, 4096), dtype=float32)

Loaded network /home/techm/Faster-RCNN_TF/VGGnet_fast_rcnn_iter_70000.ckpt
CUDA driver version is insufficient for CUDA runtime version
CUDA driver version is insufficient for CUDA runtime version
CUDA driver version is insufficient for CUDA runtime version
CUDA driver version is insufficient for CUDA runtime version
CUDA driver version is insufficient for CUDA runtime version
CUDA driver version is insufficient for CUDA runtime version
CUDA driver version is insufficient for CUDA runtime version
CUDA driver version is insufficient for CUDA runtime version
Segmentation fault (core dumped)

Should I install a CUDA version compatible with the driver GK107 [GeForce GT 640 OEM]?
How can I solve this?

Kindly help.
Thanks,
Pratosha

You probably haven’t install a GPU driver correctly. What do you get if you run:

nvidia-smi

?

Hello txbob,

I have fixed the issue by uninstalling CUDA and reinstalling it with this version of CUDA- cuda-repo-ubuntu1604-8-0-local_8.0.44-1_amd64.deb

Thanks for your response
when I run
$ nvidia-smi
It is giving me details as follows:
Thu Jun 14 14:49:55 2018
±----------------------------------------------------------------------------+
| NVIDIA-SMI 384.130 Driver Version: 384.130 |
|-------------------------------±---------------------±---------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GT 640 Off | 00000000:01:00.0 N/A | N/A |
| 16% 27C P8 N/A / N/A | 246MiB / 977MiB | N/A Default |
±------------------------------±---------------------±---------------------+

±----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 Not Supported |
±----------------------------------------------------------------------------+

The algorithm requires 3GB of GPU memory whereas GeForce provides 1GB.
That is one of the reasons the algorithm is crashing.
Any other hardware GPU that you would recommend which I can connect my PC with that has more GPU memory?
Can I run this algorithm on Nvidia GPU cloud?
If yes what are the steps to be followed?

Kindly help.
Thanks