# RuntimeError: cuDNN version mismatch: PyTorch was compiled against 7005 but linked against 7102

**URL:** <https://deeptalk.lambda.ai/t/runtimeerror-cudnn-version-mismatch-pytorch-was-compiled-against-7005-but-linked-against-7102/305>\
**Category:** Technical Help\
**Created:** [May 15, 2018, 10:18pm UTC](https://deeptalk.lambda.ai/t/runtimeerror-cudnn-version-mismatch-pytorch-was-compiled-against-7005-but-linked-against-7102/305 "2018-05-15T22:18:40Z")\
**Posts on this page:** 1\
**Showing post:** 1

<div class="post-metadata">

**Author:** ![sabalaba](https://sea1.discourse-cdn.com/flex019/user_avatar/deeptalk.lambda.ai/sabalaba/32/4_2.png) [@sabalaba](https://deeptalk.lambda.ai/u/sabalaba)\
**Post date:** [May 15, 2018, 10:18pm UTC](https://deeptalk.lambda.ai/t/runtimeerror-cudnn-version-mismatch-pytorch-was-compiled-against-7005-but-linked-against-7102/305/1 "2018-05-15T22:18:40Z")

</div>

Lambda Stack Cuda pytorch error when getting the cudnn version with `torch.backends.cudnn.versino()`. OS: Ubuntu 16.04

```shell
user@hostname$ python3
Python 3.5.2 (default, Nov 23 2017, 16:37:01)
[GCC 5.4.0 20160609] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import torch
>>> torch.version
'0.3.1'
>>> torch.backends.cudnn.version()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python3/dist-packages/torch/backends/cudnn/init.py", line 34, in version
    if _libcudnn() is None:
  File "/usr/lib/python3/dist-packages/torch/backends/cudnn/init.py", line 27, in _libcudnn
    'but linked against {}'.format(compile_version, __cudnn_version))
RuntimeError: cuDNN version mismatch: PyTorch was compiled against 7005 but linked against 7102

```

---

_[View the full topic](https://deeptalk.lambda.ai/t/runtimeerror-cudnn-version-mismatch-pytorch-was-compiled-against-7005-but-linked-against-7102/305)._