ImportError: DLL load failed: The specified module could not be found.

[bhosalems]

I have installed the right pytorch from pytorch v0.3.1
I have maxwell GPU QuadroM1200 which is of compute capability 5.0. I am using windows10 64bit.
Despite of installing all required dependencies, I am getting the below error while importing torch in python interpreter.

(fastai) C:\Users\Mahesh.Bhosale>python
Python 3.6.5 |Anaconda, Inc.| (default, Mar 29 2018, 13:32:41) [MSC v.1900 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.

>>> import torch
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "C:\Users\Mahesh.Bhosale\AppData\Local\Continuum\anaconda3\envs\fastai\lib\site-packages\torch\__init__.py", line 76, in <module>
    from torch._C import *
ImportError: DLL load failed: The specified module could not be found.
>>> exit()

I am not in a position or rather hesitating to build from the source at current moment can you please guide ?

[bhosalems]

Also when I install pytorch 0.3.1 with below command,
conda install -c peterjc123 pytorch cuda90

I am able to import torch. I don't know which exact DLL then it misses when I install pytorch 0.3.1 CI package.

Your help is much appreciated.

[peterjc123]

It's very simple. The CI packages need exactly cuda 8 and cudnn 6.1. Any other versions are rejected.

[bhosalems]

But I see from fastai forum that there are ones with cuda 9 and cudnn 7.1 working very fine.
http://forums.fast.ai/t/cuda-error-45-no-kernel-image-pytorch-0-3-1-post2/14063/57?u=maheshbhosale

Though I have Cuda 8 and cudnn 6. And all the are mentioned in environment variables.

[bhosalems]

Also, I didn't got cudnn 6.1 anywhere on NVIDIA websites. Do you have a link?

[peterjc123]

Just wait a few days and I'll rebuild for 0.4.0 with cuda 9 and cudnn 7.

[bhosalems]

Almighty. Thanks a lot.

[peterjc123]

The CI packages are regenerated now.