singa-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sai Narsi Reddy Donthi Reddy (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (SINGA-298) Check failed: input.device()->lang() == kOpencl
Date Fri, 24 Feb 2017 18:36:44 GMT

    [ https://issues.apache.org/jira/browse/SINGA-298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15883266#comment-15883266
] 

Sai Narsi Reddy Donthi Reddy edited comment on SINGA-298 at 2/24/17 6:36 PM:
-----------------------------------------------------------------------------

Hey [~wangwei.cs], 

I installed latest source code and again tested the code. 
This time, I am receiving the following error when try to train the network.

```
WARNING: Logging before InitGoogleLogging() is written to STDERR
F0224 12:18:14.541901  9631 batchnorm.cc:99] Trainning SpatialBatchNormalization has not been
implemented yet...
*** Check failure stack trace: ***
```

So, I decided to just use a single convolution layer example on layer page: https://singa.incubator.apache.org/en/docs/layer.html

and when I run the code, I received the following error at {y = conv.foward(True, x) }

```
WARNING:root:kernel 7ba57179-81f7-4155-8639-a7ee11c55296 restarted
WARNING: Logging before InitGoogleLogging() is written to STDERR
F0224 12:29:05.774775 11319 common.h:70] Check failed: initialized_ Must initialize data before
reading it
*** Check failure stack trace: ***
```

Also, the tensor initialization in the layer page example is wrong for conv2d. 

```
# Input shape for conv2d is NCHW
x = tensor.Tensor((3, 32, 32), dev)
# so the tensor should be as follows
x = tensor.Tensor((1,3, 32, 32), dev)
```

Thanks,
Narsi


was (Author: itsnarsi):
Hey [~wangwei.cs], 

I installed latest source code and again tested the code. 
This time, I am receiving the following error when try to train the network.

```
WARNING: Logging before InitGoogleLogging() is written to STDERR
F0224 12:18:14.541901  9631 batchnorm.cc:99] Trainning SpatialBatchNormalization has not been
implemented yet...
*** Check failure stack trace: ***
```

So, I decided to just use a single convolution layer example on layer page: https://singa.incubator.apache.org/en/docs/layer.html

and when I run the code, I received the following error at {y = conv.foward(True, x) }

```
WARNING:root:kernel 7ba57179-81f7-4155-8639-a7ee11c55296 restarted
WARNING: Logging before InitGoogleLogging() is written to STDERR
F0224 12:29:05.774775 11319 common.h:70] Check failed: initialized_ Must initialize data before
reading it
*** Check failure stack trace: ***
```

Also, the tensor initialization in the layer page example is wrong for conv2d. 

```
# Input shape for conv2d is NCHW
x = tensor.Tensor((3, 32, 32), dev)
# so the tensor should be as follows
x = tensor.Tensor((1,3, 32, 32), dev)
```

Thanks,
Narsi

>  Check failed: input.device()->lang() == kOpencl
> ------------------------------------------------
>
>                 Key: SINGA-298
>                 URL: https://issues.apache.org/jira/browse/SINGA-298
>             Project: Singa
>          Issue Type: Bug
>          Components: PySINGA
>         Environment: Ubuntu 16.04, AMD rx480, intel i7k 6-gen
>            Reporter: Sai Narsi Reddy Donthi Reddy
>              Labels: newbie
>
> Hello I came across the framework when I was searching for a deep learning framework
with opencl support.
> I cloned the GITHUB master branch version and downloaded as shown on the install page.
> After sucessful install, I tried to run cifar-10 example with opencl instead of CUDA,
I am receiving following errors:
> WARNING: Logging before InitGoogleLogging() is written to STDERR
> F0127 17:07:02.757550 27607 opencl_convolution.cc:30] Check failed: input.device()->lang()
== kOpencl (0 vs. 2) 
> *** Check failure stack trace: ***
> Aborted (core dumped)
> I am new to this and I am learning as I go. 
> Sorry if I posted here mistakenly.
> Thanks,
> Narsi



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Mime
View raw message