You are viewing a plain text version of this content. The canonical link for it is here.

Posted to commits@mxnet.apache.org by GitBox <gi...@apache.org> on 2017/12/11 13:42:22 UTC

[GitHub] my777777 opened a new issue #9024: issue with memory cost

my777777 opened a new issue #9024: issue with memory cost
URL: https://github.com/apache/incubator-mxnet/issues/9024
 
 
   ## Description
   (Brief description of the problem in no more than 2 sentences.)
   when training imagenet?the memory cost of resnet has been greatly increased.
   batch size =32, use resnet-34 model, the memory displayed by nvidia-smi is greater than 6000MB?but at the 0.11.0 branch, its memory cost only 2190MB.
   ## Environment info (Required)
   cuda7.5 + cudnn v5.1
   GPU: K40m, total memory: 11519MB
   Package used: Python
   
   Compiler: gcc
   
   MXNet commit hash:
   0.11.0 branch: a5edbf94094581ee27157eae4f2113115a3994e7
   newest(clone from git at 2017/12/11 21:00): 20666c3186b3d7166cae625b48c81713c6811d67
   
   Build config:
   USE_CUDA = 1
   USE_CUDNN = 1
   
   ## Minimum reproducible example
   use the example at: examples/image-classification/train_imagenet.py
   
   ## Steps to reproduce
   just run example provided by mxnet
   

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
users@infra.apache.org


With regards,
Apache Git Services