[GitHub] feevos opened a new issue #11119: Feature request: speed up asnumpy()

[GitBox <gi...@apache.org>]

feevos opened a new issue #11119: Feature request: speed up asnumpy()
URL: https://github.com/apache/incubator-mxnet/issues/11119
 
 
   Dear all, 
   
   I am trying to visualize some predictions from an NN run so I need to use ```asnumpy()``` to get ```imshow``` visualizations. For an ```nd.array``` of size 3000x3000  ```(ctx=mx.cpu())``` this takes forever. Like **unusable**-forever :(. In contrast the function ```nd.argmax(some_array_6X3KX3K)``` for some nd.array of shape (6,3000,3000) runs instantly. Surely the ```asnumpy()``` operation can be done much much faster. Could you please consider speeding up this function? It would be a great addition to mxnet usability since the majority of packages in python are using numpy (e.g. I need to save my output as a geotiff file, need to use rasterio which is numpy compatible, cannot use nd.array format). 
   
   The operation that is still running on my laptop is  the simple
   ```Python
   preds = tpreds_use.asnumpy()
   ```
   where I expect preds to be a 3K by 3K numpy array, and tpreds_use is a 3X by 3K nd.array 
   
   Thank you very much for the great work you put on 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