DominikaJedynak opened a new pull request, #21008:
URL: https://github.com/apache/incubator-mxnet/pull/21008

   ## Description ##
   This PR adds the possibility to re-run the test with all batches used for 
calibration of quantized model. 
   If test fails with only one batch used for calibration, it will be re-run 
with all batches to achieve better representation of the whole data range and 
more precise quantization scales. Thanks to that, loss in accuracy caused by 
too small calibration dataset will be avoided.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to