Github user avulanov commented on the pull request:
https://github.com/apache/spark/pull/3222#issuecomment-67250526
@witgo Thanks, I converted labels to one-hot vectors, and I was able to
train MLP. I used mnist 60K train and 10K test, topology 780x32x10, 100 steps,
batchSize = 1, LBFGS optimizer. It took 953 seconds on my laptop and produced
0.9404 accuracy and 0.060 error. Could you suggest which size of batch should I
set in MLP to get the correct comparison with our implementation, as we don't
have batchSize parameter?
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
