You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@mahout.apache.org by Renaud Grisoni <re...@gmail.com> on 2014/09/02 00:56:10 UTC

Wrong model when loading an AdaptiveLogisicRegression from a binary file

Hi,

I encounter a problem when I apply an AdaptiveLogisicRegression model that
I have loaded from a binary file.
Indeed, the AdaptiveLogisicRegression object I load is not the same object
than the one I saved using the ModelSerializer class : for instance, the
Beta matrix of the OnlineLogisticRegression objects which are contained in
the AdaptiveLogisticRegression object are very different than those I have
saved.

I tried to use the ModelSerialiser class, then the write() and read()
methods from the AdaptiveLogisticRegression object, but I have always had
the same bad results.

Is there a bug during the model (de)serialization I should have heard about
?

Thanks,

Reno

Re: Wrong model when loading an AdaptiveLogisicRegression from a binary file

Posted by Ted Dunning <te...@gmail.com>.
On Mon, Sep 1, 2014 at 3:56 PM, Renaud Grisoni <re...@gmail.com>
wrote:

> Is there a bug during the model (de)serialization I should have heard about
> ?
>

Not a known one.  But this doesn't sound good at all.