Moses-support Digest, Vol 175, Issue 4

Send Moses-support mailing list submissions to
moses-support@mit.edu

To subscribe or unsubscribe via the World Wide Web, visit
http://mailman.mit.edu/mailman/listinfo/moses-support
or, via email, send a message with subject or body 'help' to
moses-support-request@mit.edu

You can reach the person managing the list at
moses-support-owner@mit.edu

When replying, please edit your Subject line so it is more specific
than "Re: Contents of Moses-support digest..."


Today's Topics:

1. Tuning an already tuned model (Marwa Gaser)


----------------------------------------------------------------------

Message: 1
Date: Fri, 7 May 2021 19:18:54 +0200
From: Marwa Gaser <marwagaser@gmail.com>
Subject: [Moses-support] Tuning an already tuned model
To: moses-support@mit.edu
Message-ID:
<CADNwFM=jE3WH6yaZAGVW9BR-_UUWpd0+Jw-JwAqXS8XWFmpVig@mail.gmail.com>
Content-Type: text/plain; charset="utf-8"

I used MERT to tune a model using a tuning corpus X this generated a tuned
model X1. Then I took X1 and it tuned using the same exact tuning corpus X.
This resulted in a model let's call it X2. Even though I expected X1 and X2
to have the same weights since the optimal parameters were already found in
X1. X2 had different weights than X1 and generated better results. Why is
that?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/mailman/private/moses-support/attachments/20210507/3ed05077/attachment-0001.html

------------------------------

_______________________________________________
Moses-support mailing list
Moses-support@mit.edu
http://mailman.mit.edu/mailman/listinfo/moses-support


End of Moses-support Digest, Vol 175, Issue 4
*********************************************

Tweet