Hi,
I should probably have sent an update...
The hardware for training is on order and should be delivered over the
next couple of weeks. We've also been busy working on additional
interactive tools for training and quality control and improving the
Python bindings. The way it looks now, you'll be able to run the
complete OCR pipeline from Python, which should make it much easier for
people to customize things.
We're using the interactive tools to identify sources of recognition
errors, which is also taking its time. Basically, with the training and
debugging, we want to get competitive performance on UNLV data without
adaptation.
Tom
Bill Janssen wrote:
How's the access to that very large server to do the model training
on, coming?
Bill
On Nov 20 2009, 9:32 am, Bill Janssen <[email protected]> wrote:
Looking forward to it...
Bill
On Nov 20, 12:12 am, Thomas Breuel <[email protected]> wrote:
Yes, the tools support Unicode, ligatures, linked scripts, and Indic languages.
Tom
On Thu, Nov 19, 2009 at 14:01, 74yrs old <[email protected]> wrote:
Reg:"We've developed a new set of labeling and transcription tools and
used it to label millions of training samples "
Whether said developed tools will support UTF-8 as well as Indic lang?
-sriranga(77yrsold)
On Thu, Nov 19, 2009 at 4:23 PM, Thomas Breuel <[email protected]> wrote:
Well, among other things, we haven't been updating the external
repository much. It's the quiet before the beta release, I suppose.
The status right now is:
-- Most of what we wanted to accomplish for beta has been
accomplished. There is still some code cleanup to be done, however.
-- We've developed a new set of labeling and transcription tools and
used it to label millions of training samples (those will be released
hopefully in a few months).
-- We're waiting for a large server to do very large model training on.
Cheers,
Tom
On Thu, Nov 19, 2009 at 03:45, Bill Janssen <[email protected]>
wrote:
Odd to see no new messages on this group for the past few weeks...
Bill
--
You received this message because you are subscribed to the Google
Groups "ocropus" group.
To post to this group, send email to [email protected].
For more options, visit this group at
http://groups.google.com/group/ocropus?hl=.
--
You received this message because you are subscribed to the Google Groups
"ocropus" group.
To post to this group, send email to [email protected].
For more options, visit this group at
http://groups.google.com/group/ocropus?hl=.
--
You received this message because you are subscribed to the Google Groups
"ocropus" group.
To post to this group, send email to [email protected].
For more options, visit this group at
http://groups.google.com/group/ocropus?hl=.
--
You received this message because you are subscribed to the Google Groups
"ocropus" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/ocropus?hl=en.
