Batch stuck in Waiting for Class Training
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
12-12-20 09:55 AM
I'm trying to test the classification capabilities of Decipher.
I prepared a batch of 53 documents that were sent to Decipher (as a training batch). After the OCR step, the batch enters the status "Waiting for Class Training". However, it has been in this state for more than an hour now, and I see no progress being made.
Is it a matter of having even more patience, or are there certain steps to be taken to actually start the classification training ?
------------------------------
Frederic Taes
RPA Consultant
RoboRana
------------------------------
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
14-12-20 06:42 AM
Have you trained your classification model? Information can be found here, it sounds like you could have missed this step: "4. Update classification model". Make sure you mark the model for training after uploading the batches.
Best regards,
Lauris Kūms
------------------------------
Lauris Kums
------------------------------
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
14-12-20 08:37 AM
it is exactly these steps that I have been following.
It is in step 3 (after pushing the documents from the amended process to Decipher), that the document batch remains in status "Waiting for class training".
I've tried by continuing to step 4 and 5, but the batch remains in the status and the documents do not arrive in the verification step.
------------------------------
Frederic Taes
RPA Consultant
RoboRana
------------------------------
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
14-12-20 09:33 PM
when in status "Waiting for class training" these documents will never arrive in Document Verification step as these are meant only for classification training - so class verification. If your goal is to perform classification training then I recommend following (I experienced similar behavior several times and this is trial and error approach which worked, don't ask me why). Create a new document type from scratch and prepare it for classification training as specified here. Run just a few documents and perform the steps as described in help. The documents should move from waiting for class training to Completed class training.
Regards,
------------------------------
Zdeněk Kabátek
Head of Professional Services
NEOOPS
http://www.neoops.com/
Europe/Prague
------------------------------
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
15-01-21 01:11 PM
Hi Zdenek,
Unfortunately I'm still not able to get it working. It remains in status "Waiting for class training", even after restarting everything from scratch.
In the Classification Client, I'm getting the following errors:
2021-01-15 08:25:53.4603 [2028-12] INFO | Waiting for next batch...
2021-01-15 13:54:55.4299 [2028-12] ERROR | Exception occurred while signaling process heartbeat : System.IO.FileNotFoundException: Unable to find the specified file.
at System.IO.__Error.WinIOError(Int32 errorCode, String maybeFullPath)
at System.IO.MemoryMappedFiles.MemoryMappedFile.OpenCore(String mapName, HandleInheritability inheritability, Int32 desiredAccessRights, Boolean createOrOpen)
at System.IO.MemoryMappedFiles.MemoryMappedFile.OpenExisting(String mapName, MemoryMappedFileRights desiredAccessRights, HandleInheritability inheritability)
at SsiServerCommonUtils.IpcLockTracker.SignalProcessHeartBeat()
at SsiClassificationClient.ClassificationClient.DoWorkImpl(Object sender, DoWorkEventArgs e)
2021-01-15 13:54:55.4771 [2028-12] INFO | No batch available, try to perform class model training.
2021-01-15 13:54:56.3984 [2028-12] INFO | Check for classification model training...
2021-01-15 13:55:00.0245 [2028-12] WARN | Unable to lock class model with ID = 11551854-6432-429d-832a-cc37bf0ecb96 for training. Error message: Request failed for command 'LockClassificationModelForTraining' with error code '-347' and error message: 'The state of the specified classification model is not valid for this operation.'
2021-01-15 13:55:00.1493 [2028-12] INFO | No batch available for classification, perform retry...
2021-01-15 13:55:05.1741 [2028-12] INFO | Waiting for next batch...
2021-01-15 13:55:25.8409 [2028-12] ERROR | Exception occurred while signaling process heartbeat : System.IO.FileNotFoundException: Unable to find the specified file.
Unfortunately, this doesn't help a single bit. I don't know which missing file it's talking about, or why it cannot lock the model. Any suggestions ?
------------------------------
Frederic Taes
RPA Consultant
RoboRana
------------------------------
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
18-01-21 07:39 PM
https://bpdocs.blueprism.com/decipher/user-guide/classification-model-training-guide.htm?Highlight=classification
------------------------------
Ramon Nunez
Customer Solutions Engineer
Blue Prism Software Inc
UTC
------------------------------
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
19-01-21 09:00 PM
a workaround just to enable you to work with Decipher :). In your batch do you plan to use more than 1 document type? If you have mapping - 1 batch to 1 document type then in my opinion you don't need to perform classification and you can begin to use Decipher straight away. You just don't specify classification model in batch definition and it should work. At least I did it few times in my case. Of course, when your batch contains more document types (invoices, orders, etc.) then you need to perform classification and train the classification model. However, if it is not the case and your mapping is 1to1 go straight with data verification skipping the classification model. I can't say it is 100% safe but it will get you moving behind the point where you are stuck currently.
Regards,
------------------------------
Zdeněk Kabátek
Head of Professional Services
NEOOPS
http://www.neoops.com/
Europe/Prague
------------------------------
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
01-02-21 02:12 PM
I had this problem too ... and the solution was so obvious in my case 😞
I just had to swith on the classification option in the Admin panel / configuration / classification (yes, just a single option). That worked perfectly ...
As i remember this option is set to off by default. So quite logical in fact... the configuration said i don't want to manage classification, why training the classifier ? 😉
Hope that help.
NB: So obvious ... however i spent many hours to figure out the point 😄
------------------------------
Benoit Cayla
SC
BLUEPRISM
Europe/Paris
------------------------------
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
02-12-21 08:32 AM
Thank you
------------------------------
Sandro Sarango
Lead Developer
AutomationBot
Europe/Paris
------------------------------
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
03-08-22 09:44 AM
I am trying to train my classification model for a unique type of document. But it gets stuck at " in class training" stage forever. Any idea what is going wrong here??
RPA Developer
------------------------------
Shweta Yadav
------------------------------
