![]() |
KnowBrainer Speech Recognition | ![]() |
Topic Title: Retraining Hell! Topic Summary: I have to retrain everything after importing Dragon profile, vocabulary and commands Created On: 01/16/2021 09:49 PM Status: Post and Reply |
|
![]() |
![]() |
- Todachen | - 01/16/2021 09:49 PM |
![]() |
![]() |
- Scribe | - 01/16/2021 10:38 PM |
![]() |
![]() |
- dilligence | - 01/16/2021 10:50 PM |
![]() |
![]() |
- Edgar | - 01/17/2021 11:25 AM |
![]() |
![]() |
- Alan Cantor | - 01/18/2021 11:34 AM |
![]() |
![]() |
- Ag | - 01/21/2021 07:54 PM |
![]() |
![]() |
- Lunis Orcutt | - 01/21/2021 08:42 PM |
![]() |
![]() |
- mppcarey | - 01/22/2021 07:59 PM |
![]() |
![]() |
- Lunis Orcutt | - 01/22/2021 09:57 PM |
![]() |
![]() |
- mppcarey | - 01/22/2021 11:07 PM |
![]() |
![]() |
- kkkwj | - 01/22/2021 10:16 PM |
![]() |
![]() |
- Stephan Kuepper | - 01/25/2021 04:15 AM |
![]() |
|
I'm learning the hard way that when you import your Dragon profile, custom word list and commands you have to start from scratch.
I use Acronis True Image to backup my system. I recently had to restore an old backup or I had installed Dragon. Before doing so, I backed up my custom vocabulary, my commands and my user profile. After recovering the system from the Acronis backup, I installed DPI 15.61, imported my user profile, my commands, as well as the custom word list backup files. Now I'm stuck in retraining hell. I have to retrain all of my custom words, and my commands. Dragon does not recognize them. Also, Dragon appears to be making tons of errors when I dictate. It's as if I'm starting with a completely new user profile.
Going back to old backups of my system is something I do regularly. Are there any other users out there who also do this, and does anybody have a solution for escaping from retraining hell?
My assumption is that there is no workaround for this. However I'm hoping that somebody might have some advice. Any advice you can give would be much appreciated. :-) ------------------------- My voice computer: W10/DPI 15.6/KB2017/VC 2020 on a Intel Core i7-8559U CPU @ 2.70GHz, 16gb DDR4, Sennheiser ME 3D 431 II / Roccat Juke – ROC-14-11-AM |
|
|
|
![]() |
|
My suggestion would be to create a new user profile and import your custom words and commands into it. It sounds like the old profile is corrupted.
I haven't had to do this in a while, but my memory is that you have to retrain the custom words after importing them into the new profile. (I can't speak to commands, as I do not have custom commands.) ------------------------- The woman who dueled with Aaron Burr and won: www.MmeJumel.com
Crohn's News Blog: www.crohns-news.net |
|
|
|
![]() |
|
In addition to Scribe's sage advice, let me start by congratulating you on using a powerful system imaging tool like Acronis (I've used it for years but eventually switched to the more reliable Macrium Reflect). It's the most crucial step any physically challenged Dragon user can take to avoid the disaster of having to reinstall Windows all over again (and possibly losing a Dragon program activation).
The second best step is to not import a user profile belonging to a previous Dragon version. This usually causes Dragon hangups and/or quick deterioration of the user profile. If you install/upgrade Dragon you should always create a fresh user profile and then import the MyCmds.dat and your custom word list (as Scribe is suggesting).
After that it is time to create a new system image!
Hope that helps. ------------------------- Auto Box - SP Editor - HyperNotes - Hotkeys for Dragon - Quick Correct - Press Utility - Easy Guide to Dragon Scripting |
|
|
|
![]() |
|
I take a system image (just my C:\ drive) every Saturday night/Sunday morning - starting it when I go to bed. I use the Windows® built in disk imaging software (used to be easy to find in Windows 7 but in Windows 10 you must use the Run box: sdclt.exe /BLBBACKUPWIZARD). After I built my computer, immediately after installing Windows 10 (and ABSOLUTELY nothing else) I took my first image; then I installed Dragon and KnowBrainer, got them both working just like I wanted them, and took my second image; I then installed a handful of critical applications/utilities and took a third image; thereafter, I started my once-a-week imaging schedule. I use a dedicated external (USB) 2 TB Seagate hard drive ($50), it has a LOT of images of my system drive (~200 MB used) and is about half full. I have restored from it many times (I frequently restore after installing, testing then removing an application - just to be sure I have removed all artifacts). ------------------------- -Edgar |
|
|
|
![]() |
|
I have to retrain all of my custom words, and my commands. Dragon does not recognize them. Also, Dragon appears to be making tons of errors when I dictate. It's as if I'm starting with a completely new user profile. New profiles don't need training, and Dragon shouldn't be making errors that it wasn't making before. This sounds like a problem with your audio system. Test it and ensure the signal is clear and strong, and that the microphone isn't picking up environmental noises and/or adding distortions and interference. |
|
|
|
![]() |
|
Following the advice of @AlanCantor and many others on this forum I have almost completely given up training Dragon. ------------------------- DPG15.6 (also DPI 15.3) + KB, Sennheiser MB Pro 1 UC ML, BTD 800 dongle, Windows 10 Pro, MS Surface Book 3, Intel Core i7-1065G7 CPU @ 1.3/1.5GHz (4 cores, 8 logical, GPU=NVIDIA Quadro RTX 3000 with Max-Q Design. |
|
|
|
![]() |
|
You can train Dragon 15 EXACTLY the way you trained previous versions. Say the KnowBrainer Train Dragon command, click the Next button and you should see the following:
------------------------- Forum Mission Statement |
|
|
|
![]() |
|
Will that work with DMPE 4.3 as well?
------------------------- Dragon Medical Practice Edition 4.3; Dragon Capture, Speechproductivity & Knowbrainer 2017 |
|
|
|
![]() |
|
Yes, it will work in DMPE 4.3
------------------------- Forum Mission Statement |
|
|
|
![]() |
|
That is great to hear! I'll give it a try, I still can get DMPE 4.3 to work from home. ------------------------- Dragon Medical Practice Edition 4.3; Dragon Capture, Speechproductivity & Knowbrainer 2017 |
|
|
|
![]() |
|
I had to smile when I read Edgar's procedure for building up a new machine. Install Windows and take an image immediately so you do not have to spend hours if something goes wrong. Then your critical apps - put in another piton! And so on. Regular system images are a good practice. I was surprised to hear that Windows has its own imaging software. I have never heard of that before. I learned that system restores are not good enough because if you mess with device drivers and such, the restores do not always work. You need a full disk image. Like the one I took just now, before upgrading to Dragon 15.6.1. We all have such confidence in new software installations, don't we? :-)
Anyhow, 15.6 seems to be working fine so far. I installed it for the new Basic engine. At least for the moment, I cannot really feel a difference in speed. I am just grateful that the installation worked without problems. Hopefully, I should be good for another couple of years until the next version comes out. ------------------------- Win10/x64, AMD Ryzen 7 3700X, 64GB RAM, Dragon 15.61, SP 6 PRO, SpeechStart, Office 365, KB 2017, Dragon Capture, Samson Meteor USB Desk Mic, Klim and JUKSTG earbuds with microphones, 3 BenQ 2560x1440 monitors, Microsoft Sculpt Keyboard and fat mouse |
|
|
|
![]() |
|
I use Acronis True Image to backup my system. I recently had to restore an old backup or I had installed Dragon. Before doing so, I backed up my custom vocabulary, my commands and my user profile. After recovering the system from the Acronis backup, I installed DPI 15.61, imported my user profile, my commands, as well as the custom word list backup files. Now I'm stuck in retraining hell. I have to retrain all of my custom words, and my commands. Dragon does not recognize them. Also, Dragon appears to be making tons of errors when I dictate. It's as if I'm starting with a completely new user profile.
Going back to old backups of my system is something I do regularly. Are there any other users out there who also do this, and does anybody have a solution for escaping from retraining hell?
My assumption is that there is no workaround for this. However I'm hoping that somebody might have some advice. Any advice you can give would be much appreciated. :-)
1. When you export a whole profile, you export all your custom words and commands with it. Hence, no need to import them a second time into the profile. 2. Your custom words and commands come handy when you create a new profile from scratch. Import them and dictate away. Training should not be necessary except for a few cases when two words are so similar to each other that they get easily confused.
To get working, I recommend creating a new profile and importing your custom words and commands, then see what happens. If your audio device is indeed broken, you'll get bad recognition from the start. If the audio device is fine, your dictation should be fine, too.
As to backups, everything has been said.
Hope that helps, Stephan ------------------------- |
|
|
FuseTalk Standard Edition v4.0 - © 1999-2021 FuseTalk™ Inc. All rights reserved.