Home › Forums › OpenEars › [Resolved] Dynamic Language Model creation failling when dictionary >250 entries

This topic has 2 voices, contains 5 replies, and was last updated by  Halle 188 days ago.

Viewing 6 posts - 1 through 6 (of 6 total)
Author Posts
Author Posts
November 11, 2011 at 10:39 pm #7941

culov

Hi, first of all, thanks for the great library — it’s really be a lifesaver for me.

Here’s the stack trace I’m getting:

OPENEARSLOGGING: Starting dynamic language model generation
2011-11-11 13:33:04.530 FlexBud[20709:707] OPENEARSLOGGING: Running MITLM
2011-11-11 13:33:04.837 FlexBud[20709:707] (
0 FlexBud 0x0000a699 -[OpenEarsListener createDynamicLanguage] + 2340
1 FlexBud 0x00002c9d -[OpenEarsSampleProjectAppDelegate application:didFinishLaunchingWithOptions:] + 296
2 UIKit 0x3170f7eb -[UIApplication _callInitializationDelegatesForURL:payload:suspended:] + 1186
3 UIKit 0x317093bd -[UIApplication _runWithURL:payload:launchOrientation:statusBarStyle:statusBarHidden:] + 396
4 UIKit 0x316d7921 -[UIApplication handleEvent:withNewEvent:] + 1016
5 UIKit 0x316d73bf -[UIApplication sendEvent:] + 54
6 UIKit 0x316d6d2d _UIApplicationHandleEvent + 5808
7 GraphicsServices 0x32731e13 PurpleEventCallback + 882
8 CoreFoundation 0x36ad3553 __CFRUNLOOP_IS_CALLING_OUT_TO_A_SOURCE1_PERFORM_FUNCTION__ + 38
9 CoreFoundation 0x36ad34f5 __CFRunLoopDoSource1 + 140
10 CoreFoundation 0x36ad2343 __CFRunLoopRun + 1370
11 CoreFoundation 0x36a554dd CFRunLoopRunSpecific + 300
12 CoreFoundation 0x36a553a5 CFRunLoopRunInMode + 104
13 UIKit 0×31708457 -[UIApplication _run] + 550
14 UIKit 0×31705743 UIApplicationMain + 1090
15 FlexBud 0x000029ed main + 60
16 FlexBud 0x000029ac start + 40
) EXCEPTION INFO: *** -[__NSCFString substringToIndex:]: Range or index out of bounds

It’s very strange because I was previously able to generate a language of over 600 entries using this dictionary, but the limit has decreased. I’m using a database with a number of phrases to generate the language model. Does anyone have any idea what may be causing this?
November 11, 2011 at 10:57 pm #7942

culov

One more quick point — Most of the time 250 and 200 are too large of a data set for the language to properly understand, so I have to reduce to size to 100. It’s really strange how the maximum size of the language seems to keep changing so often.
November 11, 2011 at 11:04 pm #7943

Halle

OK, that isn’t related to the size of the language model, it has something to do with the search through the cmu07a.dic file. Can you go to the contact form and email me a sample corpus that you are trying to create a language model for so I can attempt to replicate? Also, are you using GCC LLVM or Apple LLVM? OpenEars hasn’t been made compatible with Apple LLVM yet.
November 11, 2011 at 11:54 pm #7944

culov

Thanks so much for the reply, I’m sending over my sample corpus right now. I was using Apple LLVM but I reverted to GCC LLVM and I’m seeing the same error.
November 12, 2011 at 12:47 am #7945

culov

Halle,

I figured it out — some words in my corpus has an extra space between them. Thanks so much for your help!
November 12, 2011 at 10:43 am #7946

Halle

Excellent, glad it’s working.
Viewing 6 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic.