Click here to register.

Audio and Prompts Discussions

Flat
Sequitur (G2P) symbol out of range error
User: Noah
Date: 11/28/2013 12:20 am
Views: 4186
Rating: 6

Dear all

I try to run sequitur (G2P) on my language dictionary ( 44 alphabet, 30400 word).

But when I run, it return error

 

Traceback (most recent call last):
  File "/usr/local/lib/python2.7/site-packages/sequitur.py", line 662, in run
    shouldStop = self.iterate(context)
  File "/usr/local/lib/python2.7/site-packages/sequitur.py", line 575, in iterate
    self.shallUseMaximumApproximation)
  File "/usr/local/lib/python2.7/site-packages/sequitur.py", line 260, in evidence
    for eg in self.graphs(model):
  File "/usr/local/lib/python2.7/site-packages/sequitur.py", line 202, in makeGraphs
    eg = self.builder.create(left, right)
  File "/usr/local/lib/python2.7/site-packages/sequitur_.py", line 145, in create
    def create(self, *args): return _sequitur_.EstimationGraphBuilder_create(self, *args)
ValueError: symbol out of range: 256
iteration failed.
failed to estimate or load model

This error is come from more alphabet or word in dictionary ???

So, could you mind to help me solve this error.

Regard.

 

--- (Edited on 11/28/2013 12:20 am [GMT-0600] by Noah) ---

Re: Sequitur (G2P) symbol out of range error
User: nsh
Date: 11/28/2013 7:01 pm
Views: 139
Rating: 5

This is a bug in Sequitur

 

To fix it you need to change the following code in Multigram.hh:

#if (MULTIGRAM_SIZE < 3)

typedef u8 Symbol;

#else

typedef u16 Symbol;

#endif

 

To the following:

 

#if (MULTIGRAM_SIZE < 2)

typedef u8 Symbol;

#else

typedef u16 Symbol;

#endif

 

Essentially you need to change 3 to 2

 

--- (Edited on 11/29/2013 04:02 [GMT+0300] by nsh) ---

Re: Sequitur (G2P) symbol out of range error
User: Noah
Date: 11/28/2013 10:32 pm
Views: 1000
Rating: 6

Dear  nsh

Thank you for your help.

This problem have been solved.

--- (Edited on 11/28/2013 10:32 pm [GMT-0600] by Noah) ---

Re: Sequitur (G2P) symbol out of range error
User: mousmi
Date: 7/31/2014 10:34 pm
Views: 144
Rating: 6

Same problem arises with me but my input data is 1,063,989 words..

 I changed MUltigram size from 3 to 2, still it doesn't work..

Please suggest me...

Thanks in advance..

--- (Edited on 7/31/2014 10:34 pm [GMT-0500] by mousmi) ---

Re: Sequitur (G2P) symbol out of range error
User: mousmi
Date: 7/31/2014 10:35 pm
Views: 1315
Rating: 7

Same problem arises with me but my input data is 1,063,989 words..

 I changed Multigram size from 3 to 2, still it doesn't work..

Please suggest me...

Thanks in advance..

--- (Edited on 7/31/2014 10:36 pm [GMT-0500] by mousmi) ---

PreviousNext