For some of the language models i create, I get a warning message saying "-mmap specified, but tseg_base is not word-aligned" I checked the code in ngram_model_dmp.c and see where it is happening but I am not able to figure out why this is happening. The output I get is complete junk too. Are they related? Any help on this will be wonderful.
Thank you,
Jagadeesh
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I did more testing with this and I get garbage output for more than 50% of my test sentences and out of 14 sentences only one is recognized correctly. These are the same sentence by different speakers and each sentence consists of the same word repeated twice. What does the garbage characters mean? Does it mean that the search did not converge to any word or is there something going wrong drastically?
I am using a 1-gram language model with 5000 words. The hmm model is adapted from the Communicator 8khz open source model.
Thanks for your help,
Jagadeesh
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi,
For some of the language models i create, I get a warning message saying "-mmap specified, but tseg_base is not word-aligned" I checked the code in ngram_model_dmp.c and see where it is happening but I am not able to figure out why this is happening. The output I get is complete junk too. Are they related? Any help on this will be wonderful.
Thank you,
Jagadeesh
Hi Nickolay,
I uploaded both the language model file to http://w15.easy-share.com/1701977175.html . Is there any other way I can send the file over to you?
I did more testing with this and I get garbage output for more than 50% of my test sentences and out of 14 sentences only one is recognized correctly. These are the same sentence by different speakers and each sentence consists of the same word repeated twice. What does the garbage characters mean? Does it mean that the search did not converge to any word or is there something going wrong drastically?
I am using a 1-gram language model with 5000 words. The hmm model is adapted from the Communicator 8khz open source model.
Thanks for your help,
Jagadeesh
Is there any chance to get the model that caused this?