<HTML>
<font style="font-family:tahoma;font-size:10pt;"><font
style="font-family:tahoma;font-size:10pt;">hi
<br />I made a bigram <span style="border-bottom: 2px dotted rgb(54, 99,
136); cursor: pointer; background: none repeat scroll 0% 0% transparent;"
class="yshortcuts" id="lw_1308169721_0">language model</span> on a small
document with both HTK toolkit and SRILM toolkit.
<br />but unfortunately their accuracy in HTK based ASR are so different. I
mean SRILM bigram has 10 percent accuracy lower than one with HTK.
<br />Can you help me where is my mistake?
<br />
<br />Here is my command for build bigram in SRILM:
<br /> ngram-count -text sentences.txt -order 2
-wbdiscount 1 -wbdiscount 2 -lm bigram.txt
<br /> sentences.txt has 405 sentences.
<br />
<br />I build my <span style="border-bottom: 2px dotted rgb(54, 99, 136);
cursor: pointer; background: none repeat scroll 0% 0% transparent;"
class="yshortcuts" id="lw_1308169721_1">acoustic model</span> based on left
to right HMMs with 2 Gaussian mixture for triphones using HTK.
<br />Someone said sentences.txt and i build my model on that.
<br />
<br />Best Regards </font></font>
</HTML>