[hts-users:03329] About the Monophone Models
- Subject: [hts-users:03329] About the Monophone Models
- From: Kwan Lisa <lisakwan1102@xxxxxxxxx>
- Date: Wed, 6 Jun 2012 17:07:15 +0800
- Delivered-to: hts-users@xxxxxxxxxxxxxxx
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:date:message-id:subject:from:to:content-type; bh=ISSdbD1VlkeX7gKHMwd6EIhvchGN0q8j3l3Duyd1iEQ=; b=UK/fXikmbQA/kckcipMgSN507csPh9l7YZUPfcuWHgZOeFWBuvrp0QrBX0OmBUq8WC AUeNnAoEMOeWhV3DkPZ6511xAciUFRVdzazW4OhnVfD8zHUV7cOkK7ZJmtCxvK3Ii/tx ssbZYfxQDf/Q6+1aos53D7FiFYebQ7n2iIMTB0u6HfDQTRwt8ojI0aqJk41IXQhoTtv2 qItK/fZvLAPTW29BUhkFnxYn5MwlZJqSXDA6Mdri71Vstpu6fbaygN3cthm1v09tJ8g1 Tek334R2XlT93J0VmMxH/nhsZMlDi2pPVLkVzweXj6z1vEWGsx+hM7YIbMXUCXSx8CqF UOrQ==
Hi,
I'm conducting a experiment using monophone models. I found that the
<TRANSP> was not trained after the embedded re-estimation in monophone
models(models/cmp/monophone.mmf). The output was like:
<TRANSP> 7
0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00
0.000000e+00 0.000000e+00
0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00 0.000000e+00
0.000000e+00 0.000000e+00
0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00 0.000000e+00
0.000000e+00 0.000000e+00
0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 1.000000e+00
0.000000e+00 0.000000e+00
0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00
1.000000e+00 0.000000e+00
0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00
0.000000e+00 1.000000e+00
0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00 0.000000e+00
0.000000e+00 0.000000e+00
I'm wondering why this matrix was not trained.
--
Lisa Kwan
lisakwan1102(at)gmail.com
Advanced Speech Technology Lab, ASTL
- Follow-Ups
-
- [hts-users:03330] Re: About the Monophone Models, Simon King