Problem with CMU language modeling toolkit

[xavic383]

Hi,
I get the following error while executing "cat a.text | text2idngram -vocab a.vocab > a.idngram":

rr_fopen: problems opening /usr/tmp/text2idngram.temp for "w".

Thanks !!!

[xavic383]

DON'T PAY ATTENTION TO MY LAST QUESTION !! I wrote the command wrong.

To solution the main problem, I changed the path of the temporary files through the -temp flag:

cat a.text | bin/text2idngram -vocab a.vocab -temp aaa/sphinx/CMU-Cam_Toolkit_v2/tmp > a.idngram

but the problem persists...:

Sorting n-grams...
Writing sorted n-grams to temporary file aaa/sphinx/CMU-Cam_Toolkit_v2/tmp/text2idngram.temp.tes-150.6050.1
rr_fopen: problems opening aaa/sphinx/CMU-Cam_Toolkit_v2/tmp/text2idngram.temp.tes-150.6050.1 for "w".

Any idea ? Thanks !!

[xavic383]

All right,

I wrote the path to the temporary files wrong.

Problem fixed ;)

[Nickolay V. Shmyrev]

> rr_fopen: problems opening /usr/tmp/text2idngram.temp for "w".

You don't have permission to open this file, change permissions for the folder /usr/tmp/ or create it if it's missing

> Thanks !!!

NP!!!

[xavic383]

Thanks, this problem seems to be solved, but I get this new error message when I try to execute the same command:

" rr_iopen: None of 'a.vocab' 'a.vocab.Z' or 'a.vocab.gz' exist "

What can I do ?

Thanks !!

[Joseph Kurtz]

I'm getting the same error.
Were you able to resolve this error?

[Joseph Kurtz]

Resolved by using the -temp /home/xxxx

[Вадим]

hi,
try to just create a folder /usr/tmp - it helped me.

[jason]

If you're using Unix, just use the -temp flag with a . to signify using the
current directory for the temp files.

I built my models via

cat corpus.txt | ./text2idngram -vocab corpus.vocab -temp . -buffer 1000 > corpus.idngram

Since I used pretty large sources for model building, I needed to increase the
memory allocation to the program, thus the -buffer flag.