|
From: Dominic W. <dwi...@cs...> - 2007-09-11 14:24:07
|
Dear Kaveh, Thanks for persevering, sorry it's not working for you yet. Is your data in the public domain? The next thing you could try would be to send me the data or a link to it, and see if I get the same model_params.bin data. Good luck. Best wishes, Dominic On Mon, 10 Sep 2007, Kaveh Piroozram wrote: > > Hello & Thanks for your answer, > > I did a 'make clean' and later on even removed the whole > structure. and build it again and observed the whole > building process mentioned by you. nr(i). Everything went on > smoothly. > > later on i seperated the data directory (as ii) and > it is not located under nlp directory. I as well reset > one of the environmental variables to point to correct > locaion. > > As for (iii), the answer is positive. Those files contain > (word-statictics/counts) and list of all words. > > > (iv) > I did a 'strings' on model_params.bin and it contains > the following: > =================================================== > /home/kultur/C/Text_Processing/phase3/data/EU/dic > \Device\NamedPipe\Win32Pipes.00000e04.00000001 > EU.txt > invited > w aH > `Mh@ > aMh@ > a(| > aMh@ > aMh@ > aMh@ > =================================================== > (The file is 2+ kb though) > > The generated files are: > corpus_format.bin > dic > model_info.bin > model_params.bin > numDocs > wordlist > > and they seem to be ok. > > > Anyways, i better study some more math and implement something > on my own. Since most of people in this list never hit similar problems, > one can conclude that i obviously do something out of order ;-) > > Thanks for your time and commitment, > kave > > > -- > Kaveh Piroozram > pa...@em... > > -- > http://www.fastmail.fm - Accessible with your email software > or over the web > > |