This is an instruction video on how you can run the transcription with the software Whisper on audio data inside TSD.
I will continue in this video that you have a TSD project that you have managed to log into.
What we are going to do first is to copy the Whisper software via a server.
To log in to the server, you must go into a shell and log in to the server.
We use PuTTY in this example. There are several options, but we use PuTTY, which is inside all Windows machines in TSD.
To find it, I can use search and search for PuTTY.
I start PuTTY and it looks like this.
What we are going to log in to now is a machine called Submit.
In front of Submit, you must enter your project number.
For me, in this example, it will be p896-submit.
This is the server I'm going to log in to to run the commands and copy the software.
I press open and log in as my user.
In here, I have to run some Unix commands.
What I do in my example is that I want to find the Durable map, because that's where I want to put Whisper on the computer.
Default, I came into a user map. To see where it is, I can type p-double-d.
Then I see that it is a home area.
I can go into Durable and type cd-prick-prick. Then I come up one level.
Then I can do it again and type p-b-d.
Then I see that I'm right in my project.
If I type ls, I can get some areas here.
Then I know that if I type cd-data, then I come into data and can check again with ls.
Then I see that I have Durable. I type cd-durable.
I can do an ls and I see that I recognize my, for example, network attachments and network data folders that I have in Durable today.
Now I'm going to copy this Whisper catalog, which is in a shared folder on a server, into this area.
Until then, everyone must run this software locally.
What can be done is to simply copy the command that is in the video and paste it.
Because it is a copy command called cp-r to be able to have with some sub-catalogs.
It is in the shared software Whisper.
It is very copied. If it says in the middle of a point, it will end up where I am right now, which is in the Durable folder.
This means that the catalog is now copied right under Durable.
This will take some time, as there are some gigabytes of data, but that was it.
Now I would like to go over to use explorers, because that's what you can do with File Explorer.
You have probably used it before.
I can go into my data folder on Durable.
There is now a catalog called Whisper.
Inside this catalog, there is now a script and a program.
You can see that there is something called large.pt.
This is the largest output of Whisper, which needs the most possible data power.
In this case, it is the one with the best solution.
We think that's the best right now.
The data folder is mapped and the script is set up to take all the files that are in the folder and transcribe them.
I'm going to find some data and I have some data that I have ready.
A sound file and a video file.
I can copy that and go into the data folder and paste them in.
Now I want these to be transcribed and I have to run the script called transcribe data.
This script can only be run from the Linux server that I have logged in to.
I can't run it in Windows Explorer, but on this right server.
It is important that you have access to Colossus, i.e. the IT system, to be allowed to run this program.
The way to run this program is to write.score or transcribe data.
The command is together with the video here and you can paste it into PuTTY.
You can always paste data into TST, but you can't copy data out.
That's a trick to get commands.
If I run the command now, which is.score or transcribe data.
I got an error message here, but when I type.pbd it is because I'm in the wrong folder.
So I have to type cd whisper.
I have to be in the whisper catalog.
If I type the same command now, it says submitted batch job and a number.
It means that it has been put in queue at the computer on Colossus.
I have to wait for it to be finished.
This can depend on the traffic on the computer and how many files you have put into transcription.
Here I have put in both a sound file and a movie file.
Then it's just a matter of waiting a little while and I can check in the file explorer if this has arrived.
After a while, I can go into the folder data in the explorer and there I see that there are now three files.
Each file that I have put in.
This is what whisper today automatically sets up.
It's a text file, a VTT file and a SRT file.
With that I say good luck to transcribe data into TST.
It's smart to take all the data from here and move them afterwards so that you don't transcribe everything again.
Because in the data folder there must always be finished data.
I can cut this and then I could just make a folder here called finished data, for example.
And then I could say that I put those files there.
So that I know that when I run the script, everything that is in the data folder is transcribed.
Good luck.