Speech:Spring 2016 Saverna Ahmad Log
- Information - General Project Information
- Experiments - List of speech experiments
Week Ending February 9, 2016
I begun this course with the expectation that Capstone was a class. I quickly learnt that this is a team project that involves heavy collaboration with group members and it is expected that we each give our best to be able to perform successfully.
- During the past two weeks I have been reading logs from previous semesters. My goal for this week is to become more familiar with redhat and using linux. I also would like to go back and read some more on the systems group. To also confirm which machines Zack would be using for his project.
- So Far: Installed Putty and was able to access Caesars remotely.
Obelix, Idefix, and Brutus are going to be used by Zack. There has been useful resources for redhat and Linux, the next task would be to apply it in our project.
- To work on contents of a project proposal with the rest of the team within the next few weeks.
- None so far.
Week Ending February 16, 2016
Task: This week I spent some time with my group in the server room. One thing that came to our attention was the documentation on server labeling on Wiki seemed to be incorrect. The label placement was different on the servers in the server room. It is possible that changes were made on the servers but the labels were not correctly placed. This is something that needs to be confirmed. I formatted and shared the group log with my team members to further collaborate and edit before final upload. Today I was able to pick up the FOBS which is now attached to my UNH ID that would allow access the server room. I tested it and able to gain entry in the server room. I am currently am rereading documentation on wiki, the next thing to do is to run a train and decode.
Results: After comparing images of the server layout on Spring 2015 Capstone poster and reading documentation on Wiki, our suspicions were correct that the labels do not match the servers. Received an email from Professor Jonas that our server room access was granted and access cards were to be picked up from the security desk. I will get that tomorrow since I was not on campus last Friday. Having readily access to the server room will make things easier because we would be able to move our project along faster. We are hoping to clone the drones in a very short time with the plan of using a thumb drive for each server upload.
Plan: Upon completing confirmation of identifying correct servers we will then attempt to update the server layout on wiki since the current documentation about labels do not match the labels in the server room. At the moment we are not sure server is which is which. We will certainly do the necessary update on wiki as soon as we figure out and confirm.
Concerns: Hoping that cloning of the servers is a smooth run. Hoping that I can get my train to run next. Unfortunately Putty keeps timing out with it's connection. After numerous tries, I will have to continue to run the train experiment on campus.
Week Ending February 23, 2016
Task: This week our goal is to spend some time in the server room. By next week the System Group needs to install 5 machines and have 1 running with Red Hat. We have to figure out the IP for Asterix. Do documentation on all the steps taken while doing server installation. I also will be spending some time with my team members to work on our proposal for this project.
Results: Yesterday we spent the rest of the class time in the server room. We begun the process of removing the old servers and installing the new ones. There was just one screw driver to use. We were able to borrow a pair of pliers from Donna in the Graduate Office. Professor Jonas lent us a label machine which we have used to print and attach new labels on the new servers. So far Asterix was replaced with a new machine and named Asterix, we now have a new Obelix called Obelix 2 which replaced Traubadix and a new Edafix which we call Edafix 2 and a new machine replaced Miraculix with the same name Miraculix. We installed Red Hat 6.6 32bit on Asterix. We left it to upload, Neil and I checked on it during our other class. There were no issues. Aaron checked on it this morning and found no issues. He did a reboot to finish the installation process.
Plan: I will drop by the server room in the morning on my way to work to check if the installation was successful. On Saturday I spent sometime in the server room. My main goal was to continue installation of Red Hat on Miraculix. However, the software failed to install after continuous numerous tries. I kept getting an error that disc was not found. Met with Professor Jonas and did yet 2 more reboots, logged in as root and attempted to run Red Hat, but was still unsuccessful. Professor Jonas suggested to replace the disc driver and try again. So far only Asterix is running with Red Hat.
Concerns: There lights appear to be orange on the drivers and this is a possible sign of low battery which we will be replacing during this week.
Week Ending March 1, 2016
Task: So last week we were unsuccessful at installing Red Hat in all of the machines with the exception of Asterix which was already completed and had Red Hat running. We kept getting battery error message and upon running Red Hat we got the message 'No disk has been found.' We swapped old drives and installed new ones and still got the same error message of 'no disk has been found.' However, yesterday during class time the systems group configured the utility upon start up ( further documentation will be given) and then did a Red Hat install after.
Results: Our first machine attempted was Majestix and then, no error messages! It was a break through. Majestix is now on the network but has not been configured to the network because this one would be used by the other group to run experiments and tests etc on. Now each machine is clearly labeled. Thanks to the artist work of Aaron and Neil. Editing and formatting of the project proposal was a team collaboration. It was completed and sent over to Jon for final formatting and upload to the Wiki page.
Plan: To spend some time on Saturday to continue installation of Red Hat on the other machines and make last minute tweaks to the proposal before Professor Jonas looks at it on Sunday. Was not able to go on campus today. I will be doing that tomorrow to continue Red Hat installation on the other machines.
Concerns: Just a bit anxious if we would be as lucky with the future installations as we were with Majestix.
Week Ending March 8, 2016
The following are some of the things we have to work on this week. I will add more to the group log.
- NFS client/servers
- Run a 1 hour train on different servers and compare times to Caesar
- Add Banners on new servers
- Enable servers able to access /mnt/main and other data on Caesar
- Install Red Hat on all the servers
Red hat is installed on all of the servers. mnt/main was installed by using documentation form Fall 2014.Please refer to systems group log for further detailed documentation. All of the above tasks are now completed.
To move the Systems Group wiki page for semester 2015 to Information page under software system setup. I will attempt that tomorrow. I added the page and when I hit save I noticed that the documentation wasn't able to be saved. I received a message from wiki that someone else was already documenting at that moment so there was a conflict. Today, I noticed that this was completed.
I still have not been able to run a train as yet. I would be collaborating with the model group during this week to get this done.
Week Ending March 22, 2016
- Task: My task this week is to do a 5 hour train.
- Results: Unable to run trains. Server was down.
- Plan: This week we were assigned in new teams The plan is to meet with members and understand what needs to be done.
- Concerns: None
Week Ending March 29, 2016
- Task: To install Red Hat on Rome and fix issues discussed in last class.
- Results: Rome is now running Red Hat.
- Plan: To continue getting better at running and decoding trains.
- Concerns:Fix connectivity issues on Majestix
Week Ending April 5, 2016
- Task: Meet with Ryan to go over train and documentation today. Check to see why Zack cannot log in to Idefix and Obelix.
- Results: Neil was able to ssh into both Idefix and Obelix and there was no issue, Zack should be all set. Met with Ryan, Mike and James who all
assisted in making sure that my train ran correctly.
I ran my first_5 hr train today. However, I got the following error
FATAL_ERROR: "mdef.c", line 680: No mdef-file. I believe this is because I changed 1000 to 500 with the idea that the train would run faster. I ended up
using the original script on the wiki page: /mnt/main/scripts/user/parseDecode.pl decode.log ../etc/hyp.trans and decoding was successful. Below is the
| =================================================================| | Sum/Avg | 500 6391 | 80.4 13.5 6.1 8.8 28.4 82.8 | |=================================================================| | Mean | 31.3 399.4 | 81.2 13.5 5.4 9.5 28.3 82.2 | | S.D. | 17.5 228.3 | 5.8 3.5 3.1 6.5 9.6 13.9 | | Median | 26.5 344.0 | 83.7 14.0 4.3 7.2 25.8 85.4 | `-----------------------------------------------------------------'
- Plan: To spend time with the Captain America group this week and get our game on.
- Concerns: After many unsuccessful attempts to log in to Caesar, I was finally able to get connected and run my first train today. Thanks to Ryan who
helped in resetting my password. Previously, I was being denied permission to log in for some unknown reason.
Week Ending April 12, 2016
This week Captain America continues to collaborate our "secret sauce." Each group member was assigned a specific task. Research was completed and a timeline was made to follow when each goal would be met. I am limited to log the details of what the group activity was.
Over the weekend students complained of power failure on the servers. The Systems Group is faced with the issue to find out what's causing the power failure. Neil was able to the network bridge documentation. The Systems group poster for the URC needs to be completed for Monday morning.
URC poster was worked on. Did some research for my group. I am limited to give details.
To continue working with the system group and Captain America group this week.
Nothing as of right now.
Week Ending April 19, 2016
My tasks for this week would be to continue doing research for Captain America and collaborate with my team via email and meeting. I have to make and document content on a spreadsheet for our train logs in my team. I also would be doing editing for the Research log along with Brian. I feel limited as to what to write in this log since most of the content and research is kept under wraps within the team.
With regards to the systems group Caesar has been really slow. I am at work so Neil is attempting to figure it out. Students have had difficulty connecting to Caesar from cisunix.unh.edu. There seems to be no issues internally with Caesar. There is definitely something within the connection between Caesar and cisunix. It was noticed that the DNS was slow. Captain America has a decode running during the time Caesar is acting up. Systems group will wait until this is finished before rebooting Caesar. The network bridge on Caesar has been disabled would be restarting Caesar within the hour. Today is the URC conference. I will be presenting my Capstone poster and another poster from Comp 730 with my team members.
After rebooting Caesar was back up and running but still slow. We can rule out that it’s not a VGA issue since that port is back up and working. Caesar is fixed for now...
For next week Cap. America Team will continue to improve and test and collaborate.
It looks like Caesar has the issue on Caesar’s side because once connection is made it takes a longtime for it to get to prompt. This is a different issue than the DNS previously.
Week Ending April 26, 2016
The goal for the Systems Group this week is to get IRC running on Rome.In order to do this there needs t be a dedicated line to the internet however, this line is currently hooked up to Majestix. Confirmation needs to be made if the Tools group is still using Majestix. An email was sent by the Systems Group, we are currently awaiting a reply from the Tools Group. With regards to our Captain America group we are collaborating via email and group meeting to communicate goals and reach our tasks. I am continuously working in keeping the spreadsheet updated with new data.
Turns out that the Tools Group no longer needs to use the connection on Majestix. The Systems Group will attempt to get Rome up on the internet today. If that is successful then we will attempt to get the IRC server up and running tomorrow. Rome is now on the internet.
To complete document for Captain America group and research parameters.
There might not be sufficient time to get the IRC up and running since there are only two weeks left in the semester.
Week Ending May 3, 2016
So this week both teams have decided to call a truce and actually come together as a class instead of having a competition between the previously
named groups - Iron Man and Captain America. We the entire class will all collaborate to the end of the semester to continue to come up with the world
class base line for the speech project where finding the lowest word error rate for unseen is a top priority. During this week and next week we will
continue to run experiments to meet this goal. Lastly we need to write the final proposal. As the semester is coming to a close, the Systems Group has
to ensure that all of the hardware and software documentation is up to date and current for the next group of students would would be taking capstone
in the future.
We continued to get different WER %unseen results from the many parameters that we have been using. All of this documentation will be in the final
The top plan is to do the experiment result write up for next Wednesday along with preparing the final report.
There are no concerns at the moment.
Week Ending May 10, 2016
I cannot believe this will be my final log. This log writing has become part of my routine for the past fifteen weeks. Capstone was an interesting
journey and a phenomenal experience. Today I begun working on the final report with Ben. By the next Wednesday we should have a document put together
to review before the final iteration is handed in on May 27th. We will be expected to complete evals for our classmates at the end of the last class.
This week was pretty much working on the Team Results Report and the Systems Group Report which will go .in to the final report. Our final evals are in.
Plan: To continue working on the final report along with the rest of the class on Wiki.
Concerns: None. It was a great semester! Wonderful capstone class.