Author Archives: underwhelmingforce

Jokebot redux

Posted on January 22, 2017 by underwhelmingforce

Jokebot has finished training. It still has some serious issues. Perhaps the issues are even more serious. DISCLAIMER: The “jokes” the network produces do not represent my views or sense of humor. The data was scraped from reddit and it shows. … Continue reading →

Posted in Uncategorized | Leave a comment

Jokebot

Posted on January 22, 2017 by underwhelmingforce

I fixed the problems! It seemed that I needed to update Cutorch, and in order to do that, I had to update CUDA, and in the process I inadvertently uninstalled my graphics drivers. What an adventure. To celebrate, I’ve been … Continue reading →

Posted in char-rnn | Leave a comment

Jeopardy!

Posted on January 12, 2017 by underwhelmingforce

While I’ve been working out some issues I’ve been having with torch, I did some training on a database of all Jeopardy questions. Unfortunately, training was cut short by said torch problems, so I’ll have to resume that tonight. Here’s … Continue reading →

Posted in Uncategorized | Leave a comment

Machine Learning vs. Human Learning

Posted on November 20, 2016 by underwhelmingforce

I’ve been a bit busy to run any experiments, unfortunately, but I’ve still been thinking about this quite a bit. Since I haven’t posted in about a month, I figured I’d share one of my motivations for getting into machine … Continue reading →

Posted in Uncategorized | Leave a comment

Encouraging results with Wavenet!

Posted on October 22, 2016 by underwhelmingforce

After doing some digging into the code and resolving an error that caused the network to devolve into white noise, then talking with some other  folks about what seemed to work for them, along with a whole bunch of hyper-parameter … Continue reading →

Posted in Wavenet | Leave a comment

Wavenet!

Posted on October 2, 2016 by underwhelmingforce

So I got wavenet working. Well, for a limited definition of “working.” I’ve been running some experiments to try to figure out how the model behaves and unfortunately I don’t seem to have a very good grasp of the underlying … Continue reading →

Posted in Wavenet | Leave a comment

Parameter Experimentation (initialization and merging styles)

Posted on September 12, 2016 by underwhelmingforce

While I wait to get my hands on Deepmind’s wavenet, I’ve been experimenting some with the parameters of neural-style. Here, I have two images that use the same style and content image, however in the first one, the result is … Continue reading →

Posted in Uncategorized | Leave a comment

Hallucinating words

Posted on August 26, 2016 by underwhelmingforce

After a few failed experiments (namely, trying to train char-rnn on the grant award database to discover that the file has been corrupted with non-unicode characters which caused loss to explode and training on the list is kernel source, which … Continue reading →

Posted in Uncategorized | Leave a comment

The adventures of Robo-Chef

Posted on August 16, 2016 by underwhelmingforce

Since my previous post, I’ve gotten the model better.  How much better? About 22%! That’s pretty exciting. Here, exclusive, and brand new, are some highlights and some dumps from this 404-star recipe bot. At temperauture=0.1, we get the least creative, … Continue reading →

Posted in char-rnn | Tagged | Leave a comment

Char-rnn retro-post

Posted on August 16, 2016 by underwhelmingforce

As with my previous post, this post is intended to serve as more permanent storage for my previous posts regarding the topic at hand– in this case, Andrej Karpathy’s infamous Char-RNN. First introduced to the world in this incredible post … Continue reading →

Posted in char-rnn | Tagged , , , , | Leave a comment