Week 3 Progress Update

Week 3 progress update

        Most of the progress I made this week was just research, I focused on getting familiar with the TensorFlow framework. I made several neural networks that work in slightly different ways and using different data sets. I have not completely decided on a data set that I want to use but I am probably going to use Image Net as it provides a relatively large sample for each class as well as plenty of classes. During my research I also designed a small script for processing "pickled" data, which has been stored for use in python as well as a script for processing some particular data sets such as Mnist, both of these are shown below. I also did a bit of research on making a CNN in TensorFlow which I will start actually building over the course of the next week or two.


Comments

Post a Comment

Popular posts from this blog

Week 14 Progress Update

Week 14 Progress Update     This week's blog is going to be very short as I made literally no progress. This week I was planning on continuing to analyze data but unfortunately I was unable to gain anymore data due to my computer running unfathomably slow. As such I have simply been trying to speed up my computer over the last week all of which is menial tasks meaning I won't have any images or anythin to really show. I think it is now at a manageable speed meaning that by at least the middle of the week I will be able to continue with my plans, unfortunately this means I will have lost a week of data collection and analyzing which is rather frustrating.
Read more

Week 13 Progress Update

Image
Week 13 Progress Update     This week was mainly used to analyze the sessions that i have run so far. I came up with a system where I apply the weights to three different images, a checkered pattern, black and white noise, and a sample from the cifar10 database, more specifically a chicken. Applying the weights to these three images allows me to see the patterns that these filters "learn" to "look for".  While I don't have a particularly precise method of analyzing these new images, I have developed a system that allows for semi-consistent analysis of the images that can be compared across sessions. This system is in the form "x/n property (c/10)" where x is the number of images that appear to have the given property out of the total number n of filters in that layer. I say that these appear to have this feature with certainty c/10 where 10 is very certain and 1 is not certain at all. The document for one of my sessions is shown below.
Read more

Week 10 Progress Update

Image
Week 10 Progress Update Its been a while since I've posted an update and as such its not actually week10 anymore, however I have decided that it will make more sense if I just continue the current numbering. Over the past few weeks I have accomplished a lot. First of all, I got the convnet working. Instead of posting tons of pictures of my code I'm just going to upload it to Github and it will be visible at this link. This code will also contain all of the later updates I made to the code. Other than making the CNN work I implemented two main features, saving and loading kernels, and testing custom images. The kernel saving and loading lets me view the convolutional weights as well as allows for makeshift loading which allows for non-continuous training sessions, however, it is not meant for this and I'm not planning on using it for this because it could cause issues with accuracy. The custom image testing is not really important but it does give me a more accesible acc...
Read more