Closed gngdb closed 9 years ago
Good example notebook would be for whatever model we manage to create our first submission with.
The notebooks Analysing Network
and Validation scoring
, along with various noteboooks looking at monitoring traces after runs should contain enough examples to close this issue.
There's a bunch of script in Pylearn2 for looking at the learned layers and other stats for models after training. We should look at these in IPython notebooks and then nbconvert them to markdown for the wiki. Then, we can keep records of anything interesting/useful in the project wiki with graphs etc for fairly little work.