uci-cbcl/DanQ

Data?

Closed this issue · 3 comments

I'd like to try your model - but I have no idea what sort of format it expect the data to be in. Any chance you can upload a sample to the GitHub?
Thanks.

Thank you for your interest in DanQ. Sorry for the delay, I've been traveling. Perhaps you can help me improve it. What exactly would you like to see? Right now, I have only made the training code available. If you want, I can write a script that will accept the weights from a fully trained model and output predictions from an input sample hdf5 file into an output hdf5 file.

Just..raw data so I can run it, and to see how to format my own data for it
On Jan 2, 2016 10:26 PM, "daquang" notifications@github.com wrote:

Thank you for your interest in DanQ. Perhaps you can help me improve it.
What exactly would you like to see? Right now, I have only made the
training code available. If you want, I can write a script that will accept
the weights from a fully trained model and output predictions from an input
sample hdf5 file into an output hdf5 file.


Reply to this email directly or view it on GitHub
#1 (comment).

Sorry about that. I did not update the README properly. I will do that later tonight. You can download the original .mat files from DeepSEA: http://deepsea.princeton.edu/media/code/deepsea_train_bundle.v0.9.tar.gz

After you extract the contents, you can move the 3 .mat files into a folder called "data/" and run the training scripts.