[Peter Herrmann]

Dear Youssef,

I took a closer look at the MNIST CNN example and would like to load the original “ubyte” files of the MNIST data set (Train- and Test Dataset) into the model instead of the CSV file.
I’ve attached a snipped below. How can I exchange for the original?
I don’t understand the “Model.valid_rate” parameter. Does “Model.batch_size” mean Mini Batch Size?

Is it also possible to save the trained model and then use the ubyte test data to test how well the model works?

Thanks

Peter

Topic

[Youssef MENJOUR]

Dear Peter,

Valid Rate is the % validation dataset between train set and test set (I added comment for the next release).

We modified CNN_LoadDATA because there was a bug (fixed for the next release coming this week)

If you want to use your dataset in this example you’ll have to connect it directly to the split VI.

 

Be carefull you’ll have to keep the same dimension organisation of your data. (I don’t know how  “ubyte” file data are organized, logically it’s same as CSV).

• This reply was modified 1 year ago by Youssef MENJOUR.

[Author]

Replies