dataset03b-barrel-sparse-small.py

#!/usr/bin/env python

# datasets for trainings starting from 2016-05-09

dataDesc = dict(

    train_files = [ '../data/2016-05-larger-window/GJet20to40/rechits-barrel-train.t7',
                    '../data/2016-05-larger-window/GJet20toInf/rechits-barrel-train.t7',
                    '../data/2016-05-larger-window/GJet40toInf/rechits-barrel-train.t7'
                    ],

    test_files  = [ '../data/2016-05-larger-window/GJet20to40/rechits-barrel-test.t7',
                    '../data/2016-05-larger-window/GJet20toInf/rechits-barrel-test.t7',
                    '../data/2016-05-larger-window/GJet40toInf/rechits-barrel-test.t7'
                    ],

    # input dimensions
    nfeats = 1,
    width = 7,
    height = 23,

    # for shifting 18,18 to 4,12
    recHitsXoffset = -18 + 4,
    recHitsYoffset = -18 + 12,

    inputDataIsSparse = True,

    # if one specifies nothing (or nil), the full sizes
    # from the input samples are taken
    # 
    # if one specifies values < 1 these are interpreted
    # as fractions of the sample
    # trsize, tesize = 10000, 1000
    # trsize, tesize = 0.1, 0.1
    # trsize, tesize = 0.01, 0.01
    # 
    # limiting the size for the moment because
    # with the full set we ran out of memory after training
    # on the first epoch
    # trsize, tesize = 0.5, 0.5
    
    trsize = 0.05, 
    tesize = 0.05,

    # trsize, tesize = 100, 100

    # DEBUG
    # trsize, tesize = 0.01, 0.01
)