minhleduc's picture
E2 run e2_rlce_sigma1_fixed_seed43 (epoch_4.pt)
5ebcbf5 verified
Raw History Blame Contribute Delete
2.59 kB
{
"raw": {
"ece": 0.0867420684248209,
"brier": 0.06649116362994949,
"nll": 0.89247768904455,
"accuracy": 0.761,
"soft_accuracy": 0.5318410232178867,
"mean_confidence": 0.6776189638525247,
"score_mae": 0.24290502518531867,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.08772538999716441,
"brier": 0.08208390673490006,
"nll": 1.025410343380645,
"accuracy": 0.7283333333333334,
"soft_accuracy": 0.4865714711571733,
"mean_confidence": 0.650977836449941,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.1000199832767248,
"brier": 0.0700340546910229,
"nll": 1.0721786250313743,
"accuracy": 0.7225,
"soft_accuracy": 0.45394872039556505,
"mean_confidence": 0.6232056948170066,
"score_mae": 0.24290502518531867,
"n": 800
},
"noul": {
"ece": 0.06818889498710633,
"brier": 0.04617456577690099,
"nll": 0.5199437867260227,
"accuracy": 0.845,
"soft_accuracy": 0.6809669790416956,
"mean_confidence": 0.7768111166357994,
"score_mae": null,
"n": 600
}
}
},
"post_temperature": {
"ece": 0.1396139353811741,
"brier": 0.05708729153524543,
"nll": 0.87030091369804,
"accuracy": 0.761,
"soft_accuracy": 0.5069854698851705,
"mean_confidence": 0.6232002761662007,
"score_mae": 0.2397035081544891,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.13807830850283304,
"brier": 0.07131863028705993,
"nll": 1.0011274497335156,
"accuracy": 0.7283333333333334,
"soft_accuracy": 0.46080951342980064,
"mean_confidence": 0.5945008865992228,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.1599178895540535,
"brier": 0.05838104838370782,
"nll": 1.045214518448338,
"accuracy": 0.7225,
"soft_accuracy": 0.4285406183823943,
"mean_confidence": 0.5639332137256861,
"score_mae": 0.2397035081544891,
"n": 800
},
"noul": {
"ece": 0.11407759229342143,
"brier": 0.04113094365214773,
"nll": 0.5062562379955003,
"accuracy": 0.845,
"soft_accuracy": 0.6577545616775751,
"mean_confidence": 0.7309224156538645,
"score_mae": null,
"n": 600
}
}
},
"fitted_temperature": {
"type0_bucket1": 1.2614489793777466,
"type2_bucket0": 1.4392832517623901,
"type1_bucket1": 1.284111738204956
}
}