minhleduc's picture
E2 run e2_ce_only_seed43 (epoch_4.pt)
0464cbd verified
Raw History Blame Contribute Delete
2.64 kB
{
"raw": {
"ece": 0.14004284296929834,
"brier": 0.05231113771935242,
"nll": 0.8615606835968792,
"accuracy": 0.7785,
"soft_accuracy": 0.5190770915709436,
"mean_confidence": 0.6384571585655212,
"score_mae": 0.22389980303356424,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.15449575329820314,
"brier": 0.06391233971989853,
"nll": 0.9867078996635974,
"accuracy": 0.7666666666666667,
"soft_accuracy": 0.47517100428541503,
"mean_confidence": 0.6121709233522415,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.1481409678608179,
"brier": 0.05756362313889895,
"nll": 1.0408445472083985,
"accuracy": 0.72375,
"soft_accuracy": 0.4375115679483861,
"mean_confidence": 0.5756090462207795,
"score_mae": 0.22389980303356424,
"n": 800
},
"noul": {
"ece": 0.11479244967301687,
"brier": 0.03370662182607759,
"nll": 0.49736831604813536,
"accuracy": 0.8633333333333333,
"soft_accuracy": 0.6717372103532155,
"mean_confidence": 0.7485408769051234,
"score_mae": null,
"n": 600
}
}
},
"post_temperature": {
"ece": 0.15468624928593636,
"brier": 0.0509832072927122,
"nll": 0.8585635291822254,
"accuracy": 0.7785,
"soft_accuracy": 0.5120872111618519,
"mean_confidence": 0.6238137583583593,
"score_mae": 0.2241187854518648,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.17406129206220305,
"brier": 0.062145273851042473,
"nll": 0.9828057477561136,
"accuracy": 0.7666666666666667,
"soft_accuracy": 0.46572454484800496,
"mean_confidence": 0.5926053783297539,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.16359852660447363,
"brier": 0.05573118854731717,
"nll": 1.0368128461157904,
"accuracy": 0.72375,
"soft_accuracy": 0.4306440041400492,
"mean_confidence": 0.5601514636352658,
"score_mae": 0.2241187854518648,
"n": 800
},
"noul": {
"ece": 0.12369074026743572,
"brier": 0.033490499061575266,
"nll": 0.4966555546969175,
"accuracy": 0.8633333333333333,
"soft_accuracy": 0.6670408201714356,
"mean_confidence": 0.7399051980177561,
"score_mae": null,
"n": 600
}
}
},
"fitted_temperature": {
"type0_bucket1": 1.0869410037994385,
"type2_bucket0": 1.0710346698760986,
"type1_bucket1": 1.0724668502807617
}
}