minhleduc's picture
E2 run e2_laya_rlce_seed42 (epoch_4.pt)
2fc16af verified
Raw History Blame Contribute Delete
2.6 kB
{
"raw": {
"ece": 0.1311461047679186,
"brier": 0.05273624225749334,
"nll": 0.8640368629889563,
"accuracy": 0.774,
"soft_accuracy": 0.5199733426012099,
"mean_confidence": 0.6428539008647204,
"score_mae": 0.22571986983093667,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.1401130052407582,
"brier": 0.06713905334344265,
"nll": 0.9959831230528653,
"accuracy": 0.7383333333333333,
"soft_accuracy": 0.47646665925780934,
"mean_confidence": 0.6193402380744616,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.1569210796058178,
"brier": 0.054211134814122486,
"nll": 1.0368544653337448,
"accuracy": 0.72875,
"soft_accuracy": 0.4359412415418774,
"mean_confidence": 0.5718288956955075,
"score_mae": 0.22571986983093667,
"n": 800
},
"noul": {
"ece": 0.10893244693676629,
"brier": 0.03636690776270516,
"nll": 0.5016671331319958,
"accuracy": 0.87,
"soft_accuracy": 0.6755228273570537,
"mean_confidence": 0.7610675705472628,
"score_mae": null,
"n": 600
}
}
},
"post_temperature": {
"ece": 0.1428269161581993,
"brier": 0.05156653890298763,
"nll": 0.8613705683182925,
"accuracy": 0.774,
"soft_accuracy": 0.5142253369018436,
"mean_confidence": 0.6311731157004833,
"score_mae": 0.2258651484287111,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.14583302150170008,
"brier": 0.06465582661260517,
"nll": 0.9904372099414468,
"accuracy": 0.7383333333333333,
"soft_accuracy": 0.4661055313050747,
"mean_confidence": 0.5978283456961314,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.15982799533754585,
"brier": 0.053903129071204606,
"nll": 1.0361163761094212,
"accuracy": 0.72875,
"soft_accuracy": 0.4346355589479208,
"mean_confidence": 0.568922016248107,
"score_mae": 0.2258651484287111,
"n": 800
},
"noul": {
"ece": 0.12248062272866568,
"brier": 0.03536179763574747,
"nll": 0.4993095163069665,
"accuracy": 0.87,
"soft_accuracy": 0.6684648464371761,
"mean_confidence": 0.7475193516413371,
"score_mae": null,
"n": 600
}
}
},
"fitted_temperature": {
"type0_bucket1": 1.0932239294052124,
"type2_bucket0": 1.109182596206665,
"type1_bucket1": 1.0129419565200806
}
}