minhleduc's picture
E2 run e2_laya_rlce_seed43 (epoch_4.pt)
8d3d55f verified
Raw History Blame Contribute Delete
2.63 kB
{
"raw": {
"ece": 0.13197172716259956,
"brier": 0.05452140746153333,
"nll": 0.8668859592825174,
"accuracy": 0.776,
"soft_accuracy": 0.5203169700633734,
"mean_confidence": 0.6449173926413059,
"score_mae": 0.22454273352865128,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.11911850492159524,
"brier": 0.0698617118084015,
"nll": 0.9974608553511402,
"accuracy": 0.7383333333333333,
"soft_accuracy": 0.47561304283638794,
"mean_confidence": 0.620018160144488,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.15881713267415762,
"brier": 0.05666545718628185,
"nll": 1.0427551303524525,
"accuracy": 0.74,
"soft_accuracy": 0.4401479551522061,
"mean_confidence": 0.5828031701594591,
"score_mae": 0.22454273352865128,
"n": 800
},
"noul": {
"ece": 0.10903105984131498,
"brier": 0.036322370148333824,
"nll": 0.5018188351206482,
"accuracy": 0.8616666666666667,
"soft_accuracy": 0.6719129171719154,
"mean_confidence": 0.7526355884472529,
"score_mae": null,
"n": 600
}
}
},
"post_temperature": {
"ece": 0.14791404638439418,
"brier": 0.05286476505563003,
"nll": 0.8630265202224254,
"accuracy": 0.776,
"soft_accuracy": 0.5129262935388833,
"mean_confidence": 0.6292819098979234,
"score_mae": 0.22493662622291594,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.13473159489532313,
"brier": 0.06779507493602827,
"nll": 0.9931945211129883,
"accuracy": 0.7383333333333333,
"soft_accuracy": 0.4687981816629569,
"mean_confidence": 0.6054597360392412,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.17830051325261592,
"brier": 0.05443568623688862,
"nll": 1.037335314094089,
"accuracy": 0.74,
"soft_accuracy": 0.43145765563938765,
"mean_confidence": 0.5632958786934614,
"score_mae": 0.22493662622291594,
"n": 800
},
"noul": {
"ece": 0.12058120141426723,
"brier": 0.03583989360022036,
"nll": 0.5004467941696445,
"accuracy": 0.8616666666666667,
"soft_accuracy": 0.6656792559474707,
"mean_confidence": 0.7410854586958885,
"score_mae": null,
"n": 600
}
}
},
"fitted_temperature": {
"type0_bucket1": 1.0623664855957031,
"type2_bucket0": 1.0938509702682495,
"type1_bucket1": 1.0897818803787231
}
}