minhleduc's picture
E2 run e2_rl_only_seed42 (epoch_4.pt)
b851f29 verified
Raw History Blame Contribute Delete
2.63 kB
{
"raw": {
"ece": 0.12822754203528167,
"brier": 0.05530109313218827,
"nll": 0.8678397416705266,
"accuracy": 0.7615,
"soft_accuracy": 0.5146520517989993,
"mean_confidence": 0.6345642500817775,
"score_mae": 0.23563454432209255,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.1322263596455256,
"brier": 0.06862509907051693,
"nll": 0.9974130208231509,
"accuracy": 0.7316666666666667,
"soft_accuracy": 0.4691468758632739,
"mean_confidence": 0.6079821938276291,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.15854124311357734,
"brier": 0.05788625737608754,
"nll": 1.0440307543426752,
"accuracy": 0.72,
"soft_accuracy": 0.4317976192757487,
"mean_confidence": 0.563935886695981,
"score_mae": 0.23563454432209255,
"n": 800
},
"noul": {
"ece": 0.09546335597832997,
"brier": 0.038530201535327255,
"nll": 0.5033451122883708,
"accuracy": 0.8466666666666667,
"soft_accuracy": 0.6706298044323922,
"mean_confidence": 0.7553174575169881,
"score_mae": null,
"n": 600
}
}
},
"post_temperature": {
"ece": 0.14291595120728018,
"brier": 0.05413706335548018,
"nll": 0.8655329882353544,
"accuracy": 0.7615,
"soft_accuracy": 0.5071216474995017,
"mean_confidence": 0.6191726033687591,
"score_mae": 0.2361672744632233,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.1422972479959329,
"brier": 0.06695349502802855,
"nll": 0.9947054195279876,
"accuracy": 0.7316666666666667,
"soft_accuracy": 0.46050154651204744,
"mean_confidence": 0.5902835857371489,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.16855653237551455,
"brier": 0.056834597829119954,
"nll": 1.0415180657990277,
"accuracy": 0.72,
"soft_accuracy": 0.4264914493076503,
"mean_confidence": 0.5522292291745543,
"score_mae": 0.2361672744632233,
"n": 800
},
"noul": {
"ece": 0.12376613413294155,
"brier": 0.03772391905141211,
"nll": 0.5017137868578235,
"accuracy": 0.8466666666666667,
"soft_accuracy": 0.6612486794094244,
"mean_confidence": 0.7373194532593091,
"score_mae": null,
"n": 600
}
}
},
"fitted_temperature": {
"type0_bucket1": 1.0749820470809937,
"type2_bucket0": 1.1464506387710571,
"type1_bucket1": 1.0550986528396606
}
}