minhleduc's picture
E2 run e2_laya_rlce_seed42_rep2 (epoch_4.pt)
07642fb verified
Raw History Blame Contribute Delete
2.61 kB
{
"raw": {
"ece": 0.1311594566702843,
"brier": 0.05490199812186093,
"nll": 0.8674021078860388,
"accuracy": 0.772,
"soft_accuracy": 0.5175125464983285,
"mean_confidence": 0.6411024855226278,
"score_mae": 0.23364423435763457,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.13066786532600722,
"brier": 0.06976302182632935,
"nll": 0.9982771100103855,
"accuracy": 0.7383333333333333,
"soft_accuracy": 0.47206922526160877,
"mean_confidence": 0.614280567864577,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.14976227495819328,
"brier": 0.05792711920497595,
"nll": 1.0437982700206339,
"accuracy": 0.72375,
"soft_accuracy": 0.43476231436245144,
"mean_confidence": 0.5739877213910222,
"score_mae": 0.23364423435763457,
"n": 800
},
"noul": {
"ece": 0.11258928567171096,
"brier": 0.03600747963990583,
"nll": 0.501332222915565,
"accuracy": 0.87,
"soft_accuracy": 0.6732895105828841,
"mean_confidence": 0.7574107553561529,
"score_mae": null,
"n": 600
}
}
},
"post_temperature": {
"ece": 0.14642349288612602,
"brier": 0.05350534239243169,
"nll": 0.8643806909918785,
"accuracy": 0.772,
"soft_accuracy": 0.5103804081380368,
"mean_confidence": 0.6261566658765078,
"score_mae": 0.23484325315104798,
"n": 2000,
"by_type": {
"choice": {
"ece": 0.14246670683224996,
"brier": 0.06797903129201587,
"nll": 0.9949558153189719,
"accuracy": 0.7383333333333333,
"soft_accuracy": 0.46516082701583705,
"mean_confidence": 0.5998148304224015,
"score_mae": null,
"n": 600
},
"score": {
"ece": 0.16657996598631142,
"brier": 0.056304361980874094,
"nll": 1.0398800939321518,
"accuracy": 0.72375,
"soft_accuracy": 0.42730253156274556,
"mean_confidence": 0.5571700439229608,
"score_mae": 0.23484325315104798,
"n": 800
},
"noul": {
"ece": 0.12551932652791342,
"brier": 0.0352996273749243,
"nll": 0.49980636274442075,
"accuracy": 0.87,
"soft_accuracy": 0.6663704913606247,
"mean_confidence": 0.7444806639353434,
"score_mae": null,
"n": 600
}
}
},
"fitted_temperature": {
"type0_bucket1": 1.0611073970794678,
"type2_bucket0": 1.1017639636993408,
"type1_bucket1": 1.077757716178894
}
}