QA-DeBERTa-v3-large-6970 / all_results.json
saiteki-kai's picture
End of training
4fd5057 verified
{
"epoch": 5.0,
"eval_accuracy": 0.6751505472934757,
"eval_animal_abuse/accuracy": 0.994776591143494,
"eval_animal_abuse/f1": 0.7687776141384389,
"eval_animal_abuse/fpr": 0.002490492377074004,
"eval_animal_abuse/precision": 0.7791044776119403,
"eval_animal_abuse/recall": 0.7587209302325582,
"eval_animal_abuse/threshold": 0.5114725828170776,
"eval_child_abuse/accuracy": 0.9965399075090661,
"eval_child_abuse/f1": 0.6876876876876877,
"eval_child_abuse/fpr": 0.0017396831769291216,
"eval_child_abuse/precision": 0.6876876876876877,
"eval_child_abuse/recall": 0.6876876876876877,
"eval_child_abuse/threshold": 0.4428774416446686,
"eval_controversial_topics,politics/accuracy": 0.966247463153342,
"eval_controversial_topics,politics/f1": 0.519990537023894,
"eval_controversial_topics,politics/fpr": 0.02206891817682588,
"eval_controversial_topics,politics/precision": 0.4607966457023061,
"eval_controversial_topics,politics/recall": 0.5966340933767644,
"eval_controversial_topics,politics/threshold": 0.2509128153324127,
"eval_discrimination,stereotype,injustice/accuracy": 0.9540872342549157,
"eval_discrimination,stereotype,injustice/f1": 0.7218302761540012,
"eval_discrimination,stereotype,injustice/fpr": 0.02821050348790973,
"eval_discrimination,stereotype,injustice/precision": 0.6964216258265267,
"eval_discrimination,stereotype,injustice/recall": 0.749163179916318,
"eval_discrimination,stereotype,injustice/threshold": 0.3674972355365753,
"eval_drug_abuse,weapons,banned_substance/accuracy": 0.9738164154772598,
"eval_drug_abuse,weapons,banned_substance/f1": 0.7758473369410425,
"eval_drug_abuse,weapons,banned_substance/fpr": 0.016076716965167083,
"eval_drug_abuse,weapons,banned_substance/precision": 0.7491749174917491,
"eval_drug_abuse,weapons,banned_substance/recall": 0.8044890726520969,
"eval_drug_abuse,weapons,banned_substance/threshold": 0.4573790431022644,
"eval_financial_crime,property_crime,theft/accuracy": 0.9607911634561,
"eval_financial_crime,property_crime,theft/f1": 0.804609135372627,
"eval_financial_crime,property_crime,theft/fpr": 0.025044689751764507,
"eval_financial_crime,property_crime,theft/precision": 0.7812298776561494,
"eval_financial_crime,property_crime,theft/recall": 0.8294308665185438,
"eval_financial_crime,property_crime,theft/threshold": 0.5081169605255127,
"eval_flagged/accuracy": 0.8544598596000932,
"eval_flagged/aucpr": 0.9034121039023629,
"eval_flagged/f1": 0.8713098477605354,
"eval_flagged/fpr": 0.1843753516108458,
"eval_flagged/precision": 0.8576475357618579,
"eval_flagged/recall": 0.8854144868613794,
"eval_hate_speech,offensive_language/accuracy": 0.949312972019829,
"eval_hate_speech,offensive_language/f1": 0.7024704618689581,
"eval_hate_speech,offensive_language/fpr": 0.023022108532797327,
"eval_hate_speech,offensive_language/precision": 0.7405806053119209,
"eval_hate_speech,offensive_language/recall": 0.6680906389301634,
"eval_hate_speech,offensive_language/threshold": 0.4561674892902374,
"eval_loss": 0.08080274611711502,
"eval_macro_f1": 0.6789235965374625,
"eval_macro_precision": 0.6705073304098963,
"eval_macro_recall": 0.6955187093258729,
"eval_micro_f1": 0.7535257504530294,
"eval_micro_precision": 0.7421112528882298,
"eval_micro_recall": 0.7652968686096342,
"eval_misinformation_regarding_ethics,laws_and_safety/accuracy": 0.9773929533885617,
"eval_misinformation_regarding_ethics,laws_and_safety/f1": 0.2577826324412889,
"eval_misinformation_regarding_ethics,laws_and_safety/fpr": 0.0145496185777074,
"eval_misinformation_regarding_ethics,laws_and_safety/precision": 0.21454545454545454,
"eval_misinformation_regarding_ethics,laws_and_safety/recall": 0.3228454172366621,
"eval_misinformation_regarding_ethics,laws_and_safety/threshold": 0.08269734680652618,
"eval_non_violent_unethical_behavior/accuracy": 0.8852846258774995,
"eval_non_violent_unethical_behavior/f1": 0.7023223689890357,
"eval_non_violent_unethical_behavior/fpr": 0.06404666998837388,
"eval_non_violent_unethical_behavior/precision": 0.7250445632798574,
"eval_non_violent_unethical_behavior/recall": 0.6809810815335677,
"eval_non_violent_unethical_behavior/threshold": 0.4007493853569031,
"eval_privacy_violation/accuracy": 0.9810193964800213,
"eval_privacy_violation/f1": 0.8114984305303156,
"eval_privacy_violation/fpr": 0.01104150626443618,
"eval_privacy_violation/precision": 0.7955944282474895,
"eval_privacy_violation/recall": 0.8280512474713418,
"eval_privacy_violation/threshold": 0.42823341488838196,
"eval_runtime": 49.1327,
"eval_samples_per_second": 1223.502,
"eval_self_harm/accuracy": 0.9969890541304854,
"eval_self_harm/f1": 0.7475592747559274,
"eval_self_harm/fpr": 0.0006532225646522834,
"eval_self_harm/precision": 0.8729641693811075,
"eval_self_harm/recall": 0.6536585365853659,
"eval_self_harm/threshold": 0.7592254281044006,
"eval_sexually_explicit,adult_content/accuracy": 0.9836976411484846,
"eval_sexually_explicit,adult_content/f1": 0.678477690288714,
"eval_sexually_explicit,adult_content/fpr": 0.009664717814103312,
"eval_sexually_explicit,adult_content/precision": 0.6458463460337289,
"eval_sexually_explicit,adult_content/recall": 0.7145818935729095,
"eval_sexually_explicit,adult_content/threshold": 0.4172714352607727,
"eval_steps_per_second": 19.132,
"eval_terrorism,organized_crime/accuracy": 0.9897694380676714,
"eval_terrorism,organized_crime/f1": 0.46937014667817084,
"eval_terrorism,organized_crime/fpr": 0.006808310834604989,
"eval_terrorism,organized_crime/precision": 0.40117994100294985,
"eval_terrorism,organized_crime/recall": 0.5654885654885655,
"eval_terrorism,organized_crime/threshold": 0.3747907280921936,
"eval_violence,aiding_and_abetting,incitement/accuracy": 0.9219150281132514,
"eval_violence,aiding_and_abetting,incitement/f1": 0.8567067586543745,
"eval_violence,aiding_and_abetting,incitement/fpr": 0.061964552830787216,
"eval_violence,aiding_and_abetting,incitement/precision": 0.8369318859596803,
"eval_violence,aiding_and_abetting,incitement/recall": 0.8774387193596799,
"eval_violence,aiding_and_abetting,incitement/threshold": 0.5012283325195312,
"test_accuracy": 0.6573990897113426,
"test_animal_abuse/accuracy": 0.9939064558629777,
"test_animal_abuse/f1": 0.7724986025712689,
"test_animal_abuse/fpr": 0.0033074402233280687,
"test_animal_abuse/precision": 0.7601760176017601,
"test_animal_abuse/recall": 0.7852272727272728,
"test_animal_abuse/threshold": 0.5018692016601562,
"test_child_abuse/accuracy": 0.9961672056533717,
"test_child_abuse/f1": 0.621301775147929,
"test_child_abuse/fpr": 0.0017158338350391304,
"test_child_abuse/precision": 0.6481481481481481,
"test_child_abuse/recall": 0.5965909090909091,
"test_child_abuse/threshold": 0.42155182361602783,
"test_controversial_topics,politics/accuracy": 0.9626152832674572,
"test_controversial_topics,politics/f1": 0.4936118434394646,
"test_controversial_topics,politics/fpr": 0.02702452568255433,
"test_controversial_topics,politics/precision": 0.4099023240148198,
"test_controversial_topics,politics/recall": 0.6202854230377166,
"test_controversial_topics,politics/threshold": 0.2030746191740036,
"test_discrimination,stereotype,injustice/accuracy": 0.9541711582225416,
"test_discrimination,stereotype,injustice/f1": 0.726964588350727,
"test_discrimination,stereotype,injustice/fpr": 0.025992685475444054,
"test_discrimination,stereotype,injustice/precision": 0.7190753485089112,
"test_discrimination,stereotype,injustice/recall": 0.73502886002886,
"test_discrimination,stereotype,injustice/threshold": 0.3965354859828949,
"test_drug_abuse,weapons,banned_substance/accuracy": 0.9732602706911008,
"test_drug_abuse,weapons,banned_substance/f1": 0.7687726566545832,
"test_drug_abuse,weapons,banned_substance/fpr": 0.016628094981453862,
"test_drug_abuse,weapons,banned_substance/precision": 0.7389248382279742,
"test_drug_abuse,weapons,banned_substance/recall": 0.8011332973556395,
"test_drug_abuse,weapons,banned_substance/threshold": 0.4615021347999573,
"test_financial_crime,property_crime,theft/accuracy": 0.9591717570966583,
"test_financial_crime,property_crime,theft/f1": 0.8037141006262146,
"test_financial_crime,property_crime,theft/fpr": 0.025494901019796,
"test_financial_crime,property_crime,theft/precision": 0.7849008857022354,
"test_financial_crime,property_crime,theft/recall": 0.8234513274336284,
"test_financial_crime,property_crime,theft/threshold": 0.4931950271129608,
"test_flagged/accuracy": 0.8443226733740568,
"test_flagged/aucpr": 0.8966769811979269,
"test_flagged/f1": 0.8643762717170136,
"test_flagged/fpr": 0.20925409668865097,
"test_flagged/precision": 0.8433443624331891,
"test_flagged/recall": 0.8864840280378833,
"test_hate_speech,offensive_language/accuracy": 0.946625344352617,
"test_hate_speech,offensive_language/f1": 0.6963631717911591,
"test_hate_speech,offensive_language/fpr": 0.02805443250172562,
"test_hate_speech,offensive_language/precision": 0.7054357204486627,
"test_hate_speech,offensive_language/recall": 0.6875210225361588,
"test_hate_speech,offensive_language/threshold": 0.39420098066329956,
"test_loss": 0.08579045534133911,
"test_macro_f1": 0.6613452546499687,
"test_macro_precision": 0.6490871217020547,
"test_macro_recall": 0.6834006165061102,
"test_micro_f1": 0.7452154487459984,
"test_micro_precision": 0.7272441228030746,
"test_micro_recall": 0.7640974780391045,
"test_misinformation_regarding_ethics,laws_and_safety/accuracy": 0.9780961791831357,
"test_misinformation_regarding_ethics,laws_and_safety/f1": 0.21554959785522787,
"test_misinformation_regarding_ethics,laws_and_safety/fpr": 0.012853158724384604,
"test_misinformation_regarding_ethics,laws_and_safety/precision": 0.19161105815061963,
"test_misinformation_regarding_ethics,laws_and_safety/recall": 0.24632352941176472,
"test_misinformation_regarding_ethics,laws_and_safety/threshold": 0.09138210862874985,
"test_non_violent_unethical_behavior/accuracy": 0.874535872559588,
"test_non_violent_unethical_behavior/f1": 0.6860717764291601,
"test_non_violent_unethical_behavior/fpr": 0.0764802939963249,
"test_non_violent_unethical_behavior/precision": 0.6918253248715625,
"test_non_violent_unethical_behavior/recall": 0.680413137167484,
"test_non_violent_unethical_behavior/threshold": 0.3647775650024414,
"test_privacy_violation/accuracy": 0.9804467600910288,
"test_privacy_violation/f1": 0.8171892497200448,
"test_privacy_violation/fpr": 0.01167345907180992,
"test_privacy_violation/precision": 0.7979770366320393,
"test_privacy_violation/recall": 0.8373493975903614,
"test_privacy_violation/threshold": 0.42298123240470886,
"test_runtime": 54.2923,
"test_samples_per_second": 1230.229,
"test_self_harm/accuracy": 0.9968259671816984,
"test_self_harm/f1": 0.7389162561576355,
"test_self_harm/fpr": 0.0007236762754794344,
"test_self_harm/precision": 0.8620689655172413,
"test_self_harm/recall": 0.646551724137931,
"test_self_harm/threshold": 0.6610482335090637,
"test_sexually_explicit,adult_content/accuracy": 0.9837106240268295,
"test_sexually_explicit,adult_content/f1": 0.6604244694132334,
"test_sexually_explicit,adult_content/fpr": 0.010166896340529766,
"test_sexually_explicit,adult_content/precision": 0.6144018583042973,
"test_sexually_explicit,adult_content/recall": 0.7139001349527665,
"test_sexually_explicit,adult_content/threshold": 0.42727720737457275,
"test_steps_per_second": 19.229,
"test_terrorism,organized_crime/accuracy": 0.9874086716972093,
"test_terrorism,organized_crime/f1": 0.40816326530612246,
"test_terrorism,organized_crime/fpr": 0.00868107977534874,
"test_terrorism,organized_crime/precision": 0.3352601156069364,
"test_terrorism,organized_crime/recall": 0.5215827338129496,
"test_terrorism,organized_crime/threshold": 0.332852303981781,
"test_violence,aiding_and_abetting,incitement/accuracy": 0.9161576236675051,
"test_violence,aiding_and_abetting,incitement/f1": 0.8492922116367942,
"test_violence,aiding_and_abetting,incitement/fpr": 0.06753316085581687,
"test_violence,aiding_and_abetting,incitement/precision": 0.8275120620935599,
"test_violence,aiding_and_abetting,incitement/recall": 0.8722498618021006,
"test_violence,aiding_and_abetting,incitement/threshold": 0.48133188486099243,
"total_flos": 5.970539301846385e+17,
"train_loss": 0.08776032416230509,
"train_runtime": 27301.134,
"train_samples_per_second": 198.168,
"train_steps_per_second": 12.386
}