aadd / eval /eval_v3_1.json
agertel's picture
v3.1: fine-tuned with FSD50K noise and babble (fewer music false alarms); eval reports
f8aad1c verified
Raw History Blame Contribute Delete
38.8 kB
{
"ckpt": "/home/agertel/work/aadd-runs/v3/v3_ft_noise.pt",
"synthetic": {
"test_en": {
"frame": {
"per_class": {
"noise": {
"support": 78639,
"auc": 0.9748964886451037,
"ap": 0.8976065214379684,
"f1": 0.8599536530978095,
"precision": 0.845073534827764,
"recall": 0.8753671842215694
},
"music": {
"support": 84000,
"auc": 0.9982148080235507,
"ap": 0.9845543440714614,
"f1": 0.9639112208908787,
"precision": 0.9615247932905304,
"recall": 0.9663095238095238
},
"reverb": {
"support": 78236,
"auc": 0.9963041275923323,
"ap": 0.9852604445664134,
"f1": 0.9391880634591534,
"precision": 0.9078989751971464,
"recall": 0.9727107725343831
},
"clipping": {
"support": 53342,
"auc": 0.9994286471649927,
"ap": 0.9927754549430884,
"f1": 0.9705925953378053,
"precision": 0.9594980765708006,
"recall": 0.9819466836639046
},
"lowpass": {
"support": 103755,
"auc": 0.9995742578524216,
"ap": 0.9960508073167798,
"f1": 0.9422761788261361,
"precision": 0.904190290574061,
"recall": 0.9837116283552599
},
"highpass": {
"support": 79278,
"auc": 0.9979229022050684,
"ap": 0.9907064618975518,
"f1": 0.9080643851036916,
"precision": 0.840823124865678,
"recall": 0.9869951310577966
},
"codec": {
"support": 83463,
"auc": 0.9896105009604881,
"ap": 0.9247977729877249,
"f1": 0.8202199283048466,
"precision": 0.7314849588763286,
"recall": 0.9334555431748199
},
"packet_loss": {
"support": 8226,
"auc": 0.99943930480011,
"ap": 0.9912479097263671,
"f1": 0.9640305050141115,
"precision": 0.9525335231992406,
"recall": 0.975808412351082
},
"hum": {
"support": 86573,
"auc": 0.9745201488600769,
"ap": 0.7258485163859276,
"f1": 0.7263919610060753,
"precision": 0.6408307064713518,
"recall": 0.8383214166079493
},
"clicks": {
"support": 8410,
"auc": 0.9981566646263414,
"ap": 0.9760219150839053,
"f1": 0.9428954097163035,
"precision": 0.9660678642714571,
"recall": 0.920808561236623
},
"pitch_shift": {
"support": 63368,
"auc": 0.9999088463472918,
"ap": 0.9988250912073104,
"f1": 0.9864443105997199,
"precision": 0.9779759596742924,
"recall": 0.9950605984092917
},
"time_stretch": {
"support": 69035,
"auc": 0.9999041807420969,
"ap": 0.9986634889395487,
"f1": 0.9857753322016038,
"precision": 0.9960224164189918,
"recall": 0.9757369450278844
}
},
"macro_auc": 0.9939900731516563,
"macro_ap": 0.9551965607136705,
"macro_f1": 0.9174786286298446
},
"segment_1s": {
"per_class": {
"noise": {
"support": 1795,
"auc": 0.9730336024346444,
"ap": 0.8981832534267176,
"f1": 0.8553020134228188,
"precision": 0.8253886010362694,
"recall": 0.8874651810584958
},
"music": {
"support": 1904,
"auc": 0.9976960852452855,
"ap": 0.9817721733199246,
"f1": 0.9557291666666667,
"precision": 0.9478305785123967,
"recall": 0.9637605042016807
},
"reverb": {
"support": 1783,
"auc": 0.9945803654773061,
"ap": 0.9822968762408003,
"f1": 0.9184161573212862,
"precision": 0.8727272727272727,
"recall": 0.9691531127313516
},
"clipping": {
"support": 1701,
"auc": 0.9988264900750315,
"ap": 0.992502790686163,
"f1": 0.9614721104082805,
"precision": 0.9409116488463702,
"recall": 0.9829512051734274
},
"lowpass": {
"support": 2347,
"auc": 0.9994477326667189,
"ap": 0.9954529399371632,
"f1": 0.9226793877956668,
"precision": 0.8647540983606558,
"recall": 0.9889220281210055
},
"highpass": {
"support": 1780,
"auc": 0.9985627509380153,
"ap": 0.9912740635971926,
"f1": 0.8913592723597775,
"precision": 0.8099173553719008,
"recall": 0.9910112359550561
},
"codec": {
"support": 1860,
"auc": 0.9868970311118882,
"ap": 0.9234671473758682,
"f1": 0.7992625028808481,
"precision": 0.6994755949979831,
"recall": 0.932258064516129
},
"packet_loss": {
"support": 851,
"auc": 0.9965731259613282,
"ap": 0.9826343446078135,
"f1": 0.9327251995438997,
"precision": 0.9058693244739756,
"recall": 0.9612220916568742
},
"hum": {
"support": 1900,
"auc": 0.9738075970469158,
"ap": 0.7529680753561515,
"f1": 0.7298772169167804,
"precision": 0.6425140112089672,
"recall": 0.8447368421052631
},
"clicks": {
"support": 1009,
"auc": 0.9922892100715139,
"ap": 0.9726803139403241,
"f1": 0.9336683417085426,
"precision": 0.9469928644240571,
"recall": 0.9207135777998018
},
"pitch_shift": {
"support": 1568,
"auc": 0.9998543074170645,
"ap": 0.9979678400602622,
"f1": 0.9780426599749059,
"precision": 0.9623456790123457,
"recall": 0.9942602040816326
},
"time_stretch": {
"support": 1687,
"auc": 0.9997216276666899,
"ap": 0.9965234658250193,
"f1": 0.9815585960737656,
"precision": 0.9850746268656716,
"recall": 0.978067575577949
}
},
"macro_auc": 0.9926074938427001,
"macro_ap": 0.9556436070311167,
"macro_f1": 0.9050077187561033
},
"clip": {
"per_class": {
"noise": {
"support": 367,
"auc": 0.963324436956632,
"ap": 0.905089091203634,
"f1": 0.8688741721854305,
"precision": 0.845360824742268,
"recall": 0.8937329700272479
},
"music": {
"support": 377,
"auc": 0.9954346926950027,
"ap": 0.9783943883815625,
"f1": 0.9537648612945839,
"precision": 0.95,
"recall": 0.9575596816976127
},
"reverb": {
"support": 358,
"auc": 0.9892476074076266,
"ap": 0.9762289042439476,
"f1": 0.9181208053691275,
"precision": 0.8837209302325582,
"recall": 0.9553072625698324
},
"clipping": {
"support": 349,
"auc": 0.9957944182819047,
"ap": 0.9898288664613474,
"f1": 0.9619181946403386,
"precision": 0.9472222222222222,
"recall": 0.9770773638968482
},
"lowpass": {
"support": 460,
"auc": 0.9994355528928449,
"ap": 0.9972786876594931,
"f1": 0.9385245901639344,
"precision": 0.8875968992248062,
"recall": 0.9956521739130435
},
"highpass": {
"support": 345,
"auc": 0.9989000791506317,
"ap": 0.996165397009855,
"f1": 0.9014454664914586,
"precision": 0.8245192307692307,
"recall": 0.9942028985507246
},
"codec": {
"support": 358,
"auc": 0.977194777953049,
"ap": 0.9249253511399831,
"f1": 0.8014616321559074,
"precision": 0.7105831533477321,
"recall": 0.9189944134078212
},
"packet_loss": {
"support": 328,
"auc": 0.9939960018986417,
"ap": 0.9853876106835922,
"f1": 0.9404761904761905,
"precision": 0.9186046511627907,
"recall": 0.9634146341463414
},
"hum": {
"support": 348,
"auc": 0.9680093965083822,
"ap": 0.8051283442591923,
"f1": 0.7779237844940867,
"precision": 0.7167070217917676,
"recall": 0.8505747126436781
},
"clicks": {
"support": 346,
"auc": 0.9860838259187789,
"ap": 0.9767021233784513,
"f1": 0.9588235294117647,
"precision": 0.9760479041916168,
"recall": 0.9421965317919075
},
"pitch_shift": {
"support": 317,
"auc": 0.9998406840123172,
"ap": 0.998785812289383,
"f1": 0.9767441860465116,
"precision": 0.9603658536585366,
"recall": 0.9936908517350158
},
"time_stretch": {
"support": 334,
"auc": 0.9996209756031823,
"ap": 0.9978055080559207,
"f1": 0.9805097451274363,
"precision": 0.9819819819819819,
"recall": 0.9790419161676647
}
},
"macro_auc": 0.9889068707732496,
"macro_ap": 0.9609766737305301,
"macro_f1": 0.914882263154731,
"false_alarm_rate_on_clean_clips": 0.21452702702702703,
"n_clean_clips": 592
},
"n_windows": 3000
},
"test_en_mit_rir": {
"frame": {
"per_class": {
"noise": {
"support": 27091,
"auc": 0.9802963515360084,
"ap": 0.8938768856333474,
"f1": 0.8503153366628997,
"precision": 0.8399049368045803,
"recall": 0.8609870436676387
},
"music": {
"support": 23749,
"auc": 0.9992527100189356,
"ap": 0.9897075342751966,
"f1": 0.9653627577052308,
"precision": 0.9512364602493358,
"recall": 0.9799149437871069
},
"reverb": {
"support": 25239,
"auc": 0.9593939537065209,
"ap": 0.8597329722803817,
"f1": 0.8265924965494166,
"precision": 0.8752823419992029,
"recall": 0.7830341931138318
},
"clipping": {
"support": 17596,
"auc": 0.9991981410087114,
"ap": 0.9921262846750708,
"f1": 0.9711072217599632,
"precision": 0.9806455351451585,
"recall": 0.9617526710616049
},
"lowpass": {
"support": 29902,
"auc": 0.9985748331305384,
"ap": 0.9858976652516711,
"f1": 0.872889906184589,
"precision": 0.7939460347897548,
"recall": 0.9692662698147281
},
"highpass": {
"support": 28059,
"auc": 0.9994425358495167,
"ap": 0.9933166588239004,
"f1": 0.8774136738159632,
"precision": 0.7850972226130511,
"recall": 0.9943333689725222
},
"codec": {
"support": 28632,
"auc": 0.9873384526802057,
"ap": 0.8882356438371874,
"f1": 0.7901135149365833,
"precision": 0.7097743524108956,
"recall": 0.890961162335848
},
"packet_loss": {
"support": 2372,
"auc": 0.9992059133703438,
"ap": 0.9940390877209412,
"f1": 0.9731768650461023,
"precision": 0.9675,
"recall": 0.978920741989882
},
"hum": {
"support": 25147,
"auc": 0.982139816331797,
"ap": 0.7872021323215781,
"f1": 0.7180796901864557,
"precision": 0.6187859608745685,
"recall": 0.855330655744224
},
"clicks": {
"support": 2915,
"auc": 0.9986746897615487,
"ap": 0.9659265695946404,
"f1": 0.9244819107832807,
"precision": 0.947103274559194,
"recall": 0.9029159519725557
},
"pitch_shift": {
"support": 24972,
"auc": 0.999934540532545,
"ap": 0.9991805599564639,
"f1": 0.9874772773177136,
"precision": 0.9962099600619447,
"recall": 0.9788963639275989
},
"time_stretch": {
"support": 25199,
"auc": 0.9999569313284481,
"ap": 0.9994726898157393,
"f1": 0.9888163474624747,
"precision": 0.9829803921568627,
"recall": 0.9947220127782849
}
},
"macro_auc": 0.9919507391045933,
"macro_ap": 0.9457262236821767,
"macro_f1": 0.8954855832008893
},
"segment_1s": {
"per_class": {
"noise": {
"support": 613,
"auc": 0.9784062473844641,
"ap": 0.8934139346690724,
"f1": 0.8408551068883612,
"precision": 0.816923076923077,
"recall": 0.866231647634584
},
"music": {
"support": 546,
"auc": 0.9987599155160195,
"ap": 0.9875060178051387,
"f1": 0.9490616621983915,
"precision": 0.9267015706806283,
"recall": 0.9725274725274725
},
"reverb": {
"support": 578,
"auc": 0.9615999474115577,
"ap": 0.8707462377835973,
"f1": 0.8158590308370044,
"precision": 0.8312387791741472,
"recall": 0.801038062283737
},
"clipping": {
"support": 556,
"auc": 0.9983278340340507,
"ap": 0.990963045006726,
"f1": 0.957847533632287,
"precision": 0.9552772808586762,
"recall": 0.960431654676259
},
"lowpass": {
"support": 686,
"auc": 0.9979409670817818,
"ap": 0.9839239320015734,
"f1": 0.8538608806636887,
"precision": 0.7593643586833144,
"recall": 0.9752186588921283
},
"highpass": {
"support": 619,
"auc": 0.9994763439902348,
"ap": 0.9940875947218047,
"f1": 0.8477366255144033,
"precision": 0.7365911799761621,
"recall": 0.9983844911147012
},
"codec": {
"support": 640,
"auc": 0.9861543605638587,
"ap": 0.8974806080360286,
"f1": 0.7851351351351351,
"precision": 0.6916666666666667,
"recall": 0.9078125
},
"packet_loss": {
"support": 264,
"auc": 0.999572478604319,
"ap": 0.9881067691214678,
"f1": 0.931159420289855,
"precision": 0.8923611111111112,
"recall": 0.9734848484848485
},
"hum": {
"support": 556,
"auc": 0.9818620520026752,
"ap": 0.8069172696536959,
"f1": 0.7208427389014297,
"precision": 0.6196636481241915,
"recall": 0.8615107913669064
},
"clicks": {
"support": 356,
"auc": 0.9943005945159834,
"ap": 0.9734882903798431,
"f1": 0.9293785310734464,
"precision": 0.9346590909090909,
"recall": 0.9241573033707865
},
"pitch_shift": {
"support": 597,
"auc": 0.9998884512164135,
"ap": 0.998795757080104,
"f1": 0.985797827903091,
"precision": 0.9833333333333333,
"recall": 0.9882747068676717
},
"time_stretch": {
"support": 603,
"auc": 0.9998410453253986,
"ap": 0.9979569198029565,
"f1": 0.9811937857726901,
"precision": 0.967741935483871,
"recall": 0.9950248756218906
}
},
"macro_auc": 0.9913441864705632,
"macro_ap": 0.9486155313385006,
"macro_f1": 0.8832273565674821
},
"clip": {
"per_class": {
"noise": {
"support": 124,
"auc": 0.9736660406539992,
"ap": 0.9073127734083682,
"f1": 0.8412698412698412,
"precision": 0.828125,
"recall": 0.8548387096774194
},
"music": {
"support": 114,
"auc": 0.9980297809987724,
"ap": 0.9889043716696614,
"f1": 0.9519650655021834,
"precision": 0.9478260869565217,
"recall": 0.956140350877193
},
"reverb": {
"support": 121,
"auc": 0.9664203311426396,
"ap": 0.9082585439194265,
"f1": 0.8333333333333334,
"precision": 0.8403361344537815,
"recall": 0.8264462809917356
},
"clipping": {
"support": 112,
"auc": 0.9982655646718147,
"ap": 0.9921330278351584,
"f1": 0.9596412556053813,
"precision": 0.963963963963964,
"recall": 0.9553571428571429
},
"lowpass": {
"support": 145,
"auc": 0.9961322847348255,
"ap": 0.9838920302386135,
"f1": 0.8909657320872274,
"precision": 0.8125,
"recall": 0.9862068965517241
},
"highpass": {
"support": 118,
"auc": 0.9992505476766979,
"ap": 0.9948923241330445,
"f1": 0.8309859154929577,
"precision": 0.7108433734939759,
"recall": 1.0
},
"codec": {
"support": 123,
"auc": 0.9783769502461275,
"ap": 0.9169064654684342,
"f1": 0.7943262411347518,
"precision": 0.7044025157232704,
"recall": 0.9105691056910569
},
"packet_loss": {
"support": 100,
"auc": 0.999638888888889,
"ap": 0.9971642966473001,
"f1": 0.9658536585365853,
"precision": 0.9428571428571428,
"recall": 0.99
},
"hum": {
"support": 106,
"auc": 0.9780349077708834,
"ap": 0.8407594921956165,
"f1": 0.7731092436974791,
"precision": 0.696969696969697,
"recall": 0.8679245283018868
},
"clicks": {
"support": 122,
"auc": 0.990781022442959,
"ap": 0.9772391332963529,
"f1": 0.946938775510204,
"precision": 0.943089430894309,
"recall": 0.9508196721311475
},
"pitch_shift": {
"support": 117,
"auc": 0.9990610873963082,
"ap": 0.9957238130986623,
"f1": 0.9871244635193134,
"precision": 0.9913793103448276,
"recall": 0.9829059829059829
},
"time_stretch": {
"support": 121,
"auc": 0.9997273385421073,
"ap": 0.9983568109499554,
"f1": 0.9876543209876544,
"precision": 0.9836065573770492,
"recall": 0.9917355371900827
}
},
"macro_auc": 0.9897820620971686,
"macro_ap": 0.9584619235717161,
"macro_f1": 0.8969306538897427,
"false_alarm_rate_on_clean_clips": 0.22522522522522523,
"n_clean_clips": 222
},
"n_windows": 1000
},
"test_en_real_rir": {
"frame": {
"per_class": {
"noise": {
"support": 24330,
"auc": 0.984740497257449,
"ap": 0.8929008700979433,
"f1": 0.8503115138765271,
"precision": 0.8371703975145384,
"recall": 0.8638717632552404
},
"music": {
"support": 22023,
"auc": 0.9991426038903324,
"ap": 0.9816566979828463,
"f1": 0.968618037960617,
"precision": 0.9486308824169605,
"recall": 0.9894655587340507
},
"reverb": {
"support": 27310,
"auc": 0.9772851277991343,
"ap": 0.9234581469117255,
"f1": 0.8575948216298775,
"precision": 0.8517408263507656,
"recall": 0.8635298425485171
},
"clipping": {
"support": 20978,
"auc": 0.9997724109205438,
"ap": 0.9969362469879345,
"f1": 0.9744491326769807,
"precision": 0.9586292777419058,
"recall": 0.9907998855944322
},
"lowpass": {
"support": 40361,
"auc": 0.9992496841041979,
"ap": 0.9935824457854006,
"f1": 0.9257951345612333,
"precision": 0.8664967272255946,
"recall": 0.9938059017368251
},
"highpass": {
"support": 20151,
"auc": 0.9997136767891224,
"ap": 0.9951016649357688,
"f1": 0.8755439512619668,
"precision": 0.779573017164555,
"recall": 0.9984616148081981
},
"codec": {
"support": 25015,
"auc": 0.9899468474486349,
"ap": 0.9281334114543798,
"f1": 0.8392804044008326,
"precision": 0.7842183864133644,
"recall": 0.9026584049570258
},
"packet_loss": {
"support": 3157,
"auc": 0.999274002914915,
"ap": 0.9934424228491409,
"f1": 0.9703343274211269,
"precision": 0.9617299315494711,
"recall": 0.9790940766550522
},
"hum": {
"support": 26535,
"auc": 0.9838169043270708,
"ap": 0.762511297406088,
"f1": 0.7806408544726302,
"precision": 0.700554141081324,
"recall": 0.8814019219898248
},
"clicks": {
"support": 2587,
"auc": 0.9995316932567736,
"ap": 0.987235094964985,
"f1": 0.9619404456714652,
"precision": 0.9818840579710145,
"recall": 0.9427908774642443
},
"pitch_shift": {
"support": 27914,
"auc": 0.9998206265751255,
"ap": 0.9977361777155498,
"f1": 0.9799812710921073,
"precision": 0.9668444723355297,
"recall": 0.9934799742064914
},
"time_stretch": {
"support": 27908,
"auc": 0.9996012374639198,
"ap": 0.99583270369918,
"f1": 0.9789272379564703,
"precision": 0.9954439382153573,
"recall": 0.9629496918446324
}
},
"macro_auc": 0.9943246093956016,
"macro_ap": 0.9540439317325786,
"macro_f1": 0.913618094415153
},
"segment_1s": {
"per_class": {
"noise": {
"support": 569,
"auc": 0.983766054851677,
"ap": 0.8996805780735067,
"f1": 0.847659574468085,
"precision": 0.8217821782178217,
"recall": 0.875219683655536
},
"music": {
"support": 504,
"auc": 0.9990670950432825,
"ap": 0.9835833558830598,
"f1": 0.962172647914646,
"precision": 0.9411764705882353,
"recall": 0.9841269841269841
},
"reverb": {
"support": 611,
"auc": 0.9790112431027767,
"ap": 0.9302036802638258,
"f1": 0.8470031545741326,
"precision": 0.817351598173516,
"recall": 0.8788870703764321
},
"clipping": {
"support": 650,
"auc": 0.9994485294117648,
"ap": 0.9951502754142342,
"f1": 0.9675471698113207,
"precision": 0.9496296296296296,
"recall": 0.9861538461538462
},
"lowpass": {
"support": 880,
"auc": 0.9989525344739529,
"ap": 0.9917536174720621,
"f1": 0.9071094966268811,
"precision": 0.8347659980897804,
"recall": 0.9931818181818182
},
"highpass": {
"support": 459,
"auc": 0.9997693075963238,
"ap": 0.996485412697227,
"f1": 0.8492136910268271,
"precision": 0.7379421221864951,
"recall": 1.0
},
"codec": {
"support": 555,
"auc": 0.984268176840373,
"ap": 0.9236010370312216,
"f1": 0.8222940226171244,
"precision": 0.7452415812591509,
"recall": 0.9171171171171171
},
"packet_loss": {
"support": 312,
"auc": 0.9988723316377549,
"ap": 0.985323517277731,
"f1": 0.9298780487804877,
"precision": 0.8866279069767442,
"recall": 0.9775641025641025
},
"hum": {
"support": 592,
"auc": 0.9822586110340903,
"ap": 0.7883184049671518,
"f1": 0.7895522388059703,
"precision": 0.7072192513368984,
"recall": 0.893581081081081
},
"clicks": {
"support": 292,
"auc": 0.999506383689602,
"ap": 0.9906869414751253,
"f1": 0.9540034071550255,
"precision": 0.9491525423728814,
"recall": 0.958904109589041
},
"pitch_shift": {
"support": 661,
"auc": 0.9997899436393429,
"ap": 0.9978073409661652,
"f1": 0.9754646840148699,
"precision": 0.9590643274853801,
"recall": 0.9924357034795764
},
"time_stretch": {
"support": 679,
"auc": 0.9993830164360641,
"ap": 0.9945733462280906,
"f1": 0.9737827715355806,
"precision": 0.9908536585365854,
"recall": 0.9572901325478645
}
},
"macro_auc": 0.9936744356464171,
"macro_ap": 0.9564306256457834,
"macro_f1": 0.9021400756109125
},
"clip": {
"per_class": {
"noise": {
"support": 124,
"auc": 0.9766349977905435,
"ap": 0.9132270265707487,
"f1": 0.8720000000000001,
"precision": 0.8650793650793651,
"recall": 0.8790322580645161
},
"music": {
"support": 102,
"auc": 0.9976691121883051,
"ap": 0.9833609365332594,
"f1": 0.9468599033816426,
"precision": 0.9333333333333333,
"recall": 0.9607843137254902
},
"reverb": {
"support": 119,
"auc": 0.9834746611470923,
"ap": 0.9557302859978887,
"f1": 0.8744939271255061,
"precision": 0.84375,
"recall": 0.907563025210084
},
"clipping": {
"support": 114,
"auc": 0.9972674349530711,
"ap": 0.988180506110162,
"f1": 0.9523809523809523,
"precision": 0.9401709401709402,
"recall": 0.9649122807017544
},
"lowpass": {
"support": 159,
"auc": 0.9982276265900881,
"ap": 0.9912231788093401,
"f1": 0.9127906976744187,
"precision": 0.8486486486486486,
"recall": 0.9874213836477987
},
"highpass": {
"support": 90,
"auc": 0.99991452991453,
"ap": 0.9991820987654322,
"f1": 0.878048780487805,
"precision": 0.782608695652174,
"recall": 1.0
},
"codec": {
"support": 104,
"auc": 0.9678807520604396,
"ap": 0.9145124324660501,
"f1": 0.810344827586207,
"precision": 0.734375,
"recall": 0.9038461538461539
},
"packet_loss": {
"support": 126,
"auc": 0.9961815771312339,
"ap": 0.9864005360001937,
"f1": 0.9486166007905138,
"precision": 0.9448818897637795,
"recall": 0.9523809523809523
},
"hum": {
"support": 113,
"auc": 0.9722291506619709,
"ap": 0.8227366099405548,
"f1": 0.8132780082987552,
"precision": 0.765625,
"recall": 0.8672566371681416
},
"clicks": {
"support": 101,
"auc": 0.9998237866055795,
"ap": 0.9986460184479988,
"f1": 0.9852216748768472,
"precision": 0.9803921568627451,
"recall": 0.9900990099009901
},
"pitch_shift": {
"support": 127,
"auc": 0.9997203957752703,
"ap": 0.9981901104992842,
"f1": 0.980544747081712,
"precision": 0.9692307692307692,
"recall": 0.9921259842519685
},
"time_stretch": {
"support": 128,
"auc": 0.999359410837156,
"ap": 0.9964951044456638,
"f1": 0.9803921568627452,
"precision": 0.984251968503937,
"recall": 0.9765625
}
},
"macro_auc": 0.9906986196379401,
"macro_ap": 0.9623237370488814,
"macro_f1": 0.9129143563789254,
"false_alarm_rate_on_clean_clips": 0.208955223880597,
"n_clean_clips": 201
},
"n_windows": 1000
},
"test_ru": {
"frame": {
"per_class": {
"noise": {
"support": 81329,
"auc": 0.9649626195584178,
"ap": 0.7993742227463438,
"f1": 0.5226713508012518,
"precision": 0.364713492901161,
"recall": 0.921983548303803
},
"music": {
"support": 80682,
"auc": 0.9992817747981081,
"ap": 0.9902801450176862,
"f1": 0.9673274504412169,
"precision": 0.975570402117736,
"recall": 0.9592226271039389
},
"reverb": {
"support": 85886,
"auc": 0.9983230184031862,
"ap": 0.9912793848293647,
"f1": 0.9619473116623016,
"precision": 0.9793918918918919,
"recall": 0.9451132897096151
},
"clipping": {
"support": 45997,
"auc": 0.9997990793817971,
"ap": 0.9972477400268378,
"f1": 0.9838880084521924,
"precision": 0.9962778879800299,
"recall": 0.9718025088592734
},
"lowpass": {
"support": 111106,
"auc": 0.9989208378084083,
"ap": 0.9943517663965455,
"f1": 0.9583709657010965,
"precision": 0.9300503874327815,
"recall": 0.9884704696416036
},
"highpass": {
"support": 83549,
"auc": 0.9999715405896906,
"ap": 0.9997739981278555,
"f1": 0.9933327342977805,
"precision": 0.9943036001247212,
"recall": 0.9923637625824366
},
"codec": {
"support": 80596,
"auc": 0.9916246818196484,
"ap": 0.9609520509573792,
"f1": 0.8353005704256253,
"precision": 0.7485549132947977,
"recall": 0.9447863417539332
},
"packet_loss": {
"support": 8257,
"auc": 0.9995611841847145,
"ap": 0.9924402333648419,
"f1": 0.9774665042630938,
"precision": 0.9830944505696435,
"recall": 0.9719026280731501
},
"hum": {
"support": 77260,
"auc": 0.9619477123915232,
"ap": 0.8248669744452303,
"f1": 0.6682432320693247,
"precision": 0.5819850787828742,
"recall": 0.7845198032617137
},
"clicks": {
"support": 8383,
"auc": 0.9983399126828789,
"ap": 0.9785697789475905,
"f1": 0.9542403904820013,
"precision": 0.9766454352441614,
"recall": 0.9328402719790051
},
"pitch_shift": {
"support": 72566,
"auc": 0.9994983871448674,
"ap": 0.992625387015011,
"f1": 0.9677589423310893,
"precision": 0.9496829481839166,
"recall": 0.9865363944547033
},
"time_stretch": {
"support": 70396,
"auc": 0.9993842460862828,
"ap": 0.9940437414086151,
"f1": 0.9727566635534162,
"precision": 0.9840829142076344,
"recall": 0.961688164100233
}
},
"macro_auc": 0.992634582904127,
"macro_ap": 0.9596504519402752,
"macro_f1": 0.896942010373366
},
"segment_1s": {
"per_class": {
"noise": {
"support": 1826,
"auc": 0.9633508986131888,
"ap": 0.8101629152042603,
"f1": 0.5036333975975085,
"precision": 0.34533251982916413,
"recall": 0.9299014238773274
},
"music": {
"support": 1838,
"auc": 0.9990307547381586,
"ap": 0.9887602151539269,
"f1": 0.9571583514099783,
"precision": 0.9540540540540541,
"recall": 0.9602829162132753
},
"reverb": {
"support": 1938,
"auc": 0.9976141447327491,
"ap": 0.989633174102984,
"f1": 0.9540436456996149,
"precision": 0.9494123658661217,
"recall": 0.9587203302373581
},
"clipping": {
"support": 1581,
"auc": 0.9994613578867763,
"ap": 0.9947187897236077,
"f1": 0.9800443458980044,
"precision": 0.9815989847715736,
"recall": 0.978494623655914
},
"lowpass": {
"support": 2495,
"auc": 0.9982670199734507,
"ap": 0.9926824332612316,
"f1": 0.9207180254300674,
"precision": 0.8629512793550649,
"recall": 0.9867735470941884
},
"highpass": {
"support": 1879,
"auc": 0.9999483584395361,
"ap": 0.9994815998347878,
"f1": 0.9850039463299132,
"precision": 0.9739854318418314,
"recall": 0.9962746141564662
},
"codec": {
"support": 1816,
"auc": 0.9883314223371686,
"ap": 0.9559092069477285,
"f1": 0.7982415548357242,
"precision": 0.6883479648842777,
"recall": 0.9498898678414097
},
"packet_loss": {
"support": 872,
"auc": 0.9981373189000685,
"ap": 0.9871081874875492,
"f1": 0.9577142857142859,
"precision": 0.9544419134396356,
"recall": 0.9610091743119266
},
"hum": {
"support": 1735,
"auc": 0.9584374747199537,
"ap": 0.8219958817103862,
"f1": 0.6499408284023669,
"precision": 0.5514056224899598,
"recall": 0.7913544668587896
},
"clicks": {
"support": 982,
"auc": 0.995474767575129,
"ap": 0.972323004457899,
"f1": 0.9371428571428572,
"precision": 0.9565217391304348,
"recall": 0.9185336048879837
},
"pitch_shift": {
"support": 1748,
"auc": 0.9991660956384136,
"ap": 0.9899846998766798,
"f1": 0.9536167863059082,
"precision": 0.9215581643543223,
"recall": 0.9879862700228833
},
"time_stretch": {
"support": 1710,
"auc": 0.9990186069807597,
"ap": 0.9914163032971024,
"f1": 0.971294669009959,
"precision": 0.9730046948356808,
"recall": 0.9695906432748538
}
},
"macro_auc": 0.9913531850446127,
"macro_ap": 0.9578480342548451,
"macro_f1": 0.880712724481349
},
"clip": {
"per_class": {
"noise": {
"support": 358,
"auc": 0.9582744788737161,
"ap": 0.8377555509788619,
"f1": 0.5728987993138936,
"precision": 0.41336633663366334,
"recall": 0.9329608938547486
},
"music": {
"support": 369,
"auc": 0.9983380354518102,
"ap": 0.9886908120162,
"f1": 0.9578231292517005,
"precision": 0.9617486338797814,
"recall": 0.9539295392953929
},
"reverb": {
"support": 383,
"auc": 0.9915824529512297,
"ap": 0.9832630470102248,
"f1": 0.9473684210526316,
"precision": 0.9318181818181818,
"recall": 0.9634464751958225
},
"clipping": {
"support": 331,
"auc": 0.9994357278770803,
"ap": 0.9969921828515053,
"f1": 0.9863429438543247,
"precision": 0.9908536585365854,
"recall": 0.9818731117824774
},
"lowpass": {
"support": 474,
"auc": 0.9949554172471278,
"ap": 0.9878711192763637,
"f1": 0.8964181994191674,
"precision": 0.8282647584973166,
"recall": 0.9767932489451476
},
"highpass": {
"support": 365,
"auc": 0.9999500922772998,
"ap": 0.9996555429211587,
"f1": 0.9655172413793104,
"precision": 0.9357326478149101,
"recall": 0.9972602739726028
},
"codec": {
"support": 351,
"auc": 0.9738196104749522,
"ap": 0.9405614310807068,
"f1": 0.7776456599286564,
"precision": 0.6673469387755102,
"recall": 0.9316239316239316
},
"packet_loss": {
"support": 329,
"auc": 0.996272015421748,
"ap": 0.9912952028382533,
"f1": 0.9696048632218845,
"precision": 0.9696048632218845,
"recall": 0.9696048632218845
},
"hum": {
"support": 350,
"auc": 0.9458975741239892,
"ap": 0.8318459306312649,
"f1": 0.6969696969696969,
"precision": 0.6244343891402715,
"recall": 0.7885714285714286
},
"clicks": {
"support": 333,
"auc": 0.9949989359438178,
"ap": 0.9820060647671576,
"f1": 0.960122699386503,
"precision": 0.9811912225705329,
"recall": 0.93993993993994
},
"pitch_shift": {
"support": 348,
"auc": 0.9987024540143201,
"ap": 0.9888640805854124,
"f1": 0.9465020576131687,
"precision": 0.905511811023622,
"recall": 0.9913793103448276
},
"time_stretch": {
"support": 343,
"auc": 0.9960306182798944,
"ap": 0.9890571102036667,
"f1": 0.9708454810495627,
"precision": 0.9708454810495627,
"recall": 0.9708454810495627
}
},
"macro_auc": 0.9873547844114156,
"macro_ap": 0.959821506263398,
"macro_f1": 0.8873382660367084,
"false_alarm_rate_on_clean_clips": 0.43676222596964587,
"n_clean_clips": 593
},
"n_windows": 3000
}
},
"ood": {
"VoiceBank-DEMAND test (824 noisy vs 824 clean twins)": {
"n": 1648,
"noise": {
"auc_mean": 0.993627167970591,
"auc_p95": 0.9921742447450278,
"recall_at_0.5_p95": 0.5970873786407767,
"false_alarm_at_0.5_p95": 0.0
},
"absent_class_detection_rate": {
"music": 0.13895631067961164,
"packet_loss": 0.0,
"hum": 0.0,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.0
},
"clean_clips_any_class_false_alarm": 0.0,
"clean_clips_false_alarm_by_class": {
"noise": 0.0,
"music": 0.0,
"reverb": 0.0,
"clipping": 0.0,
"lowpass": 0.0,
"highpass": 0.0,
"codec": 0.0,
"packet_loss": 0.0,
"hum": 0.0,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.0
}
},
"DNS-2020 synthetic (clean / noisy, dry / reverberant)": {
"n": 600,
"noise": {
"auc_mean": 0.9988555555555555,
"auc_p95": 0.9978333333333333,
"recall_at_0.5_p95": 0.9133333333333333,
"false_alarm_at_0.5_p95": 0.0
},
"reverb": {
"auc_mean": 1.0,
"auc_p95": 1.0,
"recall_at_0.5_p95": 1.0,
"false_alarm_at_0.5_p95": 0.0
},
"absent_class_detection_rate": {
"music": 0.08166666666666667,
"packet_loss": 0.0,
"hum": 0.016666666666666666,
"clicks": 0.0,
"pitch_shift": 0.013333333333333334,
"time_stretch": 0.005
},
"clean_clips_any_class_false_alarm": 0.006666666666666667,
"clean_clips_false_alarm_by_class": {
"noise": 0.0,
"music": 0.0,
"reverb": 0.0,
"clipping": 0.0,
"lowpass": 0.0,
"highpass": 0.0,
"codec": 0.006666666666666667,
"packet_loss": 0.0,
"hum": 0.0,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.0
}
},
"Dynamic-SUPERB NoiseDetection_VCTK_MUSAN-Noise": {
"n": 200,
"noise": {
"auc_mean": 0.9930436535941123,
"auc_p95": 0.9923379372920658,
"recall_at_0.5_p95": 0.963302752293578,
"false_alarm_at_0.5_p95": 0.02197802197802198
},
"absent_class_detection_rate": {
"music": 0.005,
"packet_loss": 0.005,
"hum": 0.01,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.005
},
"clean_clips_any_class_false_alarm": 0.02197802197802198,
"clean_clips_false_alarm_by_class": {
"noise": 0.02197802197802198,
"music": 0.0,
"reverb": 0.0,
"clipping": 0.0,
"lowpass": 0.0,
"highpass": 0.0,
"codec": 0.0,
"packet_loss": 0.0,
"hum": 0.0,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.0
}
},
"Dynamic-SUPERB ReverberationDetection_VCTK_RirsNoises-SmallRoom": {
"n": 200,
"reverb": {
"auc_mean": 0.9955716586151369,
"auc_p95": 0.99381038647343,
"recall_at_0.5_p95": 0.9782608695652174,
"false_alarm_at_0.5_p95": 0.0
},
"absent_class_detection_rate": {
"music": 0.0,
"packet_loss": 0.0,
"hum": 0.0,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.045
},
"clean_clips_any_class_false_alarm": 0.027777777777777776,
"clean_clips_false_alarm_by_class": {
"noise": 0.027777777777777776,
"music": 0.0,
"reverb": 0.0,
"clipping": 0.0,
"lowpass": 0.0,
"highpass": 0.0,
"codec": 0.0,
"packet_loss": 0.0,
"hum": 0.0,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.0
}
},
"Dynamic-SUPERB ReverberationDetection_VCTK_RirsNoises-MediumRoom": {
"n": 200,
"reverb": {
"auc_mean": 1.0,
"auc_p95": 1.0,
"recall_at_0.5_p95": 1.0,
"false_alarm_at_0.5_p95": 0.0
},
"absent_class_detection_rate": {
"music": 0.015,
"packet_loss": 0.0,
"hum": 0.0,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.055
},
"clean_clips_any_class_false_alarm": 0.037037037037037035,
"clean_clips_false_alarm_by_class": {
"noise": 0.027777777777777776,
"music": 0.0,
"reverb": 0.0,
"clipping": 0.0,
"lowpass": 0.0,
"highpass": 0.0,
"codec": 0.0,
"packet_loss": 0.0,
"hum": 0.0,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.009259259259259259
}
},
"Dynamic-SUPERB ReverberationDetection_VCTK_RirsNoises-LargeRoom": {
"n": 200,
"reverb": {
"auc_mean": 0.9966787439613527,
"auc_p95": 0.9958735909822866,
"recall_at_0.5_p95": 0.967391304347826,
"false_alarm_at_0.5_p95": 0.0
},
"absent_class_detection_rate": {
"music": 0.005,
"packet_loss": 0.0,
"hum": 0.005,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.065
},
"clean_clips_any_class_false_alarm": 0.027777777777777776,
"clean_clips_false_alarm_by_class": {
"noise": 0.027777777777777776,
"music": 0.0,
"reverb": 0.0,
"clipping": 0.0,
"lowpass": 0.0,
"highpass": 0.0,
"codec": 0.0,
"packet_loss": 0.0,
"hum": 0.0,
"clicks": 0.0,
"pitch_shift": 0.0,
"time_stretch": 0.0
}
}
}
}