Training Report: unified_test

Basic Information

Model Type: wakeword
Version: 1.0.0
Author: Trixy ML Trainer
Created: 2025-10-01T18:24:41.776611+00:00
Last Modified: 2025-10-01T18:25:22.961899+00:00
Description: RepCNN wakeword detection model (standard)

Model Architecture

Architecture: RepCNN
Total Parameters: 2,586,371
Trainable Parameters: 2,586,371
Model Size: 9.88 MB
Layers: 45
Activations: ReLU

Training Results

Total Epochs: 19
Best Epoch: 19
Final Train Loss: 0.040176
Final Train Accuracy: 0.0000
Best Validation Loss: 0.114964
Best Validation Accuracy: 0.9117
Training Duration: 30.20 seconds

Dataset Information

Total Samples: 1,415
Training Samples: 989
Validation Samples: 283
Test Samples: 143
Number of Classes: 0

Hardware Information

Hostname: BPGamer
Platform: Linux 6.6.87.2-microsoft-standard-WSL2
CPU Cores: 32
CUDA Available: True
CUDA Version: 12.8
GPU Count: 1
GPUs: NVIDIA GeForce RTX 5070 Ti
Total Memory: 31.2 GB

Wakeword Information

Wakewords: custom, system_command, negative
Classes:
- 0: custom
- 1: system_command
- 2: negative
Detection Threshold: 0.5

Test Results

validation_loss: 0.431685
validation_time: 0.409211
num_samples: 283
metrics: {'accuracy': 0.508833922261484, 'precision_Class_0': np.float64(0.06060606060606061), 'recall_Class_0': np.float64(0.027777777777777776), 'f1_Class_0': np.float64(0.0380952380952381), 'support_Class_0': np.int64(72), 'precision_Class_1': np.float64(0.0), 'recall_Class_1': np.float64(0.0), 'f1_Class_1': np.float64(0.0), 'support_Class_1': np.int64(38), 'precision_Class_2': np.float64(0.568), 'recall_Class_2': np.float64(0.8208092485549133), 'f1_Class_2': np.float64(0.6713947990543735), 'support_Class_2': np.int64(173), 'precision_macro': np.float64(0.20953535353535355), 'recall_macro': np.float64(0.28286234211089706), 'f1_macro': np.float64(0.2364966790498705), 'precision_weighted': np.float64(0.3626418246064889), 'recall_weighted': np.float64(0.508833922261484), 'f1_weighted': np.float64(0.42012069745322883), 'confusion_matrix': [[2, 0, 70], [0, 0, 38], [31, 0, 142]], 'auc_roc_ovr': 0.49615625147740744, 'auc_roc_ovo': np.float64(0.5183639393119922), 'classification_report': {'Class_0': {'precision': 0.06060606060606061, 'recall': 0.027777777777777776, 'f1-score': 0.0380952380952381, 'support': 72.0}, 'Class_1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 38.0}, 'Class_2': {'precision': 0.568, 'recall': 0.8208092485549133, 'f1-score': 0.6713947990543735, 'support': 173.0}, 'accuracy': 0.508833922261484, 'macro avg': {'precision': 0.20953535353535355, 'recall': 0.28286234211089706, 'f1-score': 0.2364966790498705, 'support': 283.0}, 'weighted avg': {'precision': 0.3626418246064889, 'recall': 0.508833922261484, 'f1-score': 0.42012069745322883, 'support': 283.0}}, 'intra_class_similarity_mean': np.float32(0.9961346), 'intra_class_similarity_std': np.float32(0.0038530696), 'inter_class_similarity_mean': np.float32(0.9956819), 'inter_class_similarity_std': np.float32(0.004291968), 'class_separation': np.float32(0.00045269728), 'eer': 1.0, 'eer_threshold': 0.0}
predictions: [2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 0, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 0, 0, 2, 2, 0, 2, 0, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 0, 0, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 0, 0, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 2, 0, 0, 2, 0, 2, 2, 2, 0, 2, 2, 2, 2, 2, 0, 2, 0, 2, 2, 2, 0]
targets: [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2]
probabilities: [[0.37385547161102295, 0.2434706836938858, 0.38267385959625244], [0.3723430335521698, 0.24828585982322693, 0.3793710470199585], [0.36844250559806824, 0.24882513284683228, 0.3827323317527771], [0.3736782371997833, 0.2477319836616516, 0.37858980894088745], [0.37188708782196045, 0.2485859990119934, 0.37952694296836853], [0.3677213490009308, 0.2515013813972473, 0.3807772397994995], [0.3669265806674957, 0.2538153827190399, 0.37925806641578674], [0.36977487802505493, 0.24628160893917084, 0.38394349813461304], [0.3653533458709717, 0.2585850656032562, 0.3760615587234497], [0.3694348931312561, 0.24881771206855774, 0.38174742460250854], [0.37473851442337036, 0.2457234263420105, 0.37953805923461914], [0.37297579646110535, 0.244853675365448, 0.38217052817344666], [0.3744390308856964, 0.2479059398174286, 0.377655029296875], [0.36607152223587036, 0.2583729326725006, 0.37555551528930664], [0.3661084473133087, 0.25644418597221375, 0.37744733691215515], [0.3703378140926361, 0.24711045622825623, 0.38255172967910767], [0.36835476756095886, 0.24872726202011108, 0.38291803002357483], [0.3758406639099121, 0.2441190481185913, 0.3800402581691742], [0.36968040466308594, 0.25434374809265137, 0.3759758472442627], [0.37131384015083313, 0.24768632650375366, 0.3809998035430908], [0.3681320548057556, 0.24682055413722992, 0.38504743576049805], [0.3684448003768921, 0.2497076690196991, 0.3818475902080536], [0.3680352568626404, 0.25551339983940125, 0.37645137310028076], [0.36692601442337036, 0.26032838225364685, 0.3727455735206604], [0.3636544644832611, 0.2636449635028839, 0.3727005422115326], [0.3708900213241577, 0.24417737126350403, 0.3849325478076935], [0.3643301725387573, 0.2639106214046478, 0.37175920605659485], [0.3667532503604889, 0.2555050551891327, 0.37774166464805603], [0.37079158425331116, 0.24743646383285522, 0.3817720115184784], [0.3699779212474823, 0.2442917674779892, 0.3857303857803345], [0.36478376388549805, 0.2643769085407257, 0.370839387178421], [0.36891046166419983, 0.24775023758411407, 0.3833392262458801], [0.3637295067310333, 0.26263585686683655, 0.37363460659980774], [0.3765454888343811, 0.24151676893234253, 0.38193774223327637], [0.3740263283252716, 0.24920843541622162, 0.37676525115966797], [0.3677001893520355, 0.25019657611846924, 0.38210320472717285], [0.36789608001708984, 0.24959330260753632, 0.3825106620788574], [0.3674313426017761, 0.25192373991012573, 0.38064488768577576], [0.3704729676246643, 0.24494671821594238, 0.3845802843570709], [0.3690900206565857, 0.2505803406238556, 0.3803296685218811], [0.36457204818725586, 0.26157182455062866, 0.3738560974597931], [0.37045958638191223, 0.24619519710540771, 0.38334524631500244], [0.3670119643211365, 0.2591133117675781, 0.373874694108963], [0.3639020323753357, 0.25878486037254333, 0.3773130774497986], [0.37174931168556213, 0.25087970495224, 0.37737104296684265], [0.3697618842124939, 0.24685002863407135, 0.38338813185691833], [0.3694526255130768, 0.24480117857456207, 0.38574618101119995], [0.36744943261146545, 0.25652241706848145, 0.3760281503200531], [0.3699868619441986, 0.24535736441612244, 0.38465577363967896], [0.37020426988601685, 0.2483595311641693, 0.38143619894981384], [0.37018638849258423, 0.24904565513134003, 0.3807680010795593], [0.3678820729255676, 0.2522294819355011, 0.3798883855342865], [0.36438077688217163, 0.2615729570388794, 0.37404632568359375], [0.3681243360042572, 0.25710785388946533, 0.37476781010627747], [0.3696359694004059, 0.248332217335701, 0.38203176856040955], [0.3677433431148529, 0.2565186619758606, 0.3757379949092865], [0.3664439916610718, 0.25523942708969116, 0.37831661105155945], [0.3719596266746521, 0.25158339738845825, 0.37645697593688965], [0.3719913065433502, 0.2528061270713806, 0.37520259618759155], [0.3648066222667694, 0.2577633857727051, 0.3774299621582031], [0.37598806619644165, 0.24417787790298462, 0.37983405590057373], [0.372807115316391, 0.2462480217218399, 0.3809448778629303], [0.36986154317855835, 0.2481439858675003, 0.38199442625045776], [0.3698616921901703, 0.262130469083786, 0.3680077791213989], [0.37031319737434387, 0.2470361739397049, 0.3826506435871124], [0.3692966103553772, 0.24682410061359406, 0.38387927412986755], [0.38694483041763306, 0.237505242228508, 0.37554994225502014], [0.3703710436820984, 0.2550259530544281, 0.3746029734611511], [0.37901943922042847, 0.23973660171031952, 0.3812440037727356], [0.37068891525268555, 0.24594353139400482, 0.38336750864982605], [0.3752061724662781, 0.24648593366146088, 0.37830793857574463], [0.3774634003639221, 0.23760448396205902, 0.38493213057518005], [0.3678821921348572, 0.2576553523540497, 0.37446239590644836], [0.3625717759132385, 0.26351603865623474, 0.37391215562820435], [0.3651171326637268, 0.25801917910575867, 0.37686365842819214], [0.3703506290912628, 0.245734304189682, 0.3839150667190552], [0.3627982437610626, 0.26262757182121277, 0.374574214220047], [0.3686080574989319, 0.24964556097984314, 0.381746381521225], [0.3626352846622467, 0.26687636971473694, 0.37048837542533875], [0.3614950478076935, 0.2725071907043457, 0.3659977316856384], [0.3709426820278168, 0.24503549933433533, 0.3840218782424927], [0.36885562539100647, 0.2492569088935852, 0.3818874955177307], [0.36589837074279785, 0.25721314549446106, 0.3768884241580963], [0.37265920639038086, 0.24812105298042297, 0.37921977043151855], [0.36882448196411133, 0.24863146245479584, 0.38254404067993164], [0.37696823477745056, 0.244063600897789, 0.37896811962127686], [0.37508249282836914, 0.24541498720645905, 0.3795025646686554], [0.37120819091796875, 0.2449948936700821, 0.38379690051078796], [0.3751823902130127, 0.24589364230632782, 0.3789239525794983], [0.36495667695999146, 0.26063063740730286, 0.3744126558303833], [0.36471810936927795, 0.2595716416835785, 0.37571021914482117], [0.3692455291748047, 0.24813154339790344, 0.38262292742729187], [0.36924538016319275, 0.248735249042511, 0.3820193111896515], [0.36947664618492126, 0.2508240342140198, 0.37969934940338135], [0.36548712849617004, 0.25944191217422485, 0.3750709593296051], [0.36685511469841003, 0.2553226053714752, 0.37782230973243713], [0.36418968439102173, 0.2662654519081116, 0.3695448040962219], [0.37387892603874207, 0.2453160434961319, 0.38080504536628723], [0.36527106165885925, 0.2634718418121338, 0.37125712633132935], [0.3709157109260559, 0.24604518711566925, 0.38303905725479126], [0.36896082758903503, 0.2487162947654724, 0.38232284784317017], [0.37095099687576294, 0.24657703936100006, 0.3824719786643982], [0.36869609355926514, 0.25481411814689636, 0.3764898478984833], [0.36356121301651, 0.266404926776886, 0.37003380060195923], [0.3698275089263916, 0.24790342152118683, 0.382269024848938], [0.37416261434555054, 0.24759821593761444, 0.37823909521102905], [0.3742201030254364, 0.24665704369544983, 0.3791228234767914], [0.3714425265789032, 0.24687381088733673, 0.3816836476325989], [0.3714662492275238, 0.24621929228305817, 0.38231441378593445], [0.36688491702079773, 0.2560265064239502, 0.37708863615989685], [0.376178503036499, 0.24157775938510895, 0.38224366307258606], [0.3758912682533264, 0.24695797264575958, 0.3771507441997528], [0.3807135820388794, 0.24095498025417328, 0.37833142280578613], [0.3760104179382324, 0.24400590360164642, 0.37998366355895996], [0.37579619884490967, 0.24383993446826935, 0.38036391139030457], [0.3678652048110962, 0.2529028356075287, 0.37923189997673035], [0.37658628821372986, 0.2391061633825302, 0.38430753350257874], [0.37216830253601074, 0.2503213882446289, 0.37751030921936035], [0.374326229095459, 0.2437657117843628, 0.3819080889225006], [0.36753547191619873, 0.25308099389076233, 0.3793835937976837], [0.38121137022972107, 0.2449628859758377, 0.3738257586956024], [0.36459779739379883, 0.26131102442741394, 0.37409114837646484], [0.3823380768299103, 0.24038085341453552, 0.3772810399532318], [0.37368589639663696, 0.24661947786808014, 0.3796946406364441], [0.36809754371643066, 0.25821173191070557, 0.373690664768219], [0.37349021434783936, 0.24735553562641144, 0.379154235124588], [0.3778733015060425, 0.24398699402809143, 0.3781397044658661], [0.3785761594772339, 0.24054189026355743, 0.3808819055557251], [0.365445077419281, 0.2571285665035248, 0.3774263262748718], [0.3774394392967224, 0.2404610514640808, 0.38209953904151917], [0.37353426218032837, 0.24874641001224518, 0.37771934270858765], [0.3754233121871948, 0.24265623092651367, 0.3819204568862915], [0.36720532178878784, 0.2518213987350464, 0.38097327947616577], [0.37022870779037476, 0.2507873773574829, 0.3789839744567871], [0.3684055209159851, 0.2507518529891968, 0.3808425962924957], [0.37210020422935486, 0.24614454805850983, 0.3817552626132965], [0.37243762612342834, 0.24750004708766937, 0.3800622820854187], [0.3817303776741028, 0.23856721818447113, 0.37970244884490967], [0.3738105297088623, 0.24597345292568207, 0.3802160620689392], [0.3751688599586487, 0.24425649642944336, 0.38057461380958557], [0.36959272623062134, 0.25120291113853455, 0.37920433282852173], [0.3666173219680786, 0.2551802694797516, 0.3782024085521698], [0.37018170952796936, 0.24643611907958984, 0.3833822011947632], [0.3778906762599945, 0.24324078857898712, 0.37886857986450195], [0.376289039850235, 0.24552279710769653, 0.3781881630420685], [0.3765649199485779, 0.2433536797761917, 0.3800813853740692], [0.37590518593788147, 0.23972958326339722, 0.3843652606010437], [0.37144479155540466, 0.2523231506347656, 0.3762321174144745], [0.3844572603702545, 0.2349812239408493, 0.38056156039237976], [0.36856281757354736, 0.25159698724746704, 0.37984025478363037], [0.378235787153244, 0.24352529644966125, 0.37823888659477234], [0.3748584985733032, 0.24165259301662445, 0.3834889829158783], [0.3808934688568115, 0.24357792735099792, 0.37552860379219055], [0.3752143681049347, 0.24114128947257996, 0.38364437222480774], [0.37368518114089966, 0.24576714634895325, 0.3805476427078247], [0.37091636657714844, 0.24740448594093323, 0.38167911767959595], [0.3687674403190613, 0.24885086715221405, 0.3823816776275635], [0.37502142786979675, 0.24604164063930511, 0.37893688678741455], [0.36452266573905945, 0.26631438732147217, 0.369162917137146], [0.3788059651851654, 0.2428365796804428, 0.378357470035553], [0.3792940080165863, 0.24271737039089203, 0.37798866629600525], [0.3773920238018036, 0.24394047260284424, 0.3786674439907074], [0.37569525837898254, 0.24421660602092743, 0.38008809089660645], [0.38431230187416077, 0.23938867449760437, 0.37629905343055725], [0.36624881625175476, 0.25732457637786865, 0.376426637172699], [0.37979209423065186, 0.24171797931194305, 0.3784898817539215], [0.3735663890838623, 0.24631436169147491, 0.38011929392814636], [0.3672540485858917, 0.2578561305999756, 0.3748897910118103], [0.3516378104686737, 0.3063267767429352, 0.3420354127883911], [0.3774230182170868, 0.24255821108818054, 0.38001880049705505], [0.3777235448360443, 0.2421606332063675, 0.3801157772541046], [0.3764320909976959, 0.24307820200920105, 0.38048967719078064], [0.37888556718826294, 0.24109436571598053, 0.3800201117992401], [0.3797175884246826, 0.23737750947475433, 0.38290488719940186], [0.37639766931533813, 0.24245232343673706, 0.3811499774456024], [0.37882092595100403, 0.24032928049564362, 0.38084977865219116], [0.3584819734096527, 0.2752883732318878, 0.36622968316078186], [0.37705397605895996, 0.24328133463859558, 0.37966465950012207], [0.37337180972099304, 0.24595139920711517, 0.3806767463684082], [0.38170838356018066, 0.23888856172561646, 0.37940308451652527], [0.3759083151817322, 0.24395106732845306, 0.38014066219329834], [0.3659749925136566, 0.2572329640388489, 0.3767920732498169], [0.3694509267807007, 0.24698664247989655, 0.3835624158382416], [0.36800727248191833, 0.24975572526454926, 0.3822370171546936], [0.37147456407546997, 0.24525293707847595, 0.38327255845069885], [0.36568954586982727, 0.2602049708366394, 0.37410545349121094], [0.37815046310424805, 0.24096325039863586, 0.3808863162994385], [0.3696196973323822, 0.2524998188018799, 0.3778804838657379], [0.3760891258716583, 0.24736586213111877, 0.3765449523925781], [0.37372976541519165, 0.24527530372142792, 0.38099491596221924], [0.37266817688941956, 0.2468913495540619, 0.38044047355651855], [0.37801310420036316, 0.24585819244384766, 0.3761286735534668], [0.37153974175453186, 0.2452549785375595, 0.3832053244113922], [0.37813571095466614, 0.24338732659816742, 0.37847694754600525], [0.3692858815193176, 0.2495565116405487, 0.38115760684013367], [0.37272578477859497, 0.2512360215187073, 0.37603816390037537], [0.3747614026069641, 0.24593207240104675, 0.37930652499198914], [0.36026114225387573, 0.27267783880233765, 0.3670610189437866], [0.37452951073646545, 0.24444665014743805, 0.3810238540172577], [0.38173845410346985, 0.2361358255147934, 0.3821256756782532], [0.37933048605918884, 0.24011445045471191, 0.38055503368377686], [0.3657456040382385, 0.26040294766426086, 0.373851478099823], [0.38148579001426697, 0.24003538489341736, 0.3784787952899933], [0.36809685826301575, 0.2522083520889282, 0.37969475984573364], [0.3839885890483856, 0.23434627056121826, 0.3816651403903961], [0.38130486011505127, 0.2431599199771881, 0.37553519010543823], [0.3747657537460327, 0.24623127281665802, 0.3790029287338257], [0.3685621917247772, 0.25272202491760254, 0.37871572375297546], [0.3783615529537201, 0.24320639669895172, 0.378432035446167], [0.36657512187957764, 0.2563777565956116, 0.3770470917224884], [0.3737635910511017, 0.24645927548408508, 0.37977713346481323], [0.3752904236316681, 0.24695172905921936, 0.37775787711143494], [0.38014885783195496, 0.2400517761707306, 0.37979942560195923], [0.36818554997444153, 0.2555611729621887, 0.37625324726104736], [0.37477976083755493, 0.24211221933364868, 0.3831080496311188], [0.3596937656402588, 0.2736510932445526, 0.3666550815105438], [0.37562093138694763, 0.24516713619232178, 0.3792119026184082], [0.37705203890800476, 0.24695110321044922, 0.37599682807922363], [0.381814181804657, 0.2342572808265686, 0.383928507566452], [0.3713822364807129, 0.24857831001281738, 0.38003942370414734], [0.3758620619773865, 0.24358735978603363, 0.3805505335330963], [0.36754363775253296, 0.2564449608325958, 0.3760114014148712], [0.3722895681858063, 0.24745412170886993, 0.380256325006485], [0.3704451620578766, 0.24430499970912933, 0.3852498531341553], [0.36927276849746704, 0.24742689728736877, 0.3833003342151642], [0.37441614270210266, 0.246115580201149, 0.37946823239326477], [0.3641880452632904, 0.2638327181339264, 0.3719792366027832], [0.36599448323249817, 0.25613832473754883, 0.3778671622276306], [0.3673376441001892, 0.2540235221385956, 0.3786388337612152], [0.3751842975616455, 0.24369439482688904, 0.38112130761146545], [0.3732950687408447, 0.24731440842151642, 0.37939053773880005], [0.3687097430229187, 0.2536337673664093, 0.3776564598083496], [0.369732141494751, 0.2467540204524994, 0.38351380825042725], [0.3679351508617401, 0.2560086250305176, 0.3760562539100647], [0.37109896540641785, 0.24786046147346497, 0.3810405731201172], [0.3736385703086853, 0.24820290505886078, 0.37815847992897034], [0.37727269530296326, 0.24267399311065674, 0.3800533413887024], [0.37808117270469666, 0.24009454250335693, 0.3818242847919464], [0.3745189905166626, 0.243112713098526, 0.3823683261871338], [0.3676247298717499, 0.2575529217720032, 0.37482234835624695], [0.3775680661201477, 0.24646201729774475, 0.37596985697746277], [0.37395814061164856, 0.246392160654068, 0.37964966893196106], [0.3819643557071686, 0.2387780398130417, 0.37925755977630615], [0.3776901364326477, 0.24767157435417175, 0.37463828921318054], [0.3657582104206085, 0.2561822831630707, 0.3780595660209656], [0.36377277970314026, 0.26417064666748047, 0.3720565438270569], [0.3750295042991638, 0.24741898477077484, 0.37755152583122253], [0.3821343779563904, 0.24089162051677704, 0.3769739270210266], [0.37067365646362305, 0.2450931817293167, 0.38423317670822144], [0.3734029531478882, 0.2475130558013916, 0.3790839910507202], [0.37484946846961975, 0.24768909811973572, 0.37746143341064453], [0.3808022439479828, 0.23751592636108398, 0.3816819190979004], [0.3771003484725952, 0.2441294938325882, 0.37877026200294495], [0.3640553653240204, 0.2699897289276123, 0.3659549057483673], [0.36424294114112854, 0.2620467245578766, 0.3737103044986725], [0.368669718503952, 0.25012072920799255, 0.3812095522880554], [0.37559983134269714, 0.24611397087574005, 0.378286212682724], [0.36376041173934937, 0.2626722753047943, 0.3735673427581787], [0.3649791479110718, 0.26021572947502136, 0.37480518221855164], [0.3775886297225952, 0.24176880717277527, 0.3806426227092743], [0.3816603720188141, 0.24088162183761597, 0.37745800614356995], [0.3773977756500244, 0.24539200961589813, 0.37721022963523865], [0.366096168756485, 0.25664323568344116, 0.37726059556007385], [0.3835594952106476, 0.23794740438461304, 0.3784931004047394], [0.3790213465690613, 0.24333614110946655, 0.37764257192611694], [0.3704091012477875, 0.2536187171936035, 0.375972181558609], [0.3823903799057007, 0.23753605782985687, 0.38007354736328125], [0.377571165561676, 0.24283523857593536, 0.3795936107635498], [0.3712281882762909, 0.2456473410129547, 0.38312453031539917], [0.3695887625217438, 0.2551887035369873, 0.3752225637435913], [0.3786739110946655, 0.24281010031700134, 0.37851598858833313], [0.3743053078651428, 0.24214021861553192, 0.3835544288158417], [0.37016651034355164, 0.24982938170433044, 0.3800041079521179], [0.3764423131942749, 0.24236960709095, 0.3811880052089691], [0.3736121356487274, 0.2472166270017624, 0.379171222448349], [0.3737694323062897, 0.24308264255523682, 0.3831479251384735], [0.38124528527259827, 0.24122627079486847, 0.37752848863601685], [0.38023245334625244, 0.23742885887622833, 0.38233864307403564], [0.38173019886016846, 0.2380094975233078, 0.3802602291107178], [0.36876189708709717, 0.2519288659095764, 0.37930920720100403], [0.3745671212673187, 0.2447054237127304, 0.38072746992111206], [0.3758128583431244, 0.24497805535793304, 0.37920910120010376], [0.38147205114364624, 0.2372412532567978, 0.3812866508960724]]
standard_test: {'validation_loss': 0.42594352960586546, 'validation_time': 0.31384873390197754, 'num_samples': 143, 'metrics': {'accuracy': 0.5244755244755245, 'precision_Class_0': np.float64(0.07142857142857142), 'recall_Class_0': np.float64(0.02702702702702703), 'f1_Class_0': np.float64(0.0392156862745098), 'support_Class_0': np.int64(37), 'precision_Class_1': np.float64(0.0), 'recall_Class_1': np.float64(0.0), 'f1_Class_1': np.float64(0.0), 'support_Class_1': np.int64(19), 'precision_Class_2': np.float64(0.5736434108527132), 'recall_Class_2': np.float64(0.8505747126436781), 'f1_Class_2': np.float64(0.6851851851851852), 'support_Class_2': np.int64(87), 'precision_macro': np.float64(0.21502399409376152), 'recall_macro': np.float64(0.2925339132235684), 'f1_macro': np.float64(0.24146695715323166), 'precision_weighted': np.float64(0.3674813558534489), 'recall_weighted': np.float64(0.5244755244755245), 'f1_weighted': np.float64(0.4270076328899859), 'confusion_matrix': [[1, 0, 36], [0, 0, 19], [13, 0, 74]], 'auc_roc_ovr': 0.5203282156896898, 'auc_roc_ovo': np.float64(0.5478273191521831), 'classification_report': {'Class_0': {'precision': 0.07142857142857142, 'recall': 0.02702702702702703, 'f1-score': 0.0392156862745098, 'support': 37.0}, 'Class_1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 19.0}, 'Class_2': {'precision': 0.5736434108527132, 'recall': 0.8505747126436781, 'f1-score': 0.6851851851851852, 'support': 87.0}, 'accuracy': 0.5244755244755245, 'macro avg': {'precision': 0.21502399409376152, 'recall': 0.2925339132235684, 'f1-score': 0.24146695715323166, 'support': 143.0}, 'weighted avg': {'precision': 0.3674813558534489, 'recall': 0.5244755244755245, 'f1-score': 0.4270076328899859, 'support': 143.0}}, 'intra_class_similarity_mean': np.float32(0.9962435), 'intra_class_similarity_std': np.float32(0.003778463), 'inter_class_similarity_mean': np.float32(0.9960342), 'inter_class_similarity_std': np.float32(0.004104488), 'class_separation': np.float32(0.00020927191), 'eer': 1.0, 'eer_threshold': 0.0}, 'predictions': [2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 0, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 0, 2, 2, 2, 0, 2, 0, 2, 2, 2, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 2, 2, 2, 2, 0, 2, 2, 2, 0, 2, 2], 'targets': [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2], 'probabilities': [[0.3696979284286499, 0.24883140623569489, 0.3814707100391388], [0.3743830919265747, 0.24698880314826965, 0.37862807512283325], [0.3690684139728546, 0.2482450008392334, 0.3826865255832672], [0.3735652267932892, 0.24870266020298004, 0.3777320981025696], [0.37205788493156433, 0.24809791147708893, 0.37984421849250793], [0.3686942756175995, 0.24876481294631958, 0.3825409412384033], [0.3724340498447418, 0.24516406655311584, 0.38240185379981995], [0.36801472306251526, 0.25493231415748596, 0.3770529627799988], [0.3682364225387573, 0.25453823804855347, 0.3772253394126892], [0.36684635281562805, 0.2567999064922333, 0.37635377049446106], [0.3690694272518158, 0.2564663887023926, 0.374464213848114], [0.36454251408576965, 0.2609952986240387, 0.37446218729019165], [0.3741520941257477, 0.24662211537361145, 0.37922585010528564], [0.36916160583496094, 0.2503206133842468, 0.38051778078079224], [0.3746679723262787, 0.24447104334831238, 0.38086098432540894], [0.36721569299697876, 0.2606227397918701, 0.37216153740882874], [0.36778584122657776, 0.25411200523376465, 0.3781021237373352], [0.38233545422554016, 0.23984163999557495, 0.3778229355812073], [0.36832571029663086, 0.2566464841365814, 0.3750278651714325], [0.3685663640499115, 0.2568533420562744, 0.3745802938938141], [0.3709509074687958, 0.24964475631713867, 0.37940436601638794], [0.37304624915122986, 0.2528035640716553, 0.3741501569747925], [0.37145087122917175, 0.24578554928302765, 0.3827635645866394], [0.37096506357192993, 0.24698150157928467, 0.382053405046463], [0.3677555322647095, 0.2544492185115814, 0.3777952194213867], [0.37100762128829956, 0.24614867568016052, 0.3828437030315399], [0.3735366463661194, 0.2445944994688034, 0.3818688690662384], [0.37180212140083313, 0.24565158784389496, 0.3825463652610779], [0.36958396434783936, 0.24617137014865875, 0.3842446506023407], [0.3716345429420471, 0.24422796070575714, 0.38413751125335693], [0.3705626130104065, 0.24629828333854675, 0.38313910365104675], [0.3756616413593292, 0.24493862688541412, 0.37939974665641785], [0.36818116903305054, 0.24892908334732056, 0.3828897476196289], [0.3778320550918579, 0.24049817025661469, 0.381669819355011], [0.3747061491012573, 0.2424202710390091, 0.3828735649585724], [0.36288323998451233, 0.26875039935112, 0.3683663010597229], [0.3694359362125397, 0.2496606707572937, 0.380903422832489], [0.36972442269325256, 0.24860234558582306, 0.3816732168197632], [0.3680039048194885, 0.2471044808626175, 0.3848915696144104], [0.373210072517395, 0.2492576390504837, 0.37753236293792725], [0.37034615874290466, 0.24709655344486237, 0.38255730271339417], [0.3684896230697632, 0.2534571588039398, 0.378053218126297], [0.3674250841140747, 0.25388750433921814, 0.37868738174438477], [0.3684600293636322, 0.24911853671073914, 0.38242146372795105], [0.36980074644088745, 0.24544474482536316, 0.384754478931427], [0.3613590896129608, 0.268844336271286, 0.3697965443134308], [0.3723923861980438, 0.24662023782730103, 0.38098737597465515], [0.37218454480171204, 0.24986957013607025, 0.37794584035873413], [0.3637178838253021, 0.26565101742744446, 0.37063106894493103], [0.36858344078063965, 0.25225207209587097, 0.379164457321167], [0.36986401677131653, 0.24677309393882751, 0.38336291909217834], [0.3669643700122833, 0.25411802530288696, 0.3789176046848297], [0.3655685484409332, 0.256995290517807, 0.37743622064590454], [0.36956119537353516, 0.25102031230926514, 0.3794184923171997], [0.3670576214790344, 0.2540019452571869, 0.3789404332637787], [0.36935949325561523, 0.24674656987190247, 0.3838938772678375], [0.37280789017677307, 0.24971236288547516, 0.3774797320365906], [0.37611493468284607, 0.24403157830238342, 0.3798534870147705], [0.3684867322444916, 0.25144848227500916, 0.3800647258758545], [0.3773135244846344, 0.2449846714735031, 0.3777017891407013], [0.37713223695755005, 0.24525697529315948, 0.37761083245277405], [0.37055444717407227, 0.24623934924602509, 0.38320621848106384], [0.3816896975040436, 0.24239414930343628, 0.3759162127971649], [0.36526060104370117, 0.25952988862991333, 0.3752094805240631], [0.36595088243484497, 0.25919997692108154, 0.3748491704463959], [0.37921085953712463, 0.24074983596801758, 0.3800393342971802], [0.3758491575717926, 0.2406470775604248, 0.3835037648677826], [0.3752751350402832, 0.24652282893657684, 0.37820199131965637], [0.3774367570877075, 0.24358278512954712, 0.378980427980423], [0.37051835656166077, 0.24623870849609375, 0.38324299454689026], [0.3723600506782532, 0.24606801569461823, 0.3815719187259674], [0.37420424818992615, 0.2462836503982544, 0.37951207160949707], [0.37325260043144226, 0.24458768963813782, 0.38215968012809753], [0.3777564764022827, 0.24144676327705383, 0.38079673051834106], [0.37787318229675293, 0.24563783407211304, 0.37648892402648926], [0.37832605838775635, 0.24187244474887848, 0.379801481962204], [0.37151390314102173, 0.24909144639968872, 0.37939465045928955], [0.37623679637908936, 0.24472644925117493, 0.3790367841720581], [0.3788910210132599, 0.23811323940753937, 0.38299572467803955], [0.3708433508872986, 0.24841637909412384, 0.3807402551174164], [0.37348252534866333, 0.2472986876964569, 0.37921878695487976], [0.37689870595932007, 0.2473507970571518, 0.37575048208236694], [0.36952313780784607, 0.24712608754634857, 0.38335078954696655], [0.36452728509902954, 0.2591535747051239, 0.37631914019584656], [0.3737410604953766, 0.2380470186471939, 0.3882119059562683], [0.37209221720695496, 0.25005313754081726, 0.37785467505455017], [0.3793739974498749, 0.2371596097946167, 0.38346633315086365], [0.3735876977443695, 0.2448691874742508, 0.3815430998802185], [0.37998759746551514, 0.2427866905927658, 0.3772256374359131], [0.3774712085723877, 0.24141697585582733, 0.3811117708683014], [0.3782326281070709, 0.24357186257839203, 0.37819552421569824], [0.3769048750400543, 0.2421516627073288, 0.38094353675842285], [0.36336612701416016, 0.25949251651763916, 0.3771413564682007], [0.3755306601524353, 0.24469871819019318, 0.3797706067562103], [0.37420231103897095, 0.24704201519489288, 0.37875571846961975], [0.373231440782547, 0.24728231132030487, 0.3794862926006317], [0.3679208755493164, 0.25598326325416565, 0.37609586119651794], [0.3726634383201599, 0.24596841633319855, 0.38136810064315796], [0.3715420663356781, 0.24697241187095642, 0.3814854919910431], [0.37239065766334534, 0.24712039530277252, 0.3804889917373657], [0.36674225330352783, 0.25597208738327026, 0.3772856295108795], [0.3775917887687683, 0.24585701525211334, 0.37655121088027954], [0.36577722430229187, 0.2591440677642822, 0.3750787079334259], [0.3784436285495758, 0.24449379742145538, 0.3770625591278076], [0.37792322039604187, 0.24396736919879913, 0.3781093955039978], [0.3791361153125763, 0.23617735505104065, 0.38468655943870544], [0.3780738413333893, 0.24068112671375275, 0.3812450170516968], [0.37849271297454834, 0.24369299411773682, 0.37781429290771484], [0.3670293390750885, 0.25711819529533386, 0.37585240602493286], [0.3808387219905853, 0.23910482227802277, 0.3800565004348755], [0.377498060464859, 0.24449346959590912, 0.3780084550380707], [0.37349170446395874, 0.24856454133987427, 0.377943754196167], [0.3661789000034332, 0.2563043534755707, 0.3775166869163513], [0.38068196177482605, 0.23975136876106262, 0.3795666992664337], [0.37497225403785706, 0.24605600535869598, 0.3789716958999634], [0.3760099411010742, 0.2435959428548813, 0.38039419054985046], [0.3723418116569519, 0.24644316732883453, 0.38121497631073], [0.37620919942855835, 0.2414311170578003, 0.38235968351364136], [0.3746432960033417, 0.24318552017211914, 0.3821712136268616], [0.36868977546691895, 0.25117823481559753, 0.3801320493221283], [0.36997345089912415, 0.24749061465263367, 0.3825359642505646], [0.3726966381072998, 0.24372178316116333, 0.3835815489292145], [0.3739120066165924, 0.24387885630130768, 0.3822091817855835], [0.37831103801727295, 0.24305123090744019, 0.3786376714706421], [0.3726988136768341, 0.24492892622947693, 0.38237226009368896], [0.37514734268188477, 0.2461518496274948, 0.3787008225917816], [0.3779442310333252, 0.24392135441303253, 0.3781343698501587], [0.3772237002849579, 0.24217940866947174, 0.3805968761444092], [0.37034115195274353, 0.24552810192108154, 0.3841307461261749], [0.3728271424770355, 0.2486010640859604, 0.3785717785358429], [0.36467838287353516, 0.26036736369132996, 0.3749542832374573], [0.38605937361717224, 0.2388167530298233, 0.37512388825416565], [0.373543381690979, 0.24577130377292633, 0.38068532943725586], [0.3716607391834259, 0.24683627486228943, 0.38150301575660706], [0.3762662410736084, 0.24479922652244568, 0.3789345622062683], [0.37627536058425903, 0.24474158883094788, 0.3789830505847931], [0.37996456027030945, 0.2409827709197998, 0.37905269861221313], [0.37165579199790955, 0.2468050867319107, 0.38153910636901855], [0.3753000795841217, 0.2390146106481552, 0.3856852650642395], [0.36619412899017334, 0.2562731206417084, 0.3775327503681183], [0.3508232831954956, 0.3004007637500763, 0.3487759530544281], [0.36708930134773254, 0.2580561637878418, 0.37485453486442566], [0.37403377890586853, 0.24808408319950104, 0.3778821527957916]], 'test_type': 'standard'}
augmentation_robustness: {'noise_light': {'loss': 0.42594352960586546, 'accuracy': 0.5244755244755245, 'metrics': {'accuracy': 0.5244755244755245, 'precision_Class_0': np.float64(0.07142857142857142), 'recall_Class_0': np.float64(0.02702702702702703), 'f1_Class_0': np.float64(0.0392156862745098), 'support_Class_0': np.int64(37), 'precision_Class_1': np.float64(0.0), 'recall_Class_1': np.float64(0.0), 'f1_Class_1': np.float64(0.0), 'support_Class_1': np.int64(19), 'precision_Class_2': np.float64(0.5736434108527132), 'recall_Class_2': np.float64(0.8505747126436781), 'f1_Class_2': np.float64(0.6851851851851852), 'support_Class_2': np.int64(87), 'precision_macro': np.float64(0.21502399409376152), 'recall_macro': np.float64(0.2925339132235684), 'f1_macro': np.float64(0.24146695715323166), 'precision_weighted': np.float64(0.3674813558534489), 'recall_weighted': np.float64(0.5244755244755245), 'f1_weighted': np.float64(0.4270076328899859), 'confusion_matrix': [[1, 0, 36], [0, 0, 19], [13, 0, 74]], 'classification_report': {'Class_0': {'precision': 0.07142857142857142, 'recall': 0.02702702702702703, 'f1-score': 0.0392156862745098, 'support': 37.0}, 'Class_1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 19.0}, 'Class_2': {'precision': 0.5736434108527132, 'recall': 0.8505747126436781, 'f1-score': 0.6851851851851852, 'support': 87.0}, 'accuracy': 0.5244755244755245, 'macro avg': {'precision': 0.21502399409376152, 'recall': 0.2925339132235684, 'f1-score': 0.24146695715323166, 'support': 143.0}, 'weighted avg': {'precision': 0.3674813558534489, 'recall': 0.5244755244755245, 'f1-score': 0.4270076328899859, 'support': 143.0}}}}, 'noise_moderate': {'loss': 0.42594352960586546, 'accuracy': 0.5244755244755245, 'metrics': {'accuracy': 0.5244755244755245, 'precision_Class_0': np.float64(0.07142857142857142), 'recall_Class_0': np.float64(0.02702702702702703), 'f1_Class_0': np.float64(0.0392156862745098), 'support_Class_0': np.int64(37), 'precision_Class_1': np.float64(0.0), 'recall_Class_1': np.float64(0.0), 'f1_Class_1': np.float64(0.0), 'support_Class_1': np.int64(19), 'precision_Class_2': np.float64(0.5736434108527132), 'recall_Class_2': np.float64(0.8505747126436781), 'f1_Class_2': np.float64(0.6851851851851852), 'support_Class_2': np.int64(87), 'precision_macro': np.float64(0.21502399409376152), 'recall_macro': np.float64(0.2925339132235684), 'f1_macro': np.float64(0.24146695715323166), 'precision_weighted': np.float64(0.3674813558534489), 'recall_weighted': np.float64(0.5244755244755245), 'f1_weighted': np.float64(0.4270076328899859), 'confusion_matrix': [[1, 0, 36], [0, 0, 19], [13, 0, 74]], 'classification_report': {'Class_0': {'precision': 0.07142857142857142, 'recall': 0.02702702702702703, 'f1-score': 0.0392156862745098, 'support': 37.0}, 'Class_1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 19.0}, 'Class_2': {'precision': 0.5736434108527132, 'recall': 0.8505747126436781, 'f1-score': 0.6851851851851852, 'support': 87.0}, 'accuracy': 0.5244755244755245, 'macro avg': {'precision': 0.21502399409376152, 'recall': 0.2925339132235684, 'f1-score': 0.24146695715323166, 'support': 143.0}, 'weighted avg': {'precision': 0.3674813558534489, 'recall': 0.5244755244755245, 'f1-score': 0.4270076328899859, 'support': 143.0}}}}, 'noise_heavy': {'loss': 0.42594352960586546, 'accuracy': 0.5244755244755245, 'metrics': {'accuracy': 0.5244755244755245, 'precision_Class_0': np.float64(0.07142857142857142), 'recall_Class_0': np.float64(0.02702702702702703), 'f1_Class_0': np.float64(0.0392156862745098), 'support_Class_0': np.int64(37), 'precision_Class_1': np.float64(0.0), 'recall_Class_1': np.float64(0.0), 'f1_Class_1': np.float64(0.0), 'support_Class_1': np.int64(19), 'precision_Class_2': np.float64(0.5736434108527132), 'recall_Class_2': np.float64(0.8505747126436781), 'f1_Class_2': np.float64(0.6851851851851852), 'support_Class_2': np.int64(87), 'precision_macro': np.float64(0.21502399409376152), 'recall_macro': np.float64(0.2925339132235684), 'f1_macro': np.float64(0.24146695715323166), 'precision_weighted': np.float64(0.3674813558534489), 'recall_weighted': np.float64(0.5244755244755245), 'f1_weighted': np.float64(0.4270076328899859), 'confusion_matrix': [[1, 0, 36], [0, 0, 19], [13, 0, 74]], 'classification_report': {'Class_0': {'precision': 0.07142857142857142, 'recall': 0.02702702702702703, 'f1-score': 0.0392156862745098, 'support': 37.0}, 'Class_1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 19.0}, 'Class_2': {'precision': 0.5736434108527132, 'recall': 0.8505747126436781, 'f1-score': 0.6851851851851852, 'support': 87.0}, 'accuracy': 0.5244755244755245, 'macro avg': {'precision': 0.21502399409376152, 'recall': 0.2925339132235684, 'f1-score': 0.24146695715323166, 'support': 143.0}, 'weighted avg': {'precision': 0.3674813558534489, 'recall': 0.5244755244755245, 'f1-score': 0.4270076328899859, 'support': 143.0}}}}, 'speed_light': {'loss': 0.42594352960586546, 'accuracy': 0.5244755244755245, 'metrics': {'accuracy': 0.5244755244755245, 'precision_Class_0': np.float64(0.07142857142857142), 'recall_Class_0': np.float64(0.02702702702702703), 'f1_Class_0': np.float64(0.0392156862745098), 'support_Class_0': np.int64(37), 'precision_Class_1': np.float64(0.0), 'recall_Class_1': np.float64(0.0), 'f1_Class_1': np.float64(0.0), 'support_Class_1': np.int64(19), 'precision_Class_2': np.float64(0.5736434108527132), 'recall_Class_2': np.float64(0.8505747126436781), 'f1_Class_2': np.float64(0.6851851851851852), 'support_Class_2': np.int64(87), 'precision_macro': np.float64(0.21502399409376152), 'recall_macro': np.float64(0.2925339132235684), 'f1_macro': np.float64(0.24146695715323166), 'precision_weighted': np.float64(0.3674813558534489), 'recall_weighted': np.float64(0.5244755244755245), 'f1_weighted': np.float64(0.4270076328899859), 'confusion_matrix': [[1, 0, 36], [0, 0, 19], [13, 0, 74]], 'classification_report': {'Class_0': {'precision': 0.07142857142857142, 'recall': 0.02702702702702703, 'f1-score': 0.0392156862745098, 'support': 37.0}, 'Class_1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 19.0}, 'Class_2': {'precision': 0.5736434108527132, 'recall': 0.8505747126436781, 'f1-score': 0.6851851851851852, 'support': 87.0}, 'accuracy': 0.5244755244755245, 'macro avg': {'precision': 0.21502399409376152, 'recall': 0.2925339132235684, 'f1-score': 0.24146695715323166, 'support': 143.0}, 'weighted avg': {'precision': 0.3674813558534489, 'recall': 0.5244755244755245, 'f1-score': 0.4270076328899859, 'support': 143.0}}}}, 'speed_moderate': {'loss': 0.42594352960586546, 'accuracy': 0.5244755244755245, 'metrics': {'accuracy': 0.5244755244755245, 'precision_Class_0': np.float64(0.07142857142857142), 'recall_Class_0': np.float64(0.02702702702702703), 'f1_Class_0': np.float64(0.0392156862745098), 'support_Class_0': np.int64(37), 'precision_Class_1': np.float64(0.0), 'recall_Class_1': np.float64(0.0), 'f1_Class_1': np.float64(0.0), 'support_Class_1': np.int64(19), 'precision_Class_2': np.float64(0.5736434108527132), 'recall_Class_2': np.float64(0.8505747126436781), 'f1_Class_2': np.float64(0.6851851851851852), 'support_Class_2': np.int64(87), 'precision_macro': np.float64(0.21502399409376152), 'recall_macro': np.float64(0.2925339132235684), 'f1_macro': np.float64(0.24146695715323166), 'precision_weighted': np.float64(0.3674813558534489), 'recall_weighted': np.float64(0.5244755244755245), 'f1_weighted': np.float64(0.4270076328899859), 'confusion_matrix': [[1, 0, 36], [0, 0, 19], [13, 0, 74]], 'classification_report': {'Class_0': {'precision': 0.07142857142857142, 'recall': 0.02702702702702703, 'f1-score': 0.0392156862745098, 'support': 37.0}, 'Class_1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 19.0}, 'Class_2': {'precision': 0.5736434108527132, 'recall': 0.8505747126436781, 'f1-score': 0.6851851851851852, 'support': 87.0}, 'accuracy': 0.5244755244755245, 'macro avg': {'precision': 0.21502399409376152, 'recall': 0.2925339132235684, 'f1-score': 0.24146695715323166, 'support': 143.0}, 'weighted avg': {'precision': 0.3674813558534489, 'recall': 0.5244755244755245, 'f1-score': 0.4270076328899859, 'support': 143.0}}}}, 'volume_light': {'loss': 0.42594352960586546, 'accuracy': 0.5244755244755245, 'metrics': {'accuracy': 0.5244755244755245, 'precision_Class_0': np.float64(0.07142857142857142), 'recall_Class_0': np.float64(0.02702702702702703), 'f1_Class_0': np.float64(0.0392156862745098), 'support_Class_0': np.int64(37), 'precision_Class_1': np.float64(0.0), 'recall_Class_1': np.float64(0.0), 'f1_Class_1': np.float64(0.0), 'support_Class_1': np.int64(19), 'precision_Class_2': np.float64(0.5736434108527132), 'recall_Class_2': np.float64(0.8505747126436781), 'f1_Class_2': np.float64(0.6851851851851852), 'support_Class_2': np.int64(87), 'precision_macro': np.float64(0.21502399409376152), 'recall_macro': np.float64(0.2925339132235684), 'f1_macro': np.float64(0.24146695715323166), 'precision_weighted': np.float64(0.3674813558534489), 'recall_weighted': np.float64(0.5244755244755245), 'f1_weighted': np.float64(0.4270076328899859), 'confusion_matrix': [[1, 0, 36], [0, 0, 19], [13, 0, 74]], 'classification_report': {'Class_0': {'precision': 0.07142857142857142, 'recall': 0.02702702702702703, 'f1-score': 0.0392156862745098, 'support': 37.0}, 'Class_1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 19.0}, 'Class_2': {'precision': 0.5736434108527132, 'recall': 0.8505747126436781, 'f1-score': 0.6851851851851852, 'support': 87.0}, 'accuracy': 0.5244755244755245, 'macro avg': {'precision': 0.21502399409376152, 'recall': 0.2925339132235684, 'f1-score': 0.24146695715323166, 'support': 143.0}, 'weighted avg': {'precision': 0.3674813558534489, 'recall': 0.5244755244755245, 'f1-score': 0.4270076328899859, 'support': 143.0}}}}, 'volume_moderate': {'loss': 0.42594352960586546, 'accuracy': 0.5244755244755245, 'metrics': {'accuracy': 0.5244755244755245, 'precision_Class_0': np.float64(0.07142857142857142), 'recall_Class_0': np.float64(0.02702702702702703), 'f1_Class_0': np.float64(0.0392156862745098), 'support_Class_0': np.int64(37), 'precision_Class_1': np.float64(0.0), 'recall_Class_1': np.float64(0.0), 'f1_Class_1': np.float64(0.0), 'support_Class_1': np.int64(19), 'precision_Class_2': np.float64(0.5736434108527132), 'recall_Class_2': np.float64(0.8505747126436781), 'f1_Class_2': np.float64(0.6851851851851852), 'support_Class_2': np.int64(87), 'precision_macro': np.float64(0.21502399409376152), 'recall_macro': np.float64(0.2925339132235684), 'f1_macro': np.float64(0.24146695715323166), 'precision_weighted': np.float64(0.3674813558534489), 'recall_weighted': np.float64(0.5244755244755245), 'f1_weighted': np.float64(0.4270076328899859), 'confusion_matrix': [[1, 0, 36], [0, 0, 19], [13, 0, 74]], 'classification_report': {'Class_0': {'precision': 0.07142857142857142, 'recall': 0.02702702702702703, 'f1-score': 0.0392156862745098, 'support': 37.0}, 'Class_1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 19.0}, 'Class_2': {'precision': 0.5736434108527132, 'recall': 0.8505747126436781, 'f1-score': 0.6851851851851852, 'support': 87.0}, 'accuracy': 0.5244755244755245, 'macro avg': {'precision': 0.21502399409376152, 'recall': 0.2925339132235684, 'f1-score': 0.24146695715323166, 'support': 143.0}, 'weighted avg': {'precision': 0.3674813558534489, 'recall': 0.5244755244755245, 'f1-score': 0.4270076328899859, 'support': 143.0}}}}, 'combined_light': {'loss': 0.42594352960586546, 'accuracy': 0.5244755244755245, 'metrics': {'accuracy': 0.5244755244755245, 'precision_Class_0': np.float64(0.07142857142857142), 'recall_Class_0': np.float64(0.02702702702702703), 'f1_Class_0': np.float64(0.0392156862745098), 'support_Class_0': np.int64(37), 'precision_Class_1': np.float64(0.0), 'recall_Class_1': np.float64(0.0), 'f1_Class_1': np.float64(0.0), 'support_Class_1': np.int64(19), 'precision_Class_2': np.float64(0.5736434108527132), 'recall_Class_2': np.float64(0.8505747126436781), 'f1_Class_2': np.float64(0.6851851851851852), 'support_Class_2': np.int64(87), 'precision_macro': np.float64(0.21502399409376152), 'recall_macro': np.float64(0.2925339132235684), 'f1_macro': np.float64(0.24146695715323166), 'precision_weighted': np.float64(0.3674813558534489), 'recall_weighted': np.float64(0.5244755244755245), 'f1_weighted': np.float64(0.4270076328899859), 'confusion_matrix': [[1, 0, 36], [0, 0, 19], [13, 0, 74]], 'classification_report': {'Class_0': {'precision': 0.07142857142857142, 'recall': 0.02702702702702703, 'f1-score': 0.0392156862745098, 'support': 37.0}, 'Class_1': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 19.0}, 'Class_2': {'precision': 0.5736434108527132, 'recall': 0.8505747126436781, 'f1-score': 0.6851851851851852, 'support': 87.0}, 'accuracy': 0.5244755244755245, 'macro avg': {'precision': 0.21502399409376152, 'recall': 0.2925339132235684, 'f1-score': 0.24146695715323166, 'support': 143.0}, 'weighted avg': {'precision': 0.3674813558534489, 'recall': 0.5244755244755245, 'f1-score': 0.4270076328899859, 'support': 143.0}}}}}
performance_profile: {'inference_performance': {'mean_inference_time': np.float64(0.0006418657844187692), 'std_inference_time': np.float64(3.7465627994-05), 'min_inference_time': np.float64(0.0005991850048303604), 'max_inference_time': np.float64(0.0008548320038244128), 'median_inference_time': np.float64(0.0006331400072667748), 'throughput_samples_per_second': np.float64(1557.9581031968128)}, 'memory_usage': {'initial_memory_mb': 66.80322265625, 'peak_memory_mb': 74.4580078125, 'final_memory_mb': 66.8037109375, 'memory_increase_mb': 7.65478515625}, 'model_parameters': {'total_parameters': 2586371, 'trainable_parameters': 2586371, 'non_trainable_parameters': 0}, 'model_size': {'total_size_bytes': 10360412, 'total_size_mb': 9.88045883178711, 'parameter_size_mb': 9.866222381591797, 'buffer_size_mb': 0.0142364501953125}, 'batch_performance': {'batch_1': {'inference_time': np.float64(0.0006414205534383654), 'throughput': np.float64(1559.0395328610105)}, 'batch_8': {'inference_time': np.float64(0.001541531810653396), 'throughput': np.float64(5189.643148920234)}, 'batch_16': {'inference_time': np.float64(0.007272461851243861), 'throughput': np.float64(2200.080292929059)}, 'batch_32': {'inference_time': np.float64(0.006635047844611108), 'throughput': np.float64(4822.874039407258)}}}
robustness_tests: {'confidence_thresholds': {}, 'input_corruptions': {'zero_out_10': {'loss': 0.4275360405445099, 'accuracy': np.float64(0.5384615384615384)}, 'zero_out_25': {'loss': 0.43089011311531067, 'accuracy': np.float64(0.5384615384615384)}, 'gaussian_noise_01': {'loss': 0.4257518708705902, 'accuracy': np.float64(0.5244755244755245)}, 'gaussian_noise_02': {'loss': 0.425038743019104, 'accuracy': np.float64(0.5384615384615384)}, 'dropout_10': {'loss': 0.4238474786281586, 'accuracy': np.float64(0.5524475524475524)}, 'dropout_25': {'loss': 0.419542133808136, 'accuracy': np.float64(0.5594405594405595)}}, 'adversarial_robustness': {'epsilon_0.01': {'loss': 0.4338447034358978, 'accuracy': np.float64(0.3006993006993007)}, 'epsilon_0.05': {'loss': 0.4622612178325653, 'accuracy': np.float64(0.0)}, 'epsilon_0.1': {'loss': 0.49174827337265015, 'accuracy': np.float64(0.0)}, 'epsilon_0.2': {'loss': 0.5354408085346222, 'accuracy': np.float64(0.0)}}}

Wakeword Detection Specific Results

Classification Performance

Confusion Matrix

Actual\Predicted    custom system_c negative 
      custom         1        0       36 
system_comma         0        0       19 
    negative        13        0       74

Robustness Tests

Input Corruption Robustness

zero_out_10: 0.5385 accuracy
zero_out_25: 0.5385 accuracy
gaussian_noise_01: 0.5245 accuracy
gaussian_noise_02: 0.5385 accuracy
dropout_10: 0.5524 accuracy
dropout_25: 0.5594 accuracy

Performance Profile

Inference time: 0.64 ms
Throughput: 1558.0 samples/sec
Model size: 9.88 MB

Deployment Recommendations

Detection Threshold

Recommended threshold: 0.5
Monitor false positive/negative rates in production
Consider threshold adjustment based on use case requirements

Hardware Requirements

Suitable for real-time processing on most devices

Model Optimization

Consider using lightweight variant for edge deployment
Model has been reparameterized for efficient inference

training_report.md 50 KB Permalink Cronologia Originale

Training Report: unified_test

Basic Information

Model Architecture

Training Results

Dataset Information

Hardware Information

Wakeword Information

Test Results

Wakeword Detection Specific Results

Classification Performance

Confusion Matrix

Robustness Tests

Input Corruption Robustness

Performance Profile

Deployment Recommendations

Detection Threshold

Hardware Requirements

Model Optimization

training_report.md 50 KB

Permalink Cronologia Originale