Add files using upload-large-folder tool

decoders/crossreg_reg4/eval_v2/adhd200_dx__patch__logistic/eval_table.csv

@@ -0,0 +1,203 @@
+model,repr,clf,dataset,trial,C,split,acc,acc_std,f1,f1_std,bacc,bacc_std
+flat_mae,patch,logistic,adhd200_dx,,0.005994842503189409,train,0.7424657534246575,0.021904998996738183,0.7317351598173516,0.023336040318595548,0.7287964828723209,0.022882418202841488
+flat_mae,patch,logistic,adhd200_dx,,0.005994842503189409,test,0.5846153846153846,0.05747490525201247,0.5501153550371699,0.0649772448058626,0.556949806949807,0.05949809369980641
+flat_mae,patch,logistic,adhd200_dx,1,0.005994842503189409,train,0.7315068493150685,0.02175946294856782,0.7161111111111111,0.023687396156063766,0.7133479880319961,0.022699840669655757
+flat_mae,patch,logistic,adhd200_dx,1,0.005994842503189409,test,0.5846153846153846,0.0595373961343754,0.5699583435432491,0.061553726399329114,0.5699806949806949,0.06028514680245595
+flat_mae,patch,logistic,adhd200_dx,2,0.3593813663804626,train,0.958904109589041,0.010228133723195108,0.9580527271473556,0.010483771568235135,0.9564175367893998,0.010892055823648775
+flat_mae,patch,logistic,adhd200_dx,2,0.3593813663804626,test,0.5846153846153846,0.06186923411662306,0.578226387887527,0.06294971050534451,0.5786679536679536,0.06281953395438065
+flat_mae,patch,logistic,adhd200_dx,3,0.005994842503189409,train,0.7397260273972602,0.0229039647370395,0.7272712972211954,0.024897916295657085,0.7242168895402088,0.02425350564676892
+flat_mae,patch,logistic,adhd200_dx,3,0.005994842503189409,test,0.6,0.056654301006059615,0.570630081300813,0.0629693519564316,0.5748069498069498,0.05860606934204386
+flat_mae,patch,logistic,adhd200_dx,4,0.005994842503189409,train,0.7205479452054795,0.02427381860862127,0.7089041095890412,0.025619520916078338,0.706509128656042,0.025038369760472984
+flat_mae,patch,logistic,adhd200_dx,4,0.005994842503189409,test,0.6307692307692307,0.05718021754161582,0.6264367816091954,0.05816225509061837,0.627895752895753,0.05866661248897376
+flat_mae,patch,logistic,adhd200_dx,5,0.046415888336127774,train,0.8356164383561644,0.019340680580432475,0.8312060673325934,0.020140055010896152,0.8285400256457227,0.020331923833431437
+flat_mae,patch,logistic,adhd200_dx,5,0.046415888336127774,test,0.5692307692307692,0.06097359211093965,0.5666666666666667,0.06144937965341382,0.5694980694980695,0.06208679449276509
+flat_mae,patch,logistic,adhd200_dx,6,0.3593813663804626,train,0.9534246575342465,0.010953438363008551,0.9523018856321441,0.01130584772176831,0.9494107589912683,0.011886781727977814
+flat_mae,patch,logistic,adhd200_dx,6,0.3593813663804626,test,0.6615384615384615,0.059135956082362856,0.6575670498084292,0.059942574087972475,0.6592664092664093,0.06033058277089562
+flat_mae,patch,logistic,adhd200_dx,7,0.046415888336127774,train,0.8410958904109589,0.019123523176318054,0.8362165005879805,0.020052715949169946,0.8326769249557306,0.020238647422966088
+flat_mae,patch,logistic,adhd200_dx,7,0.046415888336127774,test,0.5846153846153846,0.05917168681650083,0.5830363506771205,0.05919782582612036,0.5873552123552124,0.05994967511895883
+flat_mae,patch,logistic,adhd200_dx,8,0.005994842503189409,train,0.7315068493150685,0.022803403827525912,0.7183109683109683,0.024956083596163036,0.7155003968980888,0.02419498300613208
+flat_mae,patch,logistic,adhd200_dx,8,0.005994842503189409,test,0.6307692307692307,0.06192692009096418,0.6198830409356726,0.06462276479873824,0.6192084942084942,0.06344098862392855
+flat_mae,patch,logistic,adhd200_dx,9,0.005994842503189409,train,0.7452054794520548,0.02305588045530142,0.7354763296317943,0.024523043363115803,0.7326586065824021,0.02422583777514946
+flat_mae,patch,logistic,adhd200_dx,9,0.005994842503189409,test,0.6307692307692307,0.054095563041976026,0.5962732919254659,0.06156816960167578,0.6018339768339769,0.055989328273165724
+flat_mae,patch,logistic,adhd200_dx,10,0.005994842503189409,train,0.7424657534246575,0.022914004548061827,0.7329212853406402,0.024156020773037724,0.7302314221163827,0.023809908653756675
+flat_mae,patch,logistic,adhd200_dx,10,0.005994842503189409,test,0.5846153846153846,0.05909105600180356,0.5578231292517006,0.06407308875457003,0.5612934362934363,0.060402412127793625
+flat_mae,patch,logistic,adhd200_dx,11,0.005994842503189409,train,0.7013698630136986,0.02221617195008163,0.683816926145801,0.02431744764948055,0.6823441411735971,0.023212493901744093
+flat_mae,patch,logistic,adhd200_dx,11,0.005994842503189409,test,0.6,0.05412110898595687,0.5626293995859213,0.062306311625395884,0.5704633204633205,0.05602142058640619
+flat_mae,patch,logistic,adhd200_dx,12,0.005994842503189409,train,0.726027397260274,0.023221860919076725,0.7139498432601881,0.024878197678231764,0.7113634975880808,0.02426849239439832
+flat_mae,patch,logistic,adhd200_dx,12,0.005994842503189409,test,0.5846153846153846,0.056831785422591874,0.5699583435432491,0.05977686042399589,0.5699806949806949,0.058273607132441826
+flat_mae,patch,logistic,adhd200_dx,13,0.005994842503189409,train,0.7589041095890411,0.020850981158566873,0.7482758620689656,0.022332588561961048,0.7447945289124992,0.021929752294809205
+flat_mae,patch,logistic,adhd200_dx,13,0.005994842503189409,test,0.49230769230769234,0.0588737773274718,0.4743935309973046,0.060543653424304386,0.47586872586872586,0.059637690826888815
+flat_mae,patch,logistic,adhd200_dx,14,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adhd200_dx,14,1291.5496650148827,test,0.6,0.057080124314125104,0.5833333333333333,0.06070381533140116,0.5834942084942085,0.058669222582869274
+flat_mae,patch,logistic,adhd200_dx,15,0.005994842503189409,train,0.726027397260274,0.022219753045615607,0.7118360386534454,0.02414592096639806,0.7092110887219881,0.023307827685588735
+flat_mae,patch,logistic,adhd200_dx,15,0.005994842503189409,test,0.5384615384615384,0.061283927446873124,0.5294401544401545,0.0623149636534173,0.5294401544401545,0.06222626850432534
+flat_mae,patch,logistic,adhd200_dx,16,0.005994842503189409,train,0.7232876712328767,0.021793275667674784,0.7133381544466995,0.02293349533438079,0.7110887219881541,0.022593524141794323
+flat_mae,patch,logistic,adhd200_dx,16,0.005994842503189409,test,0.5846153846153846,0.062481171601801194,0.5745454545454545,0.06394244463034299,0.5743243243243243,0.06344919493891837
+flat_mae,patch,logistic,adhd200_dx,17,0.046415888336127774,train,0.8356164383561644,0.018018927102044696,0.8308932542624166,0.01874021067288798,0.8278225560236918,0.018824477702043735
+flat_mae,patch,logistic,adhd200_dx,17,0.046415888336127774,test,0.6153846153846154,0.056206643251150754,0.5966741126830479,0.06061578789763127,0.597007722007722,0.05831458026482621
+flat_mae,patch,logistic,adhd200_dx,18,0.3593813663804626,train,0.9698630136986301,0.008740498956674862,0.9691886208934148,0.008994420012140777,0.9668437442755083,0.009678730503850482
+flat_mae,patch,logistic,adhd200_dx,18,0.3593813663804626,test,0.5692307692307692,0.05923596380818011,0.5608108108108107,0.06037803594609343,0.5608108108108107,0.060218470396470755
+flat_mae,patch,logistic,adhd200_dx,19,0.000774263682681127,train,0.6328767123287671,0.023010616231931746,0.5924166666666667,0.027261834319406555,0.6022928497282775,0.02406024823718003
+flat_mae,patch,logistic,adhd200_dx,19,0.000774263682681127,test,0.6307692307692307,0.05238849867002689,0.587737843551797,0.0629310560687302,0.5974903474903475,0.055048676802358774
+flat_mae,patch,logistic,adhd200_dx,20,0.005994842503189409,train,0.7095890410958904,0.0229524893775056,0.701244749196936,0.023724193633656646,0.6996702692800879,0.023507540272789867
+flat_mae,patch,logistic,adhd200_dx,20,0.005994842503189409,test,0.6307692307692307,0.05634593909121585,0.6036585365853658,0.06289556976806442,0.6061776061776062,0.058661890473261515
+flat_mae,patch,logistic,adhd200_dx,21,0.005994842503189409,train,0.726027397260274,0.023098341409458226,0.7095244158655376,0.025617669985713853,0.7070586798558954,0.02439054946883275
+flat_mae,patch,logistic,adhd200_dx,21,0.005994842503189409,test,0.7230769230769231,0.05417740746187419,0.7149122807017544,0.05636841728510411,0.7133204633204633,0.05579711314188372
+flat_mae,patch,logistic,adhd200_dx,22,0.005994842503189409,train,0.7178082191780822,0.021734313402037447,0.7028086734492217,0.023821578651592547,0.7004945960798681,0.022933319859288752
+flat_mae,patch,logistic,adhd200_dx,22,0.005994842503189409,test,0.6307692307692307,0.06171500829231663,0.6235521235521235,0.06334497042852516,0.6235521235521235,0.06326613240265587
+flat_mae,patch,logistic,adhd200_dx,23,0.000774263682681127,train,0.6602739726027397,0.02229287038061088,0.6377577154565246,0.02476154092559909,0.6387616779629969,0.02321867248197043
+flat_mae,patch,logistic,adhd200_dx,23,0.000774263682681127,test,0.6153846153846154,0.05174945558985287,0.5656241646618552,0.06406613287543482,0.5796332046332047,0.054493023528735494
+flat_mae,patch,logistic,adhd200_dx,24,0.005994842503189409,train,0.7205479452054795,0.021686796274929466,0.7068134568134568,0.02346880306398429,0.7043567197899493,0.022737337739779803
+flat_mae,patch,logistic,adhd200_dx,24,0.005994842503189409,test,0.5846153846153846,0.05588169244301624,0.5501153550371699,0.06169691487186124,0.556949806949807,0.05693022347504732
+flat_mae,patch,logistic,adhd200_dx,25,0.000774263682681127,train,0.6520547945205479,0.021670064911463217,0.6114450004610188,0.026221653172145115,0.6207180802344752,0.02278685906566446
+flat_mae,patch,logistic,adhd200_dx,25,0.000774263682681127,test,0.6,0.05099025083752161,0.5427489177489178,0.06181938637330176,0.5617760617760618,0.05255712965330673
+flat_mae,patch,logistic,adhd200_dx,26,0.005994842503189409,train,0.7205479452054795,0.021968759271678243,0.7053096504559271,0.024160845427025433,0.7029217805458875,0.023234977827076238
+flat_mae,patch,logistic,adhd200_dx,26,0.005994842503189409,test,0.6307692307692307,0.05501583204085829,0.6036585365853658,0.061749572456848666,0.6061776061776062,0.05731640220304975
+flat_mae,patch,logistic,adhd200_dx,27,0.046415888336127774,train,0.821917808219178,0.020108931346790333,0.8166268617494068,0.020990003327584165,0.8135342248275019,0.02107974775547443
+flat_mae,patch,logistic,adhd200_dx,27,0.046415888336127774,test,0.6153846153846154,0.060639308047448666,0.6018132810585641,0.06333183184019378,0.6013513513513513,0.06192042359549878
+flat_mae,patch,logistic,adhd200_dx,28,0.005994842503189409,train,0.726027397260274,0.021827360636567844,0.7146118721461188,0.0232255707691606,0.7120809672101117,0.022738620292532444
+flat_mae,patch,logistic,adhd200_dx,28,0.005994842503189409,test,0.5846153846153846,0.0546718148343194,0.5501153550371699,0.06075232455253754,0.556949806949807,0.0559985707062253
+flat_mae,patch,logistic,adhd200_dx,29,0.005994842503189409,train,0.7205479452054795,0.023599615150275935,0.7075320512820513,0.025331520243230213,0.7050741894119802,0.02460199328954945
+flat_mae,patch,logistic,adhd200_dx,29,0.005994842503189409,test,0.6153846153846154,0.05950202820868164,0.5966741126830479,0.06341831262949778,0.597007722007722,0.061315369217520205
+flat_mae,patch,logistic,adhd200_dx,30,0.3593813663804626,train,0.9506849315068493,0.011509994351675637,0.9496227456753772,0.011818357288142189,0.9477010441472797,0.012220399808541461
+flat_mae,patch,logistic,adhd200_dx,30,0.3593813663804626,test,0.5692307692307692,0.06032451884569405,0.564176245210728,0.06047048488448265,0.5651544401544402,0.06061516498236156
+flat_mae,patch,logistic,adhd200_dx,31,0.3593813663804626,train,0.9506849315068493,0.01138222928754019,0.9496227456753772,0.011690226536236527,0.9477010441472797,0.012122024347066197
+flat_mae,patch,logistic,adhd200_dx,31,0.3593813663804626,test,0.6,0.061585797179229435,0.5921814671814671,0.06290748518584183,0.5921814671814671,0.06264103536840496
+flat_mae,patch,logistic,adhd200_dx,32,0.046415888336127774,train,0.8273972602739726,0.019370352050558662,0.8219222637827288,0.020277634696192738,0.8183885937595408,0.020282488371607732
+flat_mae,patch,logistic,adhd200_dx,32,0.046415888336127774,test,0.5538461538461539,0.06633448682255583,0.543030303030303,0.0677295747450472,0.542953667953668,0.06714117747970659
+flat_mae,patch,logistic,adhd200_dx,33,0.3593813663804626,train,0.9698630136986301,0.008270471907532133,0.9692386665747275,0.008478910561300812,0.9675612138975392,0.00896411037959371
+flat_mae,patch,logistic,adhd200_dx,33,0.3593813663804626,test,0.5538461538461539,0.06428263037592144,0.5521501544309813,0.06465589680292005,0.555984555984556,0.06493353033693425
+flat_mae,patch,logistic,adhd200_dx,34,0.046415888336127774,train,0.8328767123287671,0.019746244586437308,0.8282352941176471,0.02052135708102231,0.8253953715576724,0.020590510236745532
+flat_mae,patch,logistic,adhd200_dx,34,0.046415888336127774,test,0.6,0.05947099337706104,0.5953065134099617,0.06066406774404732,0.5965250965250966,0.06074520327848402
+flat_mae,patch,logistic,adhd200_dx,35,0.005994842503189409,train,0.7232876712328767,0.022077181524813458,0.7078125371525494,0.024220170236057145,0.7053489650119069,0.023256156645309
+flat_mae,patch,logistic,adhd200_dx,35,0.005994842503189409,test,0.5846153846153846,0.05892026504638361,0.5578231292517006,0.06423842231664416,0.5612934362934363,0.06053049391752815
+flat_mae,patch,logistic,adhd200_dx,36,0.000774263682681127,train,0.6520547945205479,0.023169790284662508,0.6295304695304695,0.02543052246104817,0.6307626549429077,0.02403092433972
+flat_mae,patch,logistic,adhd200_dx,36,0.000774263682681127,test,0.5230769230769231,0.05727286293177918,0.49987589972697943,0.059762560808738124,0.502895752895753,0.05766843127488071
+flat_mae,patch,logistic,adhd200_dx,37,0.005994842503189409,train,0.7205479452054795,0.022443698495942428,0.7045238095238096,0.024447890864901657,0.7022043109238566,0.023452631688816224
+flat_mae,patch,logistic,adhd200_dx,37,0.005994842503189409,test,0.6615384615384615,0.05682256406553869,0.6425000000000001,0.0619619253833288,0.6418918918918919,0.05898029206476128
+flat_mae,patch,logistic,adhd200_dx,38,0.005994842503189409,train,0.726027397260274,0.022915081595614787,0.7170542635658914,0.024071946250538167,0.7149508456982353,0.02378857968540285
+flat_mae,patch,logistic,adhd200_dx,38,0.005994842503189409,test,0.5384615384615384,0.057453132945097955,0.5192307692307693,0.05989794466165466,0.5207528957528957,0.05786402876580633
+flat_mae,patch,logistic,adhd200_dx,39,0.005994842503189409,train,0.726027397260274,0.02197036096923327,0.7110878926038501,0.02402823161951898,0.7084936190999572,0.023113348312872126
+flat_mae,patch,logistic,adhd200_dx,39,0.005994842503189409,test,0.6615384615384615,0.0560774696607318,0.6366869918699187,0.061839426323924813,0.6375482625482626,0.05808815409906013
+flat_mae,patch,logistic,adhd200_dx,40,0.005994842503189409,train,0.7205479452054795,0.02220244965935824,0.7068134568134568,0.023900226148545648,0.7043567197899493,0.023197602316128895
+flat_mae,patch,logistic,adhd200_dx,40,0.005994842503189409,test,0.6307692307692307,0.05205004419394585,0.587737843551797,0.06211088417396741,0.5974903474903475,0.054455122466531274
+flat_mae,patch,logistic,adhd200_dx,41,0.005994842503189409,train,0.7041095890410959,0.02182693490296788,0.687974924012158,0.023838853673666244,0.6862062648836783,0.02291902835682879
+flat_mae,patch,logistic,adhd200_dx,41,0.005994842503189409,test,0.5538461538461539,0.06111759932086307,0.5469838981014179,0.06205525189766619,0.5472972972972974,0.06190522647687008
+flat_mae,patch,logistic,adhd200_dx,42,0.005994842503189409,train,0.7342465753424657,0.023362355022582017,0.7222026065328092,0.02487370188479026,0.71936252060817,0.024262331763589464
+flat_mae,patch,logistic,adhd200_dx,42,0.005994842503189409,test,0.5384615384615384,0.06275487015212963,0.5330459770114943,0.06314375194502904,0.5337837837837838,0.06324572364060993
+flat_mae,patch,logistic,adhd200_dx,43,0.046415888336127774,train,0.8328767123287671,0.019165280267290002,0.8264952269627899,0.020429761572947177,0.8218080234475178,0.020498523821229027
+flat_mae,patch,logistic,adhd200_dx,43,0.046415888336127774,test,0.6,0.05912898347164893,0.5921814671814671,0.06041647484088127,0.5921814671814671,0.06011477287930213
+flat_mae,patch,logistic,adhd200_dx,44,0.046415888336127774,train,0.8054794520547945,0.020761267110831375,0.7993092179138691,0.021871281528614218,0.7961012395432618,0.02195356415166546
+flat_mae,patch,logistic,adhd200_dx,44,0.046415888336127774,test,0.6,0.057710840100355126,0.5775,0.06206447144136149,0.5791505791505791,0.05918403745742286
+flat_mae,patch,logistic,adhd200_dx,45,0.005994842503189409,train,0.7561643835616438,0.02229662970054,0.7444962679230146,0.024160604216939094,0.740932405202418,0.02359544868750415
+flat_mae,patch,logistic,adhd200_dx,45,0.005994842503189409,test,0.5076923076923077,0.05994455031465357,0.4871794871794872,0.061936053646823204,0.48938223938223935,0.06010449539992708
+flat_mae,patch,logistic,adhd200_dx,46,0.000774263682681127,train,0.6383561643835617,0.02145938117340014,0.6000199242877067,0.02552124954726054,0.6085821579043781,0.02253711183856456
+flat_mae,patch,logistic,adhd200_dx,46,0.000774263682681127,test,0.6615384615384615,0.05471973563035206,0.6299171842650104,0.06391608106189291,0.6332046332046332,0.05748081521955596
+flat_mae,patch,logistic,adhd200_dx,47,0.005994842503189409,train,0.7589041095890411,0.021763169550492133,0.747674710910005,0.023390517237837544,0.7440770592904683,0.022835243895963475
+flat_mae,patch,logistic,adhd200_dx,47,0.005994842503189409,test,0.5538461538461539,0.0607249225031844,0.5381034060279344,0.06240298791904154,0.5386100386100386,0.06119743641907058
+flat_mae,patch,logistic,adhd200_dx,48,0.3593813663804626,train,0.9561643835616438,0.01088831512528878,0.9552202183781131,0.011179413235158527,0.9532728827013495,0.01165244919462919
+flat_mae,patch,logistic,adhd200_dx,48,0.3593813663804626,test,0.6307692307692307,0.06416737728544239,0.6299810246679317,0.06424466597952602,0.6365830115830116,0.06423716187867733
+flat_mae,patch,logistic,adhd200_dx,49,0.005994842503189409,train,0.7397260273972602,0.020729336352759125,0.7244275961821202,0.02276909768838598,0.7213470110520852,0.021807210011172997
+flat_mae,patch,logistic,adhd200_dx,49,0.005994842503189409,test,0.5076923076923077,0.06062066357333121,0.48,0.0632875825966212,0.48503861003861004,0.061067542538412366
+flat_mae,patch,logistic,adhd200_dx,50,0.005994842503189409,train,0.7397260273972602,0.02222856143605805,0.7265917602996255,0.024070064097153746,0.7234994199181779,0.023332671951150376
+flat_mae,patch,logistic,adhd200_dx,50,0.005994842503189409,test,0.6307692307692307,0.05310687226392461,0.5962732919254659,0.06153319367059387,0.6018339768339769,0.05562395056913396
+flat_mae,patch,logistic,adhd200_dx,51,0.005994842503189409,train,0.7315068493150685,0.02164172627141272,0.7196708463949844,0.022973817232989892,0.7169353361421505,0.022415529048109274
+flat_mae,patch,logistic,adhd200_dx,51,0.005994842503189409,test,0.5538461538461539,0.062097849870827906,0.5469838981014179,0.0631569322999681,0.5472972972972974,0.06323941349555502
+flat_mae,patch,logistic,adhd200_dx,52,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adhd200_dx,52,166.81005372000556,test,0.5846153846153846,0.05991440837862748,0.5842217484008528,0.0600633172409386,0.5916988416988418,0.06048731269124607
+flat_mae,patch,logistic,adhd200_dx,53,0.005994842503189409,train,0.7123287671232876,0.022800594554276937,0.6992914812984017,0.02445117657662805,0.697075166391891,0.023792587411843282
+flat_mae,patch,logistic,adhd200_dx,53,0.005994842503189409,test,0.6,0.05603781478957341,0.570630081300813,0.06217061649701836,0.5748069498069498,0.05791292411609699
+flat_mae,patch,logistic,adhd200_dx,54,0.005994842503189409,train,0.7424657534246575,0.02150399143388654,0.7329212853406402,0.022833648803330095,0.7302314221163827,0.02248525324468077
+flat_mae,patch,logistic,adhd200_dx,54,0.005994842503189409,test,0.5846153846153846,0.06114006424286208,0.5699583435432491,0.06361913080694286,0.5699806949806949,0.062218178015642917
+flat_mae,patch,logistic,adhd200_dx,55,0.005994842503189409,train,0.7287671232876712,0.02189059337794961,0.7184102863822326,0.023214704130838114,0.7159430909201929,0.02282367625310187
+flat_mae,patch,logistic,adhd200_dx,55,0.005994842503189409,test,0.5538461538461539,0.0615722983897311,0.5500119360229172,0.062283864775608454,0.5516409266409266,0.06254055242144516
+flat_mae,patch,logistic,adhd200_dx,56,0.046415888336127774,train,0.8273972602739726,0.018062338623838468,0.8226036644165864,0.018719882960673097,0.8198235330036027,0.018794635786012343
+flat_mae,patch,logistic,adhd200_dx,56,0.046415888336127774,test,0.6307692307692307,0.059886282769139425,0.6235521235521235,0.06123199746900422,0.6235521235521235,0.06081287917056077
+flat_mae,patch,logistic,adhd200_dx,57,0.046415888336127774,train,0.8273972602739726,0.019640462062968714,0.8219222637827288,0.020512098878076365,0.8183885937595408,0.020519810816742468
+flat_mae,patch,logistic,adhd200_dx,57,0.046415888336127774,test,0.47692307692307695,0.061380250471377136,0.4666988416988417,0.06210337530946148,0.4666988416988417,0.061938623429712535
+flat_mae,patch,logistic,adhd200_dx,58,10000.0,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adhd200_dx,58,10000.0,test,0.5846153846153846,0.06182958829044295,0.5810455956075435,0.06238142369611767,0.583011583011583,0.06273800949872903
+flat_mae,patch,logistic,adhd200_dx,59,0.005994842503189409,train,0.736986301369863,0.022914332122625974,0.726027397260274,0.024133596287808083,0.7232246443182512,0.023646910849835705
+flat_mae,patch,logistic,adhd200_dx,59,0.005994842503189409,test,0.5846153846153846,0.05534578981010337,0.5411764705882354,0.06181641099500292,0.5526061776061776,0.056166560066473796
+flat_mae,patch,logistic,adhd200_dx,60,0.005994842503189409,train,0.7397260273972602,0.022836827835171818,0.729787648548607,0.02391043781111711,0.7270867680283324,0.02346507037206614
+flat_mae,patch,logistic,adhd200_dx,60,0.005994842503189409,test,0.6615384615384615,0.056132252961066935,0.6425000000000001,0.06006833363648928,0.6418918918918919,0.05772516890499275
+flat_mae,patch,logistic,adhd200_dx,61,0.005994842503189409,train,0.7041095890410959,0.023303740467139702,0.6917808219178082,0.024649862280958786,0.6897936129938328,0.024067836688138063
+flat_mae,patch,logistic,adhd200_dx,61,0.005994842503189409,test,0.6,0.061166096499177924,0.5833333333333333,0.06477892257281674,0.5834942084942085,0.06280781966608415
+flat_mae,patch,logistic,adhd200_dx,62,0.005994842503189409,train,0.7150684931506849,0.02295458749102437,0.705736434108527,0.024052135037160877,0.7038071685900958,0.02377186536689152
+flat_mae,patch,logistic,adhd200_dx,62,0.005994842503189409,test,0.5692307692307692,0.057135230429333335,0.5512820512820513,0.06024928311266914,0.5521235521235521,0.05848726018596496
+flat_mae,patch,logistic,adhd200_dx,63,0.005994842503189409,train,0.7287671232876712,0.023559667964300826,0.7177857092650011,0.02486981895731557,0.715225621298162,0.02437249304770905
+flat_mae,patch,logistic,adhd200_dx,63,0.005994842503189409,test,0.6,0.054644793815369644,0.570630081300813,0.05967058236001123,0.5748069498069498,0.056116287203535634
+flat_mae,patch,logistic,adhd200_dx,64,0.005994842503189409,train,0.7178082191780822,0.022598665303170146,0.7028086734492217,0.02461694271285539,0.7004945960798681,0.02373507999284538
+flat_mae,patch,logistic,adhd200_dx,64,0.005994842503189409,test,0.6307692307692307,0.05592199977800864,0.6036585365853658,0.06227386364878947,0.6061776061776062,0.05792586044684726
+flat_mae,patch,logistic,adhd200_dx,65,0.005994842503189409,train,0.7205479452054795,0.023266382533645276,0.7045238095238096,0.025512710667900215,0.7022043109238566,0.024469664743134594
+flat_mae,patch,logistic,adhd200_dx,65,0.005994842503189409,test,0.676923076923077,0.05373611613723858,0.6612062546537603,0.05776392214398613,0.6597490347490347,0.055688565931782875
+flat_mae,patch,logistic,adhd200_dx,66,0.005994842503189409,train,0.7424657534246575,0.022769794281447535,0.7317351598173516,0.02415137953584078,0.7287964828723209,0.023651629008550848
+flat_mae,patch,logistic,adhd200_dx,66,0.005994842503189409,test,0.6153846153846154,0.05826989641660921,0.5966741126830479,0.06189744657713271,0.597007722007722,0.059638607200298205
+flat_mae,patch,logistic,adhd200_dx,67,0.046415888336127774,train,0.8191780821917808,0.020216314902893544,0.8132558139534884,0.021275944961961703,0.8096721011174208,0.021303597716987244
+flat_mae,patch,logistic,adhd200_dx,67,0.046415888336127774,test,0.6,0.0547095612729124,0.5626293995859213,0.06242501864664253,0.5704633204633205,0.056400707961241414
+flat_mae,patch,logistic,adhd200_dx,68,0.3593813663804626,train,0.9561643835616438,0.010683345157969416,0.9552915237628614,0.010946145806773175,0.9539903523233804,0.011364740577253448
+flat_mae,patch,logistic,adhd200_dx,68,0.3593813663804626,test,0.5846153846153846,0.05780803981675578,0.5644080416976918,0.06163621427551431,0.5656370656370656,0.05918942817174904
+flat_mae,patch,logistic,adhd200_dx,69,0.005994842503189409,train,0.7342465753424657,0.022627827670821474,0.7246910988250481,0.023923773200905718,0.7222323990962936,0.023594600749455144
+flat_mae,patch,logistic,adhd200_dx,69,0.005994842503189409,test,0.4461538461538462,0.06193028338381796,0.4230769230769231,0.06270234320564441,0.4266409266409266,0.0616888681227142
+flat_mae,patch,logistic,adhd200_dx,70,0.005994842503189409,train,0.7232876712328767,0.022441283031214385,0.7093238714764439,0.02428303004496929,0.7067839042559687,0.02350081520282437
+flat_mae,patch,logistic,adhd200_dx,70,0.005994842503189409,test,0.5230769230769231,0.06104710762282211,0.5062484685126194,0.0635014107595719,0.5072393822393823,0.06190383000831069
+flat_mae,patch,logistic,adhd200_dx,71,0.005994842503189409,train,0.7232876712328767,0.022100115775325526,0.7107470439156054,0.023726550328457937,0.7082188435000305,0.023076836256939837
+flat_mae,patch,logistic,adhd200_dx,71,0.005994842503189409,test,0.5692307692307692,0.05659594147742377,0.5376016260162602,0.062044019075676825,0.5434362934362934,0.057753081923810734
+flat_mae,patch,logistic,adhd200_dx,72,0.005994842503189409,train,0.7232876712328767,0.02307402400100976,0.7107470439156054,0.024822190995863188,0.7082188435000305,0.024160571553970608
+flat_mae,patch,logistic,adhd200_dx,72,0.005994842503189409,test,0.6307692307692307,0.05609029913277994,0.6036585365853658,0.06115136061865767,0.6061776061776062,0.057341517022094335
+flat_mae,patch,logistic,adhd200_dx,73,0.000774263682681127,train,0.6575342465753424,0.02307242279353255,0.6332071163848894,0.025697876621656443,0.6348995542529157,0.02398940136684095
+flat_mae,patch,logistic,adhd200_dx,73,0.000774263682681127,test,0.5692307692307692,0.060231223104685414,0.545,0.06383778070712375,0.5477799227799228,0.061274607351757174
+flat_mae,patch,logistic,adhd200_dx,74,0.005994842503189409,train,0.7123287671232876,0.021518120072669285,0.6992914812984017,0.02307937963393897,0.697075166391891,0.022467433233319872
+flat_mae,patch,logistic,adhd200_dx,74,0.005994842503189409,test,0.6,0.06086725492326306,0.588206627680312,0.06293773736882363,0.5878378378378378,0.06173305166009477
+flat_mae,patch,logistic,adhd200_dx,75,0.000774263682681127,train,0.6493150684931507,0.024211490752969772,0.6271467790333909,0.026410490894806075,0.6283354704768883,0.025028103942272818
+flat_mae,patch,logistic,adhd200_dx,75,0.000774263682681127,test,0.5692307692307692,0.06144790259828085,0.5512820512820513,0.06439173788759277,0.5521235521235521,0.06257055113749697
+flat_mae,patch,logistic,adhd200_dx,76,0.005994842503189409,train,0.736986301369863,0.021762281072142658,0.722644376899696,0.02369644272299936,0.7196372962080967,0.022865534010131413
+flat_mae,patch,logistic,adhd200_dx,76,0.005994842503189409,test,0.5846153846153846,0.05334802558970225,0.5308740978348035,0.06315106954948436,0.5482625482625483,0.05494262135581445
+flat_mae,patch,logistic,adhd200_dx,77,0.005994842503189409,train,0.7287671232876712,0.02124986611506177,0.7143500841890578,0.02297012302648463,0.7116382731880075,0.022119485509107886
+flat_mae,patch,logistic,adhd200_dx,77,0.005994842503189409,test,0.6923076923076923,0.04830603622835012,0.6564482029598309,0.059567725263369425,0.6602316602316602,0.05192352376469467
+flat_mae,patch,logistic,adhd200_dx,78,0.046415888336127774,train,0.8136986301369863,0.020500328389188124,0.8079779662065978,0.021334896646349515,0.8048177321853819,0.021261613119241796
+flat_mae,patch,logistic,adhd200_dx,78,0.046415888336127774,test,0.5076923076923077,0.060007691814658856,0.4871794871794872,0.06246715479628248,0.48938223938223935,0.06073682705315684
+flat_mae,patch,logistic,adhd200_dx,79,0.046415888336127774,train,0.8356164383561644,0.01839559943135128,0.8302325581395349,0.019292304084109084,0.82638761677963,0.019337215031205886
+flat_mae,patch,logistic,adhd200_dx,79,0.046415888336127774,test,0.5692307692307692,0.06071768797587812,0.5608108108108107,0.06181523711408662,0.5608108108108107,0.06143627732299396
+flat_mae,patch,logistic,adhd200_dx,80,0.005994842503189409,train,0.736986301369863,0.021911381934621478,0.726027397260274,0.02330160014239652,0.7232246443182512,0.02288122930677169
+flat_mae,patch,logistic,adhd200_dx,80,0.005994842503189409,test,0.5692307692307692,0.062282243615837125,0.5512820512820513,0.06525275785157907,0.5521235521235521,0.06337668133097127
+flat_mae,patch,logistic,adhd200_dx,81,0.005994842503189409,train,0.7452054794520548,0.022841902151556792,0.7354763296317943,0.024188971345863692,0.7326586065824021,0.02382873835766544
+flat_mae,patch,logistic,adhd200_dx,81,0.005994842503189409,test,0.5076923076923077,0.05577023871768183,0.4715447154471545,0.05882071122044139,0.4806949806949807,0.05586864545332324
+flat_mae,patch,logistic,adhd200_dx,82,0.005994842503189409,train,0.7150684931506849,0.022610110213810392,0.7003094801995832,0.024566268007424483,0.6980674116138487,0.02368182611539774
+flat_mae,patch,logistic,adhd200_dx,82,0.005994842503189409,test,0.6923076923076923,0.05898531358681508,0.6886973180076628,0.059702359059180116,0.6906370656370657,0.05973689950897015
+flat_mae,patch,logistic,adhd200_dx,83,0.005994842503189409,train,0.7315068493150685,0.021311968644545688,0.7190013826043238,0.02263767872958227,0.7162178665201197,0.022025541376126653
+flat_mae,patch,logistic,adhd200_dx,83,0.005994842503189409,test,0.6307692307692307,0.058050180292163286,0.6198830409356726,0.06017459518282782,0.6192084942084942,0.05946718303394383
+flat_mae,patch,logistic,adhd200_dx,84,0.005994842503189409,train,0.7452054794520548,0.020639719072161434,0.732347723240686,0.022134426808788587,0.7290712584722476,0.02146654898469891
+flat_mae,patch,logistic,adhd200_dx,84,0.005994842503189409,test,0.5538461538461539,0.06545242456667899,0.5500119360229172,0.06612552498382575,0.5516409266409266,0.06640017639902036
+flat_mae,patch,logistic,adhd200_dx,85,0.005994842503189409,train,0.7123287671232876,0.022778800559420973,0.6978119455943228,0.024617930175374354,0.6956402271478292,0.02379168819421414
+flat_mae,patch,logistic,adhd200_dx,85,0.005994842503189409,test,0.6,0.058201039052112824,0.5833333333333333,0.06141692894530669,0.5834942084942085,0.05955812962727085
+flat_mae,patch,logistic,adhd200_dx,86,0.3593813663804626,train,0.9643835616438357,0.009436119185371733,0.9635865519649449,0.00970654487022802,0.9612719057214386,0.010276125068230193
+flat_mae,patch,logistic,adhd200_dx,86,0.3593813663804626,test,0.5692307692307692,0.06266370465953001,0.5608108108108107,0.0638736425392196,0.5608108108108107,0.06345222216046117
+flat_mae,patch,logistic,adhd200_dx,87,0.3593813663804626,train,0.9616438356164384,0.009990223781983642,0.960817691080849,0.010257933471132514,0.9588447212554192,0.010741585299082484
+flat_mae,patch,logistic,adhd200_dx,87,0.3593813663804626,test,0.5692307692307692,0.06240130479594706,0.5565302144249512,0.06437824570753099,0.5564671814671815,0.06360930550576196
+flat_mae,patch,logistic,adhd200_dx,88,0.000774263682681127,train,0.6438356164383562,0.02389033241158383,0.6191080143847932,0.026266485726360658,0.6213286926787568,0.024641759416660892
+flat_mae,patch,logistic,adhd200_dx,88,0.000774263682681127,test,0.5692307692307692,0.059721634350270884,0.545,0.06365544719042884,0.5477799227799228,0.0608539734283045
+flat_mae,patch,logistic,adhd200_dx,89,0.005994842503189409,train,0.7178082191780822,0.023099302629036932,0.7063831116595465,0.02454637526937608,0.7040819441900226,0.024067561676940522
+flat_mae,patch,logistic,adhd200_dx,89,0.005994842503189409,test,0.6307692307692307,0.055941431756634984,0.61,0.060568572762768585,0.6105212355212355,0.05787304586244429
+flat_mae,patch,logistic,adhd200_dx,90,0.005994842503189409,train,0.7150684931506849,0.022625052332687735,0.6995314083080041,0.02456208746660161,0.6973499419918178,0.023608987808412706
+flat_mae,patch,logistic,adhd200_dx,90,0.005994842503189409,test,0.5846153846153846,0.06087096450300937,0.5699583435432491,0.0636951798511443,0.5699806949806949,0.06205491313918357
+flat_mae,patch,logistic,adhd200_dx,91,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adhd200_dx,91,2.782559402207126,test,0.5846153846153846,0.06010026336718279,0.5699583435432491,0.062378846189696247,0.5699806949806949,0.06085861925311064
+flat_mae,patch,logistic,adhd200_dx,92,0.046415888336127774,train,0.8328767123287671,0.019084247153022194,0.8272276497838891,0.020036944980048273,0.8232429626915796,0.02009860975432427
+flat_mae,patch,logistic,adhd200_dx,92,0.046415888336127774,test,0.5692307692307692,0.06146311541252215,0.545,0.0649635730538412,0.5477799227799228,0.06227455677545912
+flat_mae,patch,logistic,adhd200_dx,93,0.005994842503189409,train,0.7287671232876712,0.022431162812377923,0.7157879834199825,0.023833123163598657,0.7130732124320693,0.023163644158809425
+flat_mae,patch,logistic,adhd200_dx,93,0.005994842503189409,test,0.6615384615384615,0.0484503039480943,0.622093023255814,0.05827237282747697,0.6288610038610039,0.05124235322205058
+flat_mae,patch,logistic,adhd200_dx,94,0.005994842503189409,train,0.7506849315068493,0.021484262810442502,0.7399979647590196,0.022820196617862182,0.73679550589241,0.022382206926051834
+flat_mae,patch,logistic,adhd200_dx,94,0.005994842503189409,test,0.5846153846153846,0.05223686607669715,0.5308740978348035,0.06316885300951601,0.5482625482625483,0.05419932100442337
+flat_mae,patch,logistic,adhd200_dx,95,0.005994842503189409,train,0.726027397260274,0.023790210254629258,0.7176146561861731,0.024915563232096937,0.7156683153202662,0.024708669055984844
+flat_mae,patch,logistic,adhd200_dx,95,0.005994842503189409,test,0.5538461538461539,0.0584713433073474,0.5381034060279344,0.0615082809429967,0.5386100386100386,0.06012061507801641
+flat_mae,patch,logistic,adhd200_dx,96,0.000774263682681127,train,0.6438356164383562,0.023349988497744733,0.6191080143847932,0.02573508257713248,0.6213286926787568,0.024120585609006095
+flat_mae,patch,logistic,adhd200_dx,96,0.000774263682681127,test,0.5692307692307692,0.06177741300343071,0.545,0.06570498395186582,0.5477799227799228,0.0629116938929283
+flat_mae,patch,logistic,adhd200_dx,97,0.046415888336127774,train,0.8191780821917808,0.020801544466815838,0.812872856077554,0.02189961753690149,0.8089546314953899,0.021857929605452256
+flat_mae,patch,logistic,adhd200_dx,97,0.046415888336127774,test,0.5692307692307692,0.059707174992308334,0.545,0.06421285537688758,0.5477799227799228,0.061117784635458534
+flat_mae,patch,logistic,adhd200_dx,98,0.005994842503189409,train,0.7178082191780822,0.022158782162991136,0.7020266467991345,0.024032354101946648,0.6997771264578372,0.023049639705691287
+flat_mae,patch,logistic,adhd200_dx,98,0.005994842503189409,test,0.676923076923077,0.058956158618529875,0.6655231560891939,0.06125211219336189,0.6640926640926641,0.05999044654691344
+flat_mae,patch,logistic,adhd200_dx,99,0.005994842503189409,train,0.7452054794520548,0.02399249272714593,0.7354763296317943,0.02523875162762978,0.7326586065824021,0.02481433006707826
+flat_mae,patch,logistic,adhd200_dx,99,0.005994842503189409,test,0.5538461538461539,0.059182109882780574,0.5321419707123356,0.061894703318438216,0.5342664092664092,0.060017085576320815
+flat_mae,patch,logistic,adhd200_dx,100,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adhd200_dx,100,2.782559402207126,test,0.6,0.06010618611646771,0.5976190476190476,0.060400606376468,0.6008687258687259,0.060382175832968454

decoders/crossreg_reg4/eval_v2/adhd200_dx__patch__logistic/log.txt

@@ -0,0 +1,241 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: aef99c83a386cf95c3d8ca503ecc968d8d5694af, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-03-07 21:46:50
+config:
+output_root: experiments/decoders/output
+name_prefix: eval_logistic
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (adhd200_dx patch logistic)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: decoders/crossreg_reg4/eval_v2/adhd200_dx__patch__logistic
+model: flat_mae
+representation: patch
+dataset: adhd200_dx
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/adhd200_dx__patch__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=False, reg_tokens=4, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: adhd200_dx (flat)
+train (n=301):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'gender', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 301
+}),
+    labels=['ADHD' 'Control'],
+    counts=[131 170]
+)
+
+validation (n=64):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'gender', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 64
+}),
+    labels=['ADHD' 'Control'],
+    counts=[28 36]
+)
+
+test (n=65):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'gender', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 65
+}),
+    labels=['ADHD' 'Control'],
+    counts=[28 37]
+)
+
+extracting features for all splits
+extract (train)  [  0/151]  eta: 0:12:55    time: 5.1370  data: 4.0069  max mem: 2698
+extract (train)  [ 20/151]  eta: 0:00:57    time: 0.2012  data: 0.0728  max mem: 2852
+extract (train)  [ 40/151]  eta: 0:00:36    time: 0.2069  data: 0.0758  max mem: 2852
+extract (train)  [ 60/151]  eta: 0:00:25    time: 0.1998  data: 0.0713  max mem: 2852
+extract (train)  [ 80/151]  eta: 0:00:18    time: 0.1969  data: 0.0723  max mem: 2852
+extract (train)  [100/151]  eta: 0:00:12    time: 0.1763  data: 0.0595  max mem: 2852
+extract (train)  [120/151]  eta: 0:00:07    time: 0.1676  data: 0.0540  max mem: 2852
+extract (train)  [140/151]  eta: 0:00:02    time: 0.1740  data: 0.0579  max mem: 2852
+extract (train)  [150/151]  eta: 0:00:00    time: 0.1668  data: 0.0555  max mem: 2852
+extract (train) Total time: 0:00:33 (0.2224 s / it)
+extract (validation)  [ 0/32]  eta: 0:02:00    time: 3.7666  data: 3.6335  max mem: 2852
+extract (validation)  [20/32]  eta: 0:00:04    time: 0.1810  data: 0.0604  max mem: 2852
+extract (validation)  [31/32]  eta: 0:00:00    time: 0.1594  data: 0.0494  max mem: 2852
+extract (validation) Total time: 0:00:09 (0.2979 s / it)
+extract (test)  [ 0/33]  eta: 0:02:03    time: 3.7426  data: 3.6112  max mem: 2852
+extract (test)  [20/33]  eta: 0:00:04    time: 0.1942  data: 0.0679  max mem: 2852
+extract (test)  [32/33]  eta: 0:00:00    time: 0.1507  data: 0.0468  max mem: 2852
+extract (test) Total time: 0:00:09 (0.2954 s / it)
+feature extraction time: 0:00:52
+train features: (301, 768)
+validation features: (64, 768)
+test features: (65, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset    | trial   |         C | split   |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:-----------|:--------|----------:|:--------|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | adhd200_dx |         | 0.0059948 | train   | 0.74247 |  0.021905 | 0.73174 | 0.023336 | 0.7288  |   0.022882 |
+| flat_mae | patch  | logistic | adhd200_dx |         | 0.0059948 | test    | 0.58462 |  0.057475 | 0.55012 | 0.064977 | 0.55695 |   0.059498 |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 1, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.0595373961343754, "f1": 0.5699583435432491, "f1_std": 0.061553726399329114, "bacc": 0.5699806949806949, "bacc_std": 0.06028514680245595}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 2, "C": 0.3593813663804626, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.06186923411662306, "f1": 0.578226387887527, "f1_std": 0.06294971050534451, "bacc": 0.5786679536679536, "bacc_std": 0.06281953395438065}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 3, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.056654301006059615, "f1": 0.570630081300813, "f1_std": 0.0629693519564316, "bacc": 0.5748069498069498, "bacc_std": 0.05860606934204386}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 4, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05718021754161582, "f1": 0.6264367816091954, "f1_std": 0.05816225509061837, "bacc": 0.627895752895753, "bacc_std": 0.05866661248897376}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 5, "C": 0.046415888336127774, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06097359211093965, "f1": 0.5666666666666667, "f1_std": 0.06144937965341382, "bacc": 0.5694980694980695, "bacc_std": 0.06208679449276509}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 6, "C": 0.3593813663804626, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.059135956082362856, "f1": 0.6575670498084292, "f1_std": 0.059942574087972475, "bacc": 0.6592664092664093, "bacc_std": 0.06033058277089562}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 7, "C": 0.046415888336127774, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05917168681650083, "f1": 0.5830363506771205, "f1_std": 0.05919782582612036, "bacc": 0.5873552123552124, "bacc_std": 0.05994967511895883}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 8, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.06192692009096418, "f1": 0.6198830409356726, "f1_std": 0.06462276479873824, "bacc": 0.6192084942084942, "bacc_std": 0.06344098862392855}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 9, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.054095563041976026, "f1": 0.5962732919254659, "f1_std": 0.06156816960167578, "bacc": 0.6018339768339769, "bacc_std": 0.055989328273165724}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 10, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05909105600180356, "f1": 0.5578231292517006, "f1_std": 0.06407308875457003, "bacc": 0.5612934362934363, "bacc_std": 0.060402412127793625}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 11, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.05412110898595687, "f1": 0.5626293995859213, "f1_std": 0.062306311625395884, "bacc": 0.5704633204633205, "bacc_std": 0.05602142058640619}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 12, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.056831785422591874, "f1": 0.5699583435432491, "f1_std": 0.05977686042399589, "bacc": 0.5699806949806949, "bacc_std": 0.058273607132441826}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 13, "C": 0.005994842503189409, "split": "test", "acc": 0.49230769230769234, "acc_std": 0.0588737773274718, "f1": 0.4743935309973046, "f1_std": 0.060543653424304386, "bacc": 0.47586872586872586, "bacc_std": 0.059637690826888815}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 14, "C": 1291.5496650148827, "split": "test", "acc": 0.6, "acc_std": 0.057080124314125104, "f1": 0.5833333333333333, "f1_std": 0.06070381533140116, "bacc": 0.5834942084942085, "bacc_std": 0.058669222582869274}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 15, "C": 0.005994842503189409, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.061283927446873124, "f1": 0.5294401544401545, "f1_std": 0.0623149636534173, "bacc": 0.5294401544401545, "bacc_std": 0.06222626850432534}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 16, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.062481171601801194, "f1": 0.5745454545454545, "f1_std": 0.06394244463034299, "bacc": 0.5743243243243243, "bacc_std": 0.06344919493891837}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 17, "C": 0.046415888336127774, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.056206643251150754, "f1": 0.5966741126830479, "f1_std": 0.06061578789763127, "bacc": 0.597007722007722, "bacc_std": 0.05831458026482621}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 18, "C": 0.3593813663804626, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.05923596380818011, "f1": 0.5608108108108107, "f1_std": 0.06037803594609343, "bacc": 0.5608108108108107, "bacc_std": 0.060218470396470755}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 19, "C": 0.000774263682681127, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05238849867002689, "f1": 0.587737843551797, "f1_std": 0.0629310560687302, "bacc": 0.5974903474903475, "bacc_std": 0.055048676802358774}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 20, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05634593909121585, "f1": 0.6036585365853658, "f1_std": 0.06289556976806442, "bacc": 0.6061776061776062, "bacc_std": 0.058661890473261515}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 21, "C": 0.005994842503189409, "split": "test", "acc": 0.7230769230769231, "acc_std": 0.05417740746187419, "f1": 0.7149122807017544, "f1_std": 0.05636841728510411, "bacc": 0.7133204633204633, "bacc_std": 0.05579711314188372}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 22, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.06171500829231663, "f1": 0.6235521235521235, "f1_std": 0.06334497042852516, "bacc": 0.6235521235521235, "bacc_std": 0.06326613240265587}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 23, "C": 0.000774263682681127, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05174945558985287, "f1": 0.5656241646618552, "f1_std": 0.06406613287543482, "bacc": 0.5796332046332047, "bacc_std": 0.054493023528735494}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 24, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05588169244301624, "f1": 0.5501153550371699, "f1_std": 0.06169691487186124, "bacc": 0.556949806949807, "bacc_std": 0.05693022347504732}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 25, "C": 0.000774263682681127, "split": "test", "acc": 0.6, "acc_std": 0.05099025083752161, "f1": 0.5427489177489178, "f1_std": 0.06181938637330176, "bacc": 0.5617760617760618, "bacc_std": 0.05255712965330673}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 26, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05501583204085829, "f1": 0.6036585365853658, "f1_std": 0.061749572456848666, "bacc": 0.6061776061776062, "bacc_std": 0.05731640220304975}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 27, "C": 0.046415888336127774, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.060639308047448666, "f1": 0.6018132810585641, "f1_std": 0.06333183184019378, "bacc": 0.6013513513513513, "bacc_std": 0.06192042359549878}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 28, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.0546718148343194, "f1": 0.5501153550371699, "f1_std": 0.06075232455253754, "bacc": 0.556949806949807, "bacc_std": 0.0559985707062253}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 29, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05950202820868164, "f1": 0.5966741126830479, "f1_std": 0.06341831262949778, "bacc": 0.597007722007722, "bacc_std": 0.061315369217520205}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 30, "C": 0.3593813663804626, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06032451884569405, "f1": 0.564176245210728, "f1_std": 0.06047048488448265, "bacc": 0.5651544401544402, "bacc_std": 0.06061516498236156}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 31, "C": 0.3593813663804626, "split": "test", "acc": 0.6, "acc_std": 0.061585797179229435, "f1": 0.5921814671814671, "f1_std": 0.06290748518584183, "bacc": 0.5921814671814671, "bacc_std": 0.06264103536840496}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 32, "C": 0.046415888336127774, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06633448682255583, "f1": 0.543030303030303, "f1_std": 0.0677295747450472, "bacc": 0.542953667953668, "bacc_std": 0.06714117747970659}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 33, "C": 0.3593813663804626, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06428263037592144, "f1": 0.5521501544309813, "f1_std": 0.06465589680292005, "bacc": 0.555984555984556, "bacc_std": 0.06493353033693425}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 34, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.05947099337706104, "f1": 0.5953065134099617, "f1_std": 0.06066406774404732, "bacc": 0.5965250965250966, "bacc_std": 0.06074520327848402}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 35, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05892026504638361, "f1": 0.5578231292517006, "f1_std": 0.06423842231664416, "bacc": 0.5612934362934363, "bacc_std": 0.06053049391752815}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 36, "C": 0.000774263682681127, "split": "test", "acc": 0.5230769230769231, "acc_std": 0.05727286293177918, "f1": 0.49987589972697943, "f1_std": 0.059762560808738124, "bacc": 0.502895752895753, "bacc_std": 0.05766843127488071}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 37, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.05682256406553869, "f1": 0.6425000000000001, "f1_std": 0.0619619253833288, "bacc": 0.6418918918918919, "bacc_std": 0.05898029206476128}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 38, "C": 0.005994842503189409, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.057453132945097955, "f1": 0.5192307692307693, "f1_std": 0.05989794466165466, "bacc": 0.5207528957528957, "bacc_std": 0.05786402876580633}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 39, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.0560774696607318, "f1": 0.6366869918699187, "f1_std": 0.061839426323924813, "bacc": 0.6375482625482626, "bacc_std": 0.05808815409906013}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 40, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05205004419394585, "f1": 0.587737843551797, "f1_std": 0.06211088417396741, "bacc": 0.5974903474903475, "bacc_std": 0.054455122466531274}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 41, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06111759932086307, "f1": 0.5469838981014179, "f1_std": 0.06205525189766619, "bacc": 0.5472972972972974, "bacc_std": 0.06190522647687008}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 42, "C": 0.005994842503189409, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.06275487015212963, "f1": 0.5330459770114943, "f1_std": 0.06314375194502904, "bacc": 0.5337837837837838, "bacc_std": 0.06324572364060993}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 43, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.05912898347164893, "f1": 0.5921814671814671, "f1_std": 0.06041647484088127, "bacc": 0.5921814671814671, "bacc_std": 0.06011477287930213}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 44, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.057710840100355126, "f1": 0.5775, "f1_std": 0.06206447144136149, "bacc": 0.5791505791505791, "bacc_std": 0.05918403745742286}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 45, "C": 0.005994842503189409, "split": "test", "acc": 0.5076923076923077, "acc_std": 0.05994455031465357, "f1": 0.4871794871794872, "f1_std": 0.061936053646823204, "bacc": 0.48938223938223935, "bacc_std": 0.06010449539992708}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 46, "C": 0.000774263682681127, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.05471973563035206, "f1": 0.6299171842650104, "f1_std": 0.06391608106189291, "bacc": 0.6332046332046332, "bacc_std": 0.05748081521955596}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 47, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.0607249225031844, "f1": 0.5381034060279344, "f1_std": 0.06240298791904154, "bacc": 0.5386100386100386, "bacc_std": 0.06119743641907058}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 48, "C": 0.3593813663804626, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.06416737728544239, "f1": 0.6299810246679317, "f1_std": 0.06424466597952602, "bacc": 0.6365830115830116, "bacc_std": 0.06423716187867733}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 49, "C": 0.005994842503189409, "split": "test", "acc": 0.5076923076923077, "acc_std": 0.06062066357333121, "f1": 0.48, "f1_std": 0.0632875825966212, "bacc": 0.48503861003861004, "bacc_std": 0.061067542538412366}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 50, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05310687226392461, "f1": 0.5962732919254659, "f1_std": 0.06153319367059387, "bacc": 0.6018339768339769, "bacc_std": 0.05562395056913396}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 51, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.062097849870827906, "f1": 0.5469838981014179, "f1_std": 0.0631569322999681, "bacc": 0.5472972972972974, "bacc_std": 0.06323941349555502}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 52, "C": 166.81005372000556, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05991440837862748, "f1": 0.5842217484008528, "f1_std": 0.0600633172409386, "bacc": 0.5916988416988418, "bacc_std": 0.06048731269124607}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 53, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.05603781478957341, "f1": 0.570630081300813, "f1_std": 0.06217061649701836, "bacc": 0.5748069498069498, "bacc_std": 0.05791292411609699}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 54, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.06114006424286208, "f1": 0.5699583435432491, "f1_std": 0.06361913080694286, "bacc": 0.5699806949806949, "bacc_std": 0.062218178015642917}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 55, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.0615722983897311, "f1": 0.5500119360229172, "f1_std": 0.062283864775608454, "bacc": 0.5516409266409266, "bacc_std": 0.06254055242144516}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 56, "C": 0.046415888336127774, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.059886282769139425, "f1": 0.6235521235521235, "f1_std": 0.06123199746900422, "bacc": 0.6235521235521235, "bacc_std": 0.06081287917056077}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 57, "C": 0.046415888336127774, "split": "test", "acc": 0.47692307692307695, "acc_std": 0.061380250471377136, "f1": 0.4666988416988417, "f1_std": 0.06210337530946148, "bacc": 0.4666988416988417, "bacc_std": 0.061938623429712535}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 58, "C": 10000.0, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.06182958829044295, "f1": 0.5810455956075435, "f1_std": 0.06238142369611767, "bacc": 0.583011583011583, "bacc_std": 0.06273800949872903}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 59, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05534578981010337, "f1": 0.5411764705882354, "f1_std": 0.06181641099500292, "bacc": 0.5526061776061776, "bacc_std": 0.056166560066473796}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 60, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.056132252961066935, "f1": 0.6425000000000001, "f1_std": 0.06006833363648928, "bacc": 0.6418918918918919, "bacc_std": 0.05772516890499275}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 61, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.061166096499177924, "f1": 0.5833333333333333, "f1_std": 0.06477892257281674, "bacc": 0.5834942084942085, "bacc_std": 0.06280781966608415}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 62, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.057135230429333335, "f1": 0.5512820512820513, "f1_std": 0.06024928311266914, "bacc": 0.5521235521235521, "bacc_std": 0.05848726018596496}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 63, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.054644793815369644, "f1": 0.570630081300813, "f1_std": 0.05967058236001123, "bacc": 0.5748069498069498, "bacc_std": 0.056116287203535634}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 64, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05592199977800864, "f1": 0.6036585365853658, "f1_std": 0.06227386364878947, "bacc": 0.6061776061776062, "bacc_std": 0.05792586044684726}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 65, "C": 0.005994842503189409, "split": "test", "acc": 0.676923076923077, "acc_std": 0.05373611613723858, "f1": 0.6612062546537603, "f1_std": 0.05776392214398613, "bacc": 0.6597490347490347, "bacc_std": 0.055688565931782875}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 66, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05826989641660921, "f1": 0.5966741126830479, "f1_std": 0.06189744657713271, "bacc": 0.597007722007722, "bacc_std": 0.059638607200298205}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 67, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.0547095612729124, "f1": 0.5626293995859213, "f1_std": 0.06242501864664253, "bacc": 0.5704633204633205, "bacc_std": 0.056400707961241414}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 68, "C": 0.3593813663804626, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05780803981675578, "f1": 0.5644080416976918, "f1_std": 0.06163621427551431, "bacc": 0.5656370656370656, "bacc_std": 0.05918942817174904}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 69, "C": 0.005994842503189409, "split": "test", "acc": 0.4461538461538462, "acc_std": 0.06193028338381796, "f1": 0.4230769230769231, "f1_std": 0.06270234320564441, "bacc": 0.4266409266409266, "bacc_std": 0.0616888681227142}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 70, "C": 0.005994842503189409, "split": "test", "acc": 0.5230769230769231, "acc_std": 0.06104710762282211, "f1": 0.5062484685126194, "f1_std": 0.0635014107595719, "bacc": 0.5072393822393823, "bacc_std": 0.06190383000831069}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 71, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.05659594147742377, "f1": 0.5376016260162602, "f1_std": 0.062044019075676825, "bacc": 0.5434362934362934, "bacc_std": 0.057753081923810734}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 72, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05609029913277994, "f1": 0.6036585365853658, "f1_std": 0.06115136061865767, "bacc": 0.6061776061776062, "bacc_std": 0.057341517022094335}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 73, "C": 0.000774263682681127, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.060231223104685414, "f1": 0.545, "f1_std": 0.06383778070712375, "bacc": 0.5477799227799228, "bacc_std": 0.061274607351757174}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 74, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.06086725492326306, "f1": 0.588206627680312, "f1_std": 0.06293773736882363, "bacc": 0.5878378378378378, "bacc_std": 0.06173305166009477}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 75, "C": 0.000774263682681127, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06144790259828085, "f1": 0.5512820512820513, "f1_std": 0.06439173788759277, "bacc": 0.5521235521235521, "bacc_std": 0.06257055113749697}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 76, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05334802558970225, "f1": 0.5308740978348035, "f1_std": 0.06315106954948436, "bacc": 0.5482625482625483, "bacc_std": 0.05494262135581445}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 77, "C": 0.005994842503189409, "split": "test", "acc": 0.6923076923076923, "acc_std": 0.04830603622835012, "f1": 0.6564482029598309, "f1_std": 0.059567725263369425, "bacc": 0.6602316602316602, "bacc_std": 0.05192352376469467}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 78, "C": 0.046415888336127774, "split": "test", "acc": 0.5076923076923077, "acc_std": 0.060007691814658856, "f1": 0.4871794871794872, "f1_std": 0.06246715479628248, "bacc": 0.48938223938223935, "bacc_std": 0.06073682705315684}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 79, "C": 0.046415888336127774, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06071768797587812, "f1": 0.5608108108108107, "f1_std": 0.06181523711408662, "bacc": 0.5608108108108107, "bacc_std": 0.06143627732299396}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 80, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.062282243615837125, "f1": 0.5512820512820513, "f1_std": 0.06525275785157907, "bacc": 0.5521235521235521, "bacc_std": 0.06337668133097127}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 81, "C": 0.005994842503189409, "split": "test", "acc": 0.5076923076923077, "acc_std": 0.05577023871768183, "f1": 0.4715447154471545, "f1_std": 0.05882071122044139, "bacc": 0.4806949806949807, "bacc_std": 0.05586864545332324}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 82, "C": 0.005994842503189409, "split": "test", "acc": 0.6923076923076923, "acc_std": 0.05898531358681508, "f1": 0.6886973180076628, "f1_std": 0.059702359059180116, "bacc": 0.6906370656370657, "bacc_std": 0.05973689950897015}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 83, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.058050180292163286, "f1": 0.6198830409356726, "f1_std": 0.06017459518282782, "bacc": 0.6192084942084942, "bacc_std": 0.05946718303394383}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 84, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06545242456667899, "f1": 0.5500119360229172, "f1_std": 0.06612552498382575, "bacc": 0.5516409266409266, "bacc_std": 0.06640017639902036}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 85, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.058201039052112824, "f1": 0.5833333333333333, "f1_std": 0.06141692894530669, "bacc": 0.5834942084942085, "bacc_std": 0.05955812962727085}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 86, "C": 0.3593813663804626, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06266370465953001, "f1": 0.5608108108108107, "f1_std": 0.0638736425392196, "bacc": 0.5608108108108107, "bacc_std": 0.06345222216046117}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 87, "C": 0.3593813663804626, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06240130479594706, "f1": 0.5565302144249512, "f1_std": 0.06437824570753099, "bacc": 0.5564671814671815, "bacc_std": 0.06360930550576196}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 88, "C": 0.000774263682681127, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.059721634350270884, "f1": 0.545, "f1_std": 0.06365544719042884, "bacc": 0.5477799227799228, "bacc_std": 0.0608539734283045}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 89, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.055941431756634984, "f1": 0.61, "f1_std": 0.060568572762768585, "bacc": 0.6105212355212355, "bacc_std": 0.05787304586244429}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 90, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.06087096450300937, "f1": 0.5699583435432491, "f1_std": 0.0636951798511443, "bacc": 0.5699806949806949, "bacc_std": 0.06205491313918357}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 91, "C": 2.782559402207126, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.06010026336718279, "f1": 0.5699583435432491, "f1_std": 0.062378846189696247, "bacc": 0.5699806949806949, "bacc_std": 0.06085861925311064}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 92, "C": 0.046415888336127774, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06146311541252215, "f1": 0.545, "f1_std": 0.0649635730538412, "bacc": 0.5477799227799228, "bacc_std": 0.06227455677545912}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 93, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.0484503039480943, "f1": 0.622093023255814, "f1_std": 0.05827237282747697, "bacc": 0.6288610038610039, "bacc_std": 0.05124235322205058}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 94, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05223686607669715, "f1": 0.5308740978348035, "f1_std": 0.06316885300951601, "bacc": 0.5482625482625483, "bacc_std": 0.05419932100442337}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 95, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.0584713433073474, "f1": 0.5381034060279344, "f1_std": 0.0615082809429967, "bacc": 0.5386100386100386, "bacc_std": 0.06012061507801641}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 96, "C": 0.000774263682681127, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06177741300343071, "f1": 0.545, "f1_std": 0.06570498395186582, "bacc": 0.5477799227799228, "bacc_std": 0.0629116938929283}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 97, "C": 0.046415888336127774, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.059707174992308334, "f1": 0.545, "f1_std": 0.06421285537688758, "bacc": 0.5477799227799228, "bacc_std": 0.061117784635458534}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 98, "C": 0.005994842503189409, "split": "test", "acc": 0.676923076923077, "acc_std": 0.058956158618529875, "f1": 0.6655231560891939, "f1_std": 0.06125211219336189, "bacc": 0.6640926640926641, "bacc_std": 0.05999044654691344}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 99, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.059182109882780574, "f1": 0.5321419707123356, "f1_std": 0.061894703318438216, "bacc": 0.5342664092664092, "bacc_std": 0.060017085576320815}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adhd200_dx", "trial": 100, "C": 2.782559402207126, "split": "test", "acc": 0.6, "acc_std": 0.06010618611646771, "f1": 0.5976190476190476, "f1_std": 0.060400606376468, "bacc": 0.6008687258687259, "bacc_std": 0.060382175832968454}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset    | split   |   n_trials |      C |   C_std |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:-----------|:--------|-----------:|-------:|--------:|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | adhd200_dx | train   |        100 | 114.69 |  1006.9 | 0.77258 |  0.096606 | 0.76138 | 0.10314  | 0.75948 |   0.10258  |
+| flat_mae | patch  | logistic | adhd200_dx | test    |        100 | 114.69 |  1006.9 | 0.59123 |  0.047115 | 0.5718  | 0.047856 | 0.57458 |   0.047088 |
+
+
+done! total time: 0:04:48

decoders/crossreg_reg4/eval_v2/adhd200_dx__reg__logistic/config.yaml

@@ -0,0 +1,30 @@
+output_root: experiments/decoders/output
+name_prefix: eval_logistic
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (adhd200_dx reg logistic)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: decoders/crossreg_reg4/eval_v2/adhd200_dx__reg__logistic
+model: flat_mae
+representation: reg
+dataset: adhd200_dx
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/adhd200_dx__reg__logistic
+remote_dir: null

decoders/crossreg_reg4/eval_v2/adhd200_dx__reg__logistic/eval_table.csv

@@ -0,0 +1,203 @@
+model,repr,clf,dataset,trial,C,split,acc,acc_std,f1,f1_std,bacc,bacc_std
+flat_mae,reg,logistic,adhd200_dx,,0.005994842503189409,train,0.7424657534246575,0.021432767270098234,0.7329212853406402,0.0227187920646136,0.7302314221163827,0.02238807596624753
+flat_mae,reg,logistic,adhd200_dx,,0.005994842503189409,test,0.6307692307692307,0.05831768536246534,0.6036585365853658,0.06489270617713994,0.6061776061776062,0.06058231523849871
+flat_mae,reg,logistic,adhd200_dx,1,0.000774263682681127,train,0.7013698630136986,0.02304642422061071,0.682037962037962,0.02541632754350979,0.6809092019295353,0.02407710497437812
+flat_mae,reg,logistic,adhd200_dx,1,0.000774263682681127,test,0.6153846153846154,0.05411907974919205,0.5905769715293525,0.05803834559566884,0.5926640926640927,0.055124868248523945
+flat_mae,reg,logistic,adhd200_dx,2,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adhd200_dx,2,1291.5496650148827,test,0.6153846153846154,0.05982351954811971,0.61207925519217,0.05981363839382905,0.6143822393822393,0.05986497517915926
+flat_mae,reg,logistic,adhd200_dx,3,0.005994842503189409,train,0.7479452054794521,0.023249991111728194,0.7386038111844564,0.024648198527342884,0.7358032606704524,0.024378936098951278
+flat_mae,reg,logistic,adhd200_dx,3,0.005994842503189409,test,0.5538461538461539,0.060202254769152454,0.5321419707123356,0.0636341211291953,0.5342664092664092,0.061413303360600593
+flat_mae,reg,logistic,adhd200_dx,4,0.046415888336127774,train,0.8547945205479452,0.018185074684058288,0.8510330276218419,0.018842934900027218,0.8484001953959822,0.019027185030518316
+flat_mae,reg,logistic,adhd200_dx,4,0.046415888336127774,test,0.6461538461538462,0.05993241953436056,0.6407113674597452,0.061089912956029384,0.6414092664092663,0.060926841989675636
+flat_mae,reg,logistic,adhd200_dx,5,0.005994842503189409,train,0.7726027397260274,0.02138304897018664,0.7649163103616852,0.022612783110910736,0.7619527385968126,0.02249119771561826
+flat_mae,reg,logistic,adhd200_dx,5,0.005994842503189409,test,0.47692307692307695,0.06159348309491283,0.47078544061302685,0.062117338927684625,0.471042471042471,0.062499876082523505
+flat_mae,reg,logistic,adhd200_dx,6,0.005994842503189409,train,0.7397260273972602,0.020798504500418563,0.7279303878414111,0.022090635981922742,0.7249343591622397,0.02154517603989293
+flat_mae,reg,logistic,adhd200_dx,6,0.005994842503189409,test,0.7384615384615385,0.053718697504113615,0.7257383966244726,0.058208076716623855,0.7224903474903475,0.056404911743680015
+flat_mae,reg,logistic,adhd200_dx,7,0.046415888336127774,train,0.8301369863013699,0.01791263299229994,0.8255796029103466,0.018667046943875654,0.822968187091653,0.01887327657820457
+flat_mae,reg,logistic,adhd200_dx,7,0.046415888336127774,test,0.6307692307692307,0.056917080765269724,0.6198830409356726,0.058583785673106514,0.6192084942084942,0.05773996728003387
+flat_mae,reg,logistic,adhd200_dx,8,0.000774263682681127,train,0.6821917808219178,0.023722356639701497,0.6630483224040238,0.026608127465403652,0.6624839714233376,0.02515418969944185
+flat_mae,reg,logistic,adhd200_dx,8,0.000774263682681127,test,0.6461538461538462,0.06106987752426234,0.6289401836684041,0.0653821254722588,0.6283783783783784,0.06295308465640816
+flat_mae,reg,logistic,adhd200_dx,9,0.000774263682681127,train,0.7041095890410959,0.02425270350542799,0.6937919463087248,0.025495257997463203,0.6919460218599255,0.025124258012071896
+flat_mae,reg,logistic,adhd200_dx,9,0.000774263682681127,test,0.5384615384615384,0.05281294444160177,0.4846723044397463,0.06004305365181408,0.5033783783783784,0.053475815367940965
+flat_mae,reg,logistic,adhd200_dx,10,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adhd200_dx,10,2.782559402207126,test,0.5538461538461539,0.05879770554634587,0.5250692869740489,0.0635217987613546,0.5299227799227799,0.060001900121408776
+flat_mae,reg,logistic,adhd200_dx,11,0.000774263682681127,train,0.6794520547945205,0.02182284633716653,0.6499274524752232,0.025218409094760103,0.6528820907370092,0.02289642306486321
+flat_mae,reg,logistic,adhd200_dx,11,0.000774263682681127,test,0.6153846153846154,0.053510160512808304,0.5834401435529352,0.06111724552426378,0.5883204633204633,0.05553877742815464
+flat_mae,reg,logistic,adhd200_dx,12,0.005994842503189409,train,0.7589041095890411,0.023487155271573124,0.751976772918211,0.024292478379561354,0.7498168162667155,0.02408962731701068
+flat_mae,reg,logistic,adhd200_dx,12,0.005994842503189409,test,0.5538461538461539,0.05688007523359899,0.5321419707123356,0.060712565554667496,0.5342664092664092,0.05823910805825308
+flat_mae,reg,logistic,adhd200_dx,13,0.3593813663804626,train,0.9808219178082191,0.007022255054729113,0.9804551539491299,0.00717393691057325,0.9794223606277096,0.007561225095708522
+flat_mae,reg,logistic,adhd200_dx,13,0.3593813663804626,test,0.5384615384615384,0.060106383006353896,0.5294401544401545,0.061881266227136356,0.5294401544401545,0.06160227225701969
+flat_mae,reg,logistic,adhd200_dx,14,0.005994842503189409,train,0.7534246575342466,0.02216630636629756,0.7442863370282725,0.023292926587379276,0.7413750992245222,0.023006980497255344
+flat_mae,reg,logistic,adhd200_dx,14,0.005994842503189409,test,0.6461538461538462,0.05608597795418141,0.6233308138070043,0.061928870062683036,0.6240347490347491,0.05848052685892671
+flat_mae,reg,logistic,adhd200_dx,15,0.005994842503189409,train,0.7616438356164383,0.021972080750811307,0.7519935020813646,0.023376915649536808,0.7486566526225804,0.023008531648180043
+flat_mae,reg,logistic,adhd200_dx,15,0.005994842503189409,test,0.6461538461538462,0.058649890913300196,0.6407113674597452,0.05924847460360398,0.6414092664092663,0.059291018842661025
+flat_mae,reg,logistic,adhd200_dx,16,0.005994842503189409,train,0.7479452054794521,0.021429657126076224,0.7429262616364527,0.021988685828156107,0.7422604872687305,0.022045359695777537
+flat_mae,reg,logistic,adhd200_dx,16,0.005994842503189409,test,0.6307692307692307,0.05970876854717504,0.6198830409356726,0.06181351128591715,0.6192084942084942,0.06102563711942756
+flat_mae,reg,logistic,adhd200_dx,17,0.046415888336127774,train,0.8602739726027397,0.017523060644309675,0.8558418325860186,0.018274017709909614,0.8518196250839593,0.018350568811968268
+flat_mae,reg,logistic,adhd200_dx,17,0.046415888336127774,test,0.6153846153846154,0.05645310354097713,0.5905769715293525,0.062467746307242854,0.5926640926640927,0.05874038912674361
+flat_mae,reg,logistic,adhd200_dx,18,0.000774263682681127,train,0.684931506849315,0.02305297105739498,0.6625505470740982,0.025870807605397433,0.6627587470232643,0.024152588164445512
+flat_mae,reg,logistic,adhd200_dx,18,0.000774263682681127,test,0.6307692307692307,0.05155016170292421,0.587737843551797,0.06202705212374842,0.5974903474903475,0.05398135472476604
+flat_mae,reg,logistic,adhd200_dx,19,0.005994842503189409,train,0.7479452054794521,0.023945631167652776,0.7386038111844564,0.025268604078505005,0.7358032606704524,0.024901641656825742
+flat_mae,reg,logistic,adhd200_dx,19,0.005994842503189409,test,0.6923076923076923,0.053726072698215756,0.6697154471544715,0.060637238627391545,0.6689189189189189,0.0566715544766336
+flat_mae,reg,logistic,adhd200_dx,20,0.046415888336127774,train,0.8767123287671232,0.01641901693449214,0.8737402854968521,0.01694082423186053,0.871405019234292,0.017166366333024544
+flat_mae,reg,logistic,adhd200_dx,20,0.046415888336127774,test,0.6,0.055196055085861166,0.588206627680312,0.05765580387147078,0.5878378378378378,0.0566229473123185
+flat_mae,reg,logistic,adhd200_dx,21,0.005994842503189409,train,0.7506849315068493,0.022937833859816008,0.74059090447591,0.024358071143080633,0.737512975514441,0.023928879725356338
+flat_mae,reg,logistic,adhd200_dx,21,0.005994842503189409,test,0.6615384615384615,0.059167070643396824,0.6515594541910331,0.061347152699184,0.6505791505791505,0.06070435213929338
+flat_mae,reg,logistic,adhd200_dx,22,0.005994842503189409,train,0.7479452054794521,0.02131442334481628,0.7374429223744292,0.02254563754942046,0.7343683214263906,0.022112255922936624
+flat_mae,reg,logistic,adhd200_dx,22,0.005994842503189409,test,0.676923076923077,0.05792256818654922,0.6690909090909091,0.05927840706231435,0.6684362934362934,0.05876856765235649
+flat_mae,reg,logistic,adhd200_dx,23,0.005994842503189409,train,0.7616438356164383,0.02109528623934705,0.7530734597709193,0.022275355082248677,0.7500915918666422,0.02203220466244194
+flat_mae,reg,logistic,adhd200_dx,23,0.005994842503189409,test,0.5846153846153846,0.05426927383741408,0.5411764705882354,0.06215790303130002,0.5526061776061776,0.05567246337907471
+flat_mae,reg,logistic,adhd200_dx,24,0.005994842503189409,train,0.7589041095890411,0.020917929326559552,0.7515008974438324,0.021766369970395452,0.7490993466446846,0.021614116519920075
+flat_mae,reg,logistic,adhd200_dx,24,0.005994842503189409,test,0.676923076923077,0.05266031370067268,0.6500897205844656,0.0600487333055939,0.6510617760617761,0.05519130375835446
+flat_mae,reg,logistic,adhd200_dx,25,0.000774263682681127,train,0.673972602739726,0.022710543413478355,0.6486438388299722,0.02557879722762291,0.6501801306710631,0.02367818329277033
+flat_mae,reg,logistic,adhd200_dx,25,0.000774263682681127,test,0.6,0.05597844893851678,0.5626293995859213,0.06369848285291128,0.5704633204633205,0.05755100879334684
+flat_mae,reg,logistic,adhd200_dx,26,0.005994842503189409,train,0.7452054794520548,0.02214989180305054,0.7354763296317943,0.023511789538119717,0.7326586065824021,0.023139669715843768
+flat_mae,reg,logistic,adhd200_dx,26,0.005994842503189409,test,0.5538461538461539,0.058837688682233734,0.5381034060279344,0.06061713559113523,0.5386100386100386,0.0594176832945171
+flat_mae,reg,logistic,adhd200_dx,27,0.000774263682681127,train,0.684931506849315,0.02371374860226925,0.6706525747553924,0.025381640396760524,0.6692159736215424,0.024567572477019625
+flat_mae,reg,logistic,adhd200_dx,27,0.000774263682681127,test,0.5846153846153846,0.05810045575191198,0.5644080416976918,0.061223062076996924,0.5656370656370656,0.059142547955176164
+flat_mae,reg,logistic,adhd200_dx,28,0.005994842503189409,train,0.7643835616438356,0.020554927763372893,0.7576136644427971,0.021224727931052827,0.7553886548207852,0.021077732165782303
+flat_mae,reg,logistic,adhd200_dx,28,0.005994842503189409,test,0.5692307692307692,0.056173391751152095,0.545,0.05990328108932862,0.5477799227799228,0.057244263596738974
+flat_mae,reg,logistic,adhd200_dx,29,0.000774263682681127,train,0.6931506849315069,0.024383260579948367,0.6755555555555555,0.02619558065059491,0.674345118153508,0.025052829859681044
+flat_mae,reg,logistic,adhd200_dx,29,0.000774263682681127,test,0.6153846153846154,0.05356550157125276,0.5751633986928104,0.06256352385242225,0.583976833976834,0.0557865532664454
+flat_mae,reg,logistic,adhd200_dx,30,0.000774263682681127,train,0.6931506849315069,0.023546665516241544,0.6788587229763701,0.025734910444985017,0.6772149966416315,0.024854097442394734
+flat_mae,reg,logistic,adhd200_dx,30,0.000774263682681127,test,0.5846153846153846,0.05932626866613266,0.5578231292517006,0.06523901397147255,0.5612934362934363,0.06115253981183368
+flat_mae,reg,logistic,adhd200_dx,31,0.005994842503189409,train,0.7561643835616438,0.022577083986557823,0.7484298647089345,0.023434953979992876,0.7459546925566343,0.023179336800088667
+flat_mae,reg,logistic,adhd200_dx,31,0.005994842503189409,test,0.6153846153846154,0.058603196794681886,0.5905769715293525,0.0635328202944105,0.5926640926640927,0.060246074437261904
+flat_mae,reg,logistic,adhd200_dx,32,0.046415888336127774,train,0.863013698630137,0.0181830016776489,0.8588073280930866,0.019017806171317625,0.8549642791720096,0.019226617681662315
+flat_mae,reg,logistic,adhd200_dx,32,0.046415888336127774,test,0.5538461538461539,0.06127886784994003,0.5469838981014179,0.06226163907065569,0.5472972972972974,0.062280828805024564
+flat_mae,reg,logistic,adhd200_dx,33,0.3593813663804626,train,0.9780821917808219,0.006942057052981699,0.9776457618814307,0.007099990537129171,0.9762777065396593,0.007482900932862325
+flat_mae,reg,logistic,adhd200_dx,33,0.3593813663804626,test,0.6,0.061925513568877495,0.5953065134099617,0.06290342684129337,0.5965250965250966,0.06320772560998983
+flat_mae,reg,logistic,adhd200_dx,34,0.005994842503189409,train,0.7479452054794521,0.020606845991263133,0.7416446113128577,0.02121415364396825,0.7401080784026378,0.02114539141424023
+flat_mae,reg,logistic,adhd200_dx,34,0.005994842503189409,test,0.5846153846153846,0.059881342235532715,0.5810455956075435,0.06020453016925911,0.583011583011583,0.06033649823807296
+flat_mae,reg,logistic,adhd200_dx,35,0.005994842503189409,train,0.7534246575342466,0.02159453653720562,0.7425548589341693,0.02303682313429695,0.7392226903584295,0.02255250192219167
+flat_mae,reg,logistic,adhd200_dx,35,0.005994842503189409,test,0.5692307692307692,0.057396748562633146,0.5512820512820513,0.06075726221488239,0.5521235521235521,0.05890187492733762
+flat_mae,reg,logistic,adhd200_dx,36,0.005994842503189409,train,0.736986301369863,0.023348088588964106,0.7289100699387263,0.024407238139724938,0.7268119924284057,0.02425788152373696
+flat_mae,reg,logistic,adhd200_dx,36,0.005994842503189409,test,0.7230769230769231,0.05386394431381182,0.7149122807017544,0.056342636739836464,0.7133204633204633,0.05546971450442966
+flat_mae,reg,logistic,adhd200_dx,37,0.000774263682681127,train,0.6657534246575343,0.022335184219499322,0.642547521191883,0.02476569263412069,0.6436160468950357,0.023295470723575216
+flat_mae,reg,logistic,adhd200_dx,37,0.000774263682681127,test,0.676923076923077,0.04783300338696862,0.6351242983159583,0.05987041344292248,0.6423745173745173,0.051200731611956665
+flat_mae,reg,logistic,adhd200_dx,38,0.000774263682681127,train,0.6958904109589041,0.0241019732678938,0.6842781998831093,0.02542472445672463,0.6825120595957745,0.02489126691342445
+flat_mae,reg,logistic,adhd200_dx,38,0.000774263682681127,test,0.5230769230769231,0.06064992377932448,0.5115151515151515,0.061714609204104924,0.5115830115830116,0.0610801540411007
+flat_mae,reg,logistic,adhd200_dx,39,0.005994842503189409,train,0.7452054794520548,0.02238943535548907,0.734283634314163,0.02397634474925081,0.7312236673383403,0.02348873167729606
+flat_mae,reg,logistic,adhd200_dx,39,0.005994842503189409,test,0.6307692307692307,0.056621036507184586,0.6036585365853658,0.06219801627552449,0.6061776061776062,0.05833223636472194
+flat_mae,reg,logistic,adhd200_dx,40,10000.0,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adhd200_dx,40,10000.0,test,0.6,0.058673453973525636,0.5833333333333333,0.06058733046494953,0.5834942084942085,0.059312206242286496
+flat_mae,reg,logistic,adhd200_dx,41,0.005994842503189409,train,0.7863013698630137,0.020115617027929734,0.7783814920911696,0.021323852328439044,0.7748061305489405,0.021161157073443425
+flat_mae,reg,logistic,adhd200_dx,41,0.005994842503189409,test,0.5538461538461539,0.06032153687242134,0.5521501544309813,0.06052284569814894,0.555984555984556,0.06046982366147405
+flat_mae,reg,logistic,adhd200_dx,42,0.005994842503189409,train,0.7561643835616438,0.021688986372246803,0.7457122952038764,0.022763403077116686,0.7423673444464798,0.022309199254823133
+flat_mae,reg,logistic,adhd200_dx,42,0.005994842503189409,test,0.5384615384615384,0.06173270114398481,0.5374762808349146,0.0618262721673731,0.5424710424710424,0.06221652257362289
+flat_mae,reg,logistic,adhd200_dx,43,0.005994842503189409,train,0.7506849315068493,0.021999226175851577,0.74059090447591,0.023208868952126725,0.737512975514441,0.022750280865841875
+flat_mae,reg,logistic,adhd200_dx,43,0.005994842503189409,test,0.6153846153846154,0.0595263036682303,0.5966741126830479,0.06296582066821696,0.597007722007722,0.06068776921401086
+flat_mae,reg,logistic,adhd200_dx,44,0.005994842503189409,train,0.7479452054794521,0.022493717977929577,0.7380328278100231,0.023957101507183734,0.7350857910484215,0.023603705615218212
+flat_mae,reg,logistic,adhd200_dx,44,0.005994842503189409,test,0.6615384615384615,0.05573682029049267,0.6425000000000001,0.06073148828822711,0.6418918918918919,0.05809082562166728
+flat_mae,reg,logistic,adhd200_dx,45,0.005994842503189409,train,0.7726027397260274,0.022927756031113373,0.763396099686819,0.024488438140326658,0.7598003297307199,0.024196809060950585
+flat_mae,reg,logistic,adhd200_dx,45,0.005994842503189409,test,0.5230769230769231,0.06299233352432847,0.5062484685126194,0.06532016945212119,0.5072393822393823,0.06397253542919706
+flat_mae,reg,logistic,adhd200_dx,46,0.005994842503189409,train,0.7561643835616438,0.021738368897053252,0.7484298647089345,0.02269408552061702,0.7459546925566343,0.022538139806854894
+flat_mae,reg,logistic,adhd200_dx,46,0.005994842503189409,test,0.676923076923077,0.05643008975347992,0.656084656084656,0.06234477923313667,0.6554054054054055,0.05861205218561572
+flat_mae,reg,logistic,adhd200_dx,47,0.046415888336127774,train,0.8712328767123287,0.017134221166927466,0.8674071154188019,0.017995946575237984,0.8636807718141296,0.0183746087539179
+flat_mae,reg,logistic,adhd200_dx,47,0.046415888336127774,test,0.46153846153846156,0.06234833242527836,0.4484848484848485,0.062453538488762374,0.4488416988416989,0.06218658668910828
+flat_mae,reg,logistic,adhd200_dx,48,0.3593813663804626,train,0.9780821917808219,0.007980640866231637,0.9776457618814307,0.008172440564953318,0.9762777065396593,0.008675571382909449
+flat_mae,reg,logistic,adhd200_dx,48,0.3593813663804626,test,0.5692307692307692,0.06280399489799936,0.5666666666666667,0.06318806739640898,0.5694980694980695,0.06343270481017102
+flat_mae,reg,logistic,adhd200_dx,49,0.005994842503189409,train,0.7534246575342466,0.021693484930315893,0.7419400452488687,0.023095609813974294,0.7385052207363986,0.022531806730338985
+flat_mae,reg,logistic,adhd200_dx,49,0.005994842503189409,test,0.5846153846153846,0.0621158756827457,0.5699583435432491,0.06474627044543364,0.5699806949806949,0.06336489636056991
+flat_mae,reg,logistic,adhd200_dx,50,0.005994842503189409,train,0.7643835616438356,0.021669583436527546,0.7551176433876303,0.023207623751255957,0.7518013067106307,0.022898686829651068
+flat_mae,reg,logistic,adhd200_dx,50,0.005994842503189409,test,0.6,0.06033609221410216,0.5921814671814671,0.061513932952100905,0.5921814671814671,0.06099065790483331
+flat_mae,reg,logistic,adhd200_dx,51,0.005994842503189409,train,0.736986301369863,0.02251999799348507,0.7294292068198666,0.023325480869851655,0.7275294620504366,0.023157145967618722
+flat_mae,reg,logistic,adhd200_dx,51,0.005994842503189409,test,0.5846153846153846,0.06178359637529276,0.578226387887527,0.062920359380333,0.5786679536679536,0.0629802191214737
+flat_mae,reg,logistic,adhd200_dx,52,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adhd200_dx,52,21.54434690031882,test,0.6615384615384615,0.057219651525590914,0.6474358974358974,0.06013768427786563,0.6462355212355213,0.058820055283587996
+flat_mae,reg,logistic,adhd200_dx,53,0.005994842503189409,train,0.7287671232876712,0.02204774838520243,0.7164748252242072,0.02375588459307742,0.7137906820541002,0.023150719186480213
+flat_mae,reg,logistic,adhd200_dx,53,0.005994842503189409,test,0.676923076923077,0.05364590456803835,0.6431372549019607,0.06376578705762419,0.6467181467181468,0.05706349375027341
+flat_mae,reg,logistic,adhd200_dx,54,0.005994842503189409,train,0.7479452054794521,0.02089669341867548,0.7396899224806202,0.022032944345160215,0.7372381999145142,0.0218885581747688
+flat_mae,reg,logistic,adhd200_dx,54,0.005994842503189409,test,0.5692307692307692,0.058525179934713,0.5512820512820513,0.06211518037030336,0.5521235521235521,0.060094851279050426
+flat_mae,reg,logistic,adhd200_dx,55,0.046415888336127774,train,0.8684931506849315,0.01747241151632794,0.865874571288584,0.017876019713317017,0.8648409354582647,0.018043306517012383
+flat_mae,reg,logistic,adhd200_dx,55,0.046415888336127774,test,0.5692307692307692,0.06259961245319288,0.5666666666666667,0.06275244593889474,0.5694980694980695,0.0630294216601441
+flat_mae,reg,logistic,adhd200_dx,56,0.046415888336127774,train,0.8657534246575342,0.01755551022330597,0.8622758179900047,0.018139984459181076,0.8595438725041217,0.018257867290157565
+flat_mae,reg,logistic,adhd200_dx,56,0.046415888336127774,test,0.5538461538461539,0.06114817777309814,0.543030303030303,0.06277857389640336,0.542953667953668,0.061992414940413895
+flat_mae,reg,logistic,adhd200_dx,57,0.005994842503189409,train,0.7616438356164383,0.02218637106394491,0.752542372881356,0.023454888880720216,0.7493741222446113,0.023099047994873688
+flat_mae,reg,logistic,adhd200_dx,57,0.005994842503189409,test,0.6,0.05858267611528516,0.5833333333333333,0.06128363991719677,0.5834942084942085,0.059470954111315665
+flat_mae,reg,logistic,adhd200_dx,58,0.005994842503189409,train,0.7616438356164383,0.020893529342821054,0.7554693094924574,0.021611016491538963,0.7536789399767967,0.021579351029848752
+flat_mae,reg,logistic,adhd200_dx,58,0.005994842503189409,test,0.6153846153846154,0.05920079161472971,0.61207925519217,0.05991549883732456,0.6143822393822393,0.06017431376178625
+flat_mae,reg,logistic,adhd200_dx,59,0.000774263682681127,train,0.6958904109589041,0.02180147207529868,0.6805440567711414,0.023532635128757633,0.6789247114856201,0.022703938798496547
+flat_mae,reg,logistic,adhd200_dx,59,0.000774263682681127,test,0.6,0.05949921982482892,0.570630081300813,0.06512401365319602,0.5748069498069498,0.0609204679930452
+flat_mae,reg,logistic,adhd200_dx,60,0.005994842503189409,train,0.7479452054794521,0.0224685136961152,0.7416446113128577,0.023124939291092887,0.7401080784026378,0.02302797313141784
+flat_mae,reg,logistic,adhd200_dx,60,0.005994842503189409,test,0.6307692307692307,0.05391348757047567,0.61,0.057791572418128816,0.6105212355212355,0.055212965029909834
+flat_mae,reg,logistic,adhd200_dx,61,0.046415888336127774,train,0.8712328767123287,0.016785381767951853,0.8681287426300456,0.01731063454777775,0.8658331806802223,0.017536402153138092
+flat_mae,reg,logistic,adhd200_dx,61,0.046415888336127774,test,0.5692307692307692,0.05782168877080374,0.5512820512820513,0.06142783058378694,0.5521235521235521,0.0593999195454249
+flat_mae,reg,logistic,adhd200_dx,62,0.005994842503189409,train,0.7424657534246575,0.022063773548712313,0.7329212853406402,0.02343638130777813,0.7302314221163827,0.02314394581113892
+flat_mae,reg,logistic,adhd200_dx,62,0.005994842503189409,test,0.6,0.058634529479968354,0.5833333333333333,0.062153574966849055,0.5834942084942085,0.06036296920771857
+flat_mae,reg,logistic,adhd200_dx,63,0.005994842503189409,train,0.7589041095890411,0.022616865656459765,0.7515008974438324,0.02359278470886471,0.7490993466446846,0.023487318084081672
+flat_mae,reg,logistic,adhd200_dx,63,0.005994842503189409,test,0.5846153846153846,0.05859721907087602,0.5644080416976918,0.06171730498480744,0.5656370656370656,0.059677807764307776
+flat_mae,reg,logistic,adhd200_dx,64,0.005994842503189409,train,0.7506849315068493,0.02183332416080239,0.7393859504586148,0.023346164489483106,0.7360780362703792,0.02281094482993573
+flat_mae,reg,logistic,adhd200_dx,64,0.005994842503189409,test,0.6615384615384615,0.0562814482766599,0.6474358974358974,0.05967887286854398,0.6462355212355213,0.05805761168703123
+flat_mae,reg,logistic,adhd200_dx,65,0.000774263682681127,train,0.6821917808219178,0.023650343395571805,0.6630483224040238,0.025770354139461417,0.6624839714233376,0.024555299809935673
+flat_mae,reg,logistic,adhd200_dx,65,0.000774263682681127,test,0.5846153846153846,0.05821222141349264,0.5644080416976918,0.06191015276644697,0.5656370656370656,0.05957531362836289
+flat_mae,reg,logistic,adhd200_dx,66,0.005994842503189409,train,0.7479452054794521,0.02332096186160634,0.7386038111844564,0.024541357718368156,0.7358032606704524,0.024160544231172032
+flat_mae,reg,logistic,adhd200_dx,66,0.005994842503189409,test,0.6,0.05811557552277354,0.5833333333333333,0.061461961473551985,0.5834942084942085,0.0596798708976062
+flat_mae,reg,logistic,adhd200_dx,67,0.046415888336127774,train,0.852054794520548,0.018431426978645228,0.848085460599334,0.019189122910162986,0.8452555413079319,0.019436176348117066
+flat_mae,reg,logistic,adhd200_dx,67,0.046415888336127774,test,0.5230769230769231,0.05411358643341235,0.47320261437908495,0.059891286755037826,0.48986486486486486,0.05437409992054961
+flat_mae,reg,logistic,adhd200_dx,68,0.000774263682681127,train,0.6794520547945205,0.023078447726915165,0.6566818609362564,0.02599347806870773,0.6571869084691946,0.024302414786853407
+flat_mae,reg,logistic,adhd200_dx,68,0.000774263682681127,test,0.6153846153846154,0.05406548735632486,0.5834401435529352,0.060554709999589944,0.5883204633204633,0.05568895352250174
+flat_mae,reg,logistic,adhd200_dx,69,0.005994842503189409,train,0.7424657534246575,0.020449922174546727,0.732337889284154,0.02188352179916429,0.7295139524943518,0.021518657614676813
+flat_mae,reg,logistic,adhd200_dx,69,0.005994842503189409,test,0.5692307692307692,0.06026038961219067,0.5512820512820513,0.06256330745060376,0.5521235521235521,0.060852241961177354
+flat_mae,reg,logistic,adhd200_dx,70,0.000774263682681127,train,0.6767123287671233,0.023462966991437455,0.6562759369214071,0.025961263415341312,0.656194663247237,0.02450813716978358
+flat_mae,reg,logistic,adhd200_dx,70,0.000774263682681127,test,0.5538461538461539,0.059221905430678574,0.5321419707123356,0.06277704698561554,0.5342664092664092,0.0603208493686423
+flat_mae,reg,logistic,adhd200_dx,71,0.005994842503189409,train,0.7424657534246575,0.02252822320297486,0.732337889284154,0.02390465128166366,0.7295139524943518,0.023473984550725437
+flat_mae,reg,logistic,adhd200_dx,71,0.005994842503189409,test,0.6307692307692307,0.0554689528811191,0.61,0.060005201893700184,0.6105212355212355,0.05734453590027393
+flat_mae,reg,logistic,adhd200_dx,72,0.046415888336127774,train,0.8465753424657534,0.019143072545872265,0.841864207464257,0.020017202399587066,0.8382487635098004,0.020156941406506573
+flat_mae,reg,logistic,adhd200_dx,72,0.046415888336127774,test,0.6615384615384615,0.056923401246477316,0.6549227799227799,0.058000950235487084,0.6549227799227799,0.05761686566928672
+flat_mae,reg,logistic,adhd200_dx,73,0.000774263682681127,train,0.6876712328767123,0.024171066699219538,0.6723209223209223,0.02579035067267978,0.6709256884655309,0.024927662687461025
+flat_mae,reg,logistic,adhd200_dx,73,0.000774263682681127,test,0.6153846153846154,0.0615741127499759,0.5966741126830479,0.06519819896992515,0.597007722007722,0.06321107991353545
+flat_mae,reg,logistic,adhd200_dx,74,0.046415888336127774,train,0.8273972602739726,0.019368829097980066,0.8229260517014346,0.02005076725077155,0.8205410026256335,0.020147351140727375
+flat_mae,reg,logistic,adhd200_dx,74,0.046415888336127774,test,0.6461538461538462,0.06184462301970725,0.6431129147767964,0.062243731224664554,0.6457528957528957,0.062227302638423444
+flat_mae,reg,logistic,adhd200_dx,75,0.005994842503189409,train,0.7397260273972602,0.02204794992914451,0.7303675710142223,0.023184269460216296,0.7278042376503633,0.022857538155276864
+flat_mae,reg,logistic,adhd200_dx,75,0.005994842503189409,test,0.5846153846153846,0.0629935659342478,0.5830363506771205,0.0631466371586398,0.5873552123552124,0.06282279914818478
+flat_mae,reg,logistic,adhd200_dx,76,0.005994842503189409,train,0.7534246575342466,0.021475685300137163,0.7442863370282725,0.022619060555884814,0.7413750992245222,0.022269204530152315
+flat_mae,reg,logistic,adhd200_dx,76,0.005994842503189409,test,0.5692307692307692,0.05725392418271019,0.5289855072463768,0.06415230853343244,0.5390926640926641,0.0586922045248185
+flat_mae,reg,logistic,adhd200_dx,77,0.005994842503189409,train,0.7397260273972602,0.022739672553163733,0.729787648548607,0.024069764976024196,0.7270867680283324,0.023682906175033325
+flat_mae,reg,logistic,adhd200_dx,77,0.005994842503189409,test,0.7230769230769231,0.05086942095370932,0.7027439024390244,0.05787094737523379,0.7002895752895753,0.05402382190280039
+flat_mae,reg,logistic,adhd200_dx,78,0.046415888336127774,train,0.852054794520548,0.01953539853276905,0.8488682370261318,0.01997978533674775,0.8474079501740246,0.020002518716981648
+flat_mae,reg,logistic,adhd200_dx,78,0.046415888336127774,test,0.6153846153846154,0.06143279389447035,0.5966741126830479,0.06496056282255665,0.597007722007722,0.06261017182595881
+flat_mae,reg,logistic,adhd200_dx,79,0.005994842503189409,train,0.7671232876712328,0.021669735847016907,0.7592516431414847,0.022840145259893446,0.7563809000427428,0.022640456904089855
+flat_mae,reg,logistic,adhd200_dx,79,0.005994842503189409,test,0.5538461538461539,0.06175202174300945,0.5469838981014179,0.06275857001581663,0.5472972972972974,0.06263433558866377
+flat_mae,reg,logistic,adhd200_dx,80,0.005994842503189409,train,0.7506849315068493,0.021937960598437927,0.74059090447591,0.023306162882120354,0.737512975514441,0.02291692571854121
+flat_mae,reg,logistic,adhd200_dx,80,0.005994842503189409,test,0.6307692307692307,0.057260736571080904,0.5962732919254659,0.06572039762313099,0.6018339768339769,0.05971898122035945
+flat_mae,reg,logistic,adhd200_dx,81,0.000774263682681127,train,0.6821917808219178,0.024607557266989172,0.6673893916540976,0.026451070354174468,0.6660713195334921,0.025585212914469566
+flat_mae,reg,logistic,adhd200_dx,81,0.000774263682681127,test,0.676923076923077,0.04622558527187958,0.6259249109345026,0.06105849841764585,0.638030888030888,0.05003641377447831
+flat_mae,reg,logistic,adhd200_dx,82,0.005994842503189409,train,0.7315068493150685,0.022606657360777264,0.7232623630624497,0.023637004154921457,0.7212401538743359,0.023426185085978383
+flat_mae,reg,logistic,adhd200_dx,82,0.005994842503189409,test,0.6461538461538462,0.05731725475201767,0.644808743169399,0.05751726403773977,0.6500965250965252,0.057629030371378766
+flat_mae,reg,logistic,adhd200_dx,83,0.046415888336127774,train,0.8410958904109589,0.018045495349829466,0.835891472868217,0.01887964552888314,0.8319594553336997,0.018876637433863273
+flat_mae,reg,logistic,adhd200_dx,83,0.046415888336127774,test,0.7076923076923077,0.056523160671601624,0.6973780936045086,0.05901089982264924,0.6954633204633205,0.058281666753759793
+flat_mae,reg,logistic,adhd200_dx,84,0.005994842503189409,train,0.7287671232876712,0.02225445234728408,0.7177857092650011,0.023262081442072657,0.715225621298162,0.022738192589196196
+flat_mae,reg,logistic,adhd200_dx,84,0.005994842503189409,test,0.5692307692307692,0.06326329689265182,0.5608108108108107,0.06466085650148116,0.5608108108108107,0.06416369857028685
+flat_mae,reg,logistic,adhd200_dx,85,0.005994842503189409,train,0.7589041095890411,0.020941352486126767,0.747674710910005,0.022479079866986346,0.7440770592904683,0.021942427536242343
+flat_mae,reg,logistic,adhd200_dx,85,0.005994842503189409,test,0.6461538461538462,0.05842074690244087,0.6289401836684041,0.06293966442834756,0.6283783783783784,0.06063115315592142
+flat_mae,reg,logistic,adhd200_dx,86,0.005994842503189409,train,0.7506849315068493,0.022565559124972528,0.74059090447591,0.02403135160574751,0.737512975514441,0.02362074213561687
+flat_mae,reg,logistic,adhd200_dx,86,0.005994842503189409,test,0.5692307692307692,0.059017149247701246,0.5289855072463768,0.06665568052195199,0.5390926640926641,0.06069359024486291
+flat_mae,reg,logistic,adhd200_dx,87,0.000774263682681127,train,0.684931506849315,0.023000909688538004,0.6673113046786453,0.024847999616828646,0.6663460951334188,0.023811602202798765
+flat_mae,reg,logistic,adhd200_dx,87,0.000774263682681127,test,0.6615384615384615,0.05484675584953895,0.6299171842650104,0.06414568418164646,0.6332046332046332,0.05796768037251713
+flat_mae,reg,logistic,adhd200_dx,88,0.005994842503189409,train,0.7589041095890411,0.021526771333174304,0.751976772918211,0.022325447068227547,0.7498168162667155,0.022272923469937242
+flat_mae,reg,logistic,adhd200_dx,88,0.005994842503189409,test,0.6615384615384615,0.05805115882926127,0.6549227799227799,0.059667220267746526,0.6549227799227799,0.05961911702342157
+flat_mae,reg,logistic,adhd200_dx,89,0.000774263682681127,train,0.6794520547945205,0.0220234024865342,0.6632761679479597,0.023811810164930183,0.6622091958234109,0.02289076827814951
+flat_mae,reg,logistic,adhd200_dx,89,0.000774263682681127,test,0.6153846153846154,0.056099143007303114,0.5905769715293525,0.06136589239568586,0.5926640926640927,0.057735868699971354
+flat_mae,reg,logistic,adhd200_dx,90,0.000774263682681127,train,0.684931506849315,0.02287999254639028,0.6664123532730928,0.025002871088435356,0.6656286255113879,0.02384144272162364
+flat_mae,reg,logistic,adhd200_dx,90,0.000774263682681127,test,0.6,0.05530143326090255,0.5626293995859213,0.06388592380212471,0.5704633204633205,0.05731430090760336
+flat_mae,reg,logistic,adhd200_dx,91,0.005994842503189409,train,0.7616438356164383,0.021959945112439143,0.7519935020813646,0.023302449125030895,0.7486566526225804,0.022929980089565778
+flat_mae,reg,logistic,adhd200_dx,91,0.005994842503189409,test,0.6153846153846154,0.048978768495273634,0.5656241646618552,0.06027529294877572,0.5796332046332047,0.051299403091889045
+flat_mae,reg,logistic,adhd200_dx,92,0.005994842503189409,train,0.7397260273972602,0.02124530683441102,0.729787648548607,0.022458543317968416,0.7270867680283324,0.022095855987365857
+flat_mae,reg,logistic,adhd200_dx,92,0.005994842503189409,test,0.6,0.05826529004191637,0.570630081300813,0.06320303222900987,0.5748069498069498,0.05936981050674195
+flat_mae,reg,logistic,adhd200_dx,93,0.000774263682681127,train,0.6876712328767123,0.022710643227771048,0.664975845410628,0.025117962276566857,0.6651859314892837,0.023563096516942158
+flat_mae,reg,logistic,adhd200_dx,93,0.000774263682681127,test,0.6461538461538462,0.044685260929960995,0.5787545787545787,0.06231896123216173,0.6023166023166023,0.04794562836932044
+flat_mae,reg,logistic,adhd200_dx,94,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adhd200_dx,94,2.782559402207126,test,0.5076923076923077,0.06053623102136594,0.4715447154471545,0.06457680492433392,0.4806949806949807,0.06095776993353349
+flat_mae,reg,logistic,adhd200_dx,95,0.005994842503189409,train,0.7561643835616438,0.023236998212867992,0.74891985685688,0.024209439389893624,0.7466721621786652,0.024076971819372208
+flat_mae,reg,logistic,adhd200_dx,95,0.005994842503189409,test,0.5538461538461539,0.06043705904913702,0.5250692869740489,0.06469297773329087,0.5299227799227799,0.06129013832218233
+flat_mae,reg,logistic,adhd200_dx,96,0.005994842503189409,train,0.7424657534246575,0.02423066175401206,0.734031007751938,0.025373844090475557,0.7316663613604445,0.025146860490887206
+flat_mae,reg,logistic,adhd200_dx,96,0.005994842503189409,test,0.6,0.06206985185156913,0.5921814671814671,0.06370793698504675,0.5921814671814671,0.06340233085636056
+flat_mae,reg,logistic,adhd200_dx,97,0.005994842503189409,train,0.7506849315068493,0.021558314828908545,0.7411650107149814,0.02291181976520443,0.7382304451364718,0.022596383773597125
+flat_mae,reg,logistic,adhd200_dx,97,0.005994842503189409,test,0.5846153846153846,0.05306973418428373,0.5308740978348035,0.06383967896359853,0.5482625482625483,0.05516302619925288
+flat_mae,reg,logistic,adhd200_dx,98,0.3593813663804626,train,0.9808219178082191,0.007104030872990247,0.9804246060020994,0.00728531492705078,0.9787048910056787,0.007922362478625919
+flat_mae,reg,logistic,adhd200_dx,98,0.3593813663804626,test,0.6307692307692307,0.06178017146139056,0.6235521235521235,0.06267651123743039,0.6235521235521235,0.062255694841437746
+flat_mae,reg,logistic,adhd200_dx,99,0.005994842503189409,train,0.7616438356164383,0.023434882972050017,0.7545621072645906,0.024450675372950013,0.7522440007327349,0.024288332711246047
+flat_mae,reg,logistic,adhd200_dx,99,0.005994842503189409,test,0.5538461538461539,0.05793553650854149,0.5381034060279344,0.05950087924458394,0.5386100386100386,0.05837661897131186
+flat_mae,reg,logistic,adhd200_dx,100,0.3593813663804626,train,0.9835616438356164,0.006392139571516355,0.9832599523023299,0.006523881601514422,0.9825670147157599,0.0068471649720416765
+flat_mae,reg,logistic,adhd200_dx,100,0.3593813663804626,test,0.6,0.057642029373503656,0.5921814671814671,0.058544257057188034,0.5921814671814671,0.058299775323032925

decoders/crossreg_reg4/eval_v2/adhd200_dx__reg__logistic/log.txt

@@ -0,0 +1,241 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: aef99c83a386cf95c3d8ca503ecc968d8d5694af, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-03-07 21:19:16
+config:
+output_root: experiments/decoders/output
+name_prefix: eval_logistic
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (adhd200_dx reg logistic)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: decoders/crossreg_reg4/eval_v2/adhd200_dx__reg__logistic
+model: flat_mae
+representation: reg
+dataset: adhd200_dx
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/adhd200_dx__reg__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=False, reg_tokens=4, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: adhd200_dx (flat)
+train (n=301):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'gender', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 301
+}),
+    labels=['ADHD' 'Control'],
+    counts=[131 170]
+)
+
+validation (n=64):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'gender', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 64
+}),
+    labels=['ADHD' 'Control'],
+    counts=[28 36]
+)
+
+test (n=65):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'site', 'gender', 'dx', 'path', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 65
+}),
+    labels=['ADHD' 'Control'],
+    counts=[28 37]
+)
+
+extracting features for all splits
+extract (train)  [  0/151]  eta: 0:10:12    time: 4.0532  data: 3.2006  max mem: 2698
+extract (train)  [ 20/151]  eta: 0:00:55    time: 0.2408  data: 0.0889  max mem: 2852
+extract (train)  [ 40/151]  eta: 0:00:32    time: 0.1622  data: 0.0441  max mem: 2852
+extract (train)  [ 60/151]  eta: 0:00:23    time: 0.1770  data: 0.0584  max mem: 2852
+extract (train)  [ 80/151]  eta: 0:00:16    time: 0.1758  data: 0.0514  max mem: 2852
+extract (train)  [100/151]  eta: 0:00:11    time: 0.1642  data: 0.0486  max mem: 2852
+extract (train)  [120/151]  eta: 0:00:06    time: 0.1905  data: 0.0638  max mem: 2852
+extract (train)  [140/151]  eta: 0:00:02    time: 0.1611  data: 0.0479  max mem: 2852
+extract (train)  [150/151]  eta: 0:00:00    time: 0.1545  data: 0.0465  max mem: 2852
+extract (train) Total time: 0:00:31 (0.2072 s / it)
+extract (validation)  [ 0/32]  eta: 0:01:53    time: 3.5614  data: 3.3840  max mem: 2852
+extract (validation)  [20/32]  eta: 0:00:04    time: 0.1896  data: 0.0577  max mem: 2852
+extract (validation)  [31/32]  eta: 0:00:00    time: 0.1607  data: 0.0458  max mem: 2852
+extract (validation) Total time: 0:00:09 (0.2957 s / it)
+extract (test)  [ 0/33]  eta: 0:02:22    time: 4.3322  data: 4.1131  max mem: 2852
+extract (test)  [20/33]  eta: 0:00:05    time: 0.2033  data: 0.0683  max mem: 2852
+extract (test)  [32/33]  eta: 0:00:00    time: 0.1679  data: 0.0524  max mem: 2852
+extract (test) Total time: 0:00:10 (0.3232 s / it)
+feature extraction time: 0:00:51
+train features: (301, 768)
+validation features: (64, 768)
+test features: (65, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset    | trial   |         C | split   |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:-----------|:--------|----------:|:--------|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | reg    | logistic | adhd200_dx |         | 0.0059948 | train   | 0.74247 |  0.021433 | 0.73292 | 0.022719 | 0.73023 |   0.022388 |
+| flat_mae | reg    | logistic | adhd200_dx |         | 0.0059948 | test    | 0.63077 |  0.058318 | 0.60366 | 0.064893 | 0.60618 |   0.060582 |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 1, "C": 0.000774263682681127, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05411907974919205, "f1": 0.5905769715293525, "f1_std": 0.05803834559566884, "bacc": 0.5926640926640927, "bacc_std": 0.055124868248523945}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 2, "C": 1291.5496650148827, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05982351954811971, "f1": 0.61207925519217, "f1_std": 0.05981363839382905, "bacc": 0.6143822393822393, "bacc_std": 0.05986497517915926}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 3, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.060202254769152454, "f1": 0.5321419707123356, "f1_std": 0.0636341211291953, "bacc": 0.5342664092664092, "bacc_std": 0.061413303360600593}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 4, "C": 0.046415888336127774, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.05993241953436056, "f1": 0.6407113674597452, "f1_std": 0.061089912956029384, "bacc": 0.6414092664092663, "bacc_std": 0.060926841989675636}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 5, "C": 0.005994842503189409, "split": "test", "acc": 0.47692307692307695, "acc_std": 0.06159348309491283, "f1": 0.47078544061302685, "f1_std": 0.062117338927684625, "bacc": 0.471042471042471, "bacc_std": 0.062499876082523505}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 6, "C": 0.005994842503189409, "split": "test", "acc": 0.7384615384615385, "acc_std": 0.053718697504113615, "f1": 0.7257383966244726, "f1_std": 0.058208076716623855, "bacc": 0.7224903474903475, "bacc_std": 0.056404911743680015}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 7, "C": 0.046415888336127774, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.056917080765269724, "f1": 0.6198830409356726, "f1_std": 0.058583785673106514, "bacc": 0.6192084942084942, "bacc_std": 0.05773996728003387}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 8, "C": 0.000774263682681127, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.06106987752426234, "f1": 0.6289401836684041, "f1_std": 0.0653821254722588, "bacc": 0.6283783783783784, "bacc_std": 0.06295308465640816}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 9, "C": 0.000774263682681127, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.05281294444160177, "f1": 0.4846723044397463, "f1_std": 0.06004305365181408, "bacc": 0.5033783783783784, "bacc_std": 0.053475815367940965}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 10, "C": 2.782559402207126, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.05879770554634587, "f1": 0.5250692869740489, "f1_std": 0.0635217987613546, "bacc": 0.5299227799227799, "bacc_std": 0.060001900121408776}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 11, "C": 0.000774263682681127, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.053510160512808304, "f1": 0.5834401435529352, "f1_std": 0.06111724552426378, "bacc": 0.5883204633204633, "bacc_std": 0.05553877742815464}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 12, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.05688007523359899, "f1": 0.5321419707123356, "f1_std": 0.060712565554667496, "bacc": 0.5342664092664092, "bacc_std": 0.05823910805825308}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 13, "C": 0.3593813663804626, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.060106383006353896, "f1": 0.5294401544401545, "f1_std": 0.061881266227136356, "bacc": 0.5294401544401545, "bacc_std": 0.06160227225701969}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 14, "C": 0.005994842503189409, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.05608597795418141, "f1": 0.6233308138070043, "f1_std": 0.061928870062683036, "bacc": 0.6240347490347491, "bacc_std": 0.05848052685892671}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 15, "C": 0.005994842503189409, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.058649890913300196, "f1": 0.6407113674597452, "f1_std": 0.05924847460360398, "bacc": 0.6414092664092663, "bacc_std": 0.059291018842661025}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 16, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05970876854717504, "f1": 0.6198830409356726, "f1_std": 0.06181351128591715, "bacc": 0.6192084942084942, "bacc_std": 0.06102563711942756}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 17, "C": 0.046415888336127774, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05645310354097713, "f1": 0.5905769715293525, "f1_std": 0.062467746307242854, "bacc": 0.5926640926640927, "bacc_std": 0.05874038912674361}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 18, "C": 0.000774263682681127, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05155016170292421, "f1": 0.587737843551797, "f1_std": 0.06202705212374842, "bacc": 0.5974903474903475, "bacc_std": 0.05398135472476604}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 19, "C": 0.005994842503189409, "split": "test", "acc": 0.6923076923076923, "acc_std": 0.053726072698215756, "f1": 0.6697154471544715, "f1_std": 0.060637238627391545, "bacc": 0.6689189189189189, "bacc_std": 0.0566715544766336}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 20, "C": 0.046415888336127774, "split": "test", "acc": 0.6, "acc_std": 0.055196055085861166, "f1": 0.588206627680312, "f1_std": 0.05765580387147078, "bacc": 0.5878378378378378, "bacc_std": 0.0566229473123185}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 21, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.059167070643396824, "f1": 0.6515594541910331, "f1_std": 0.061347152699184, "bacc": 0.6505791505791505, "bacc_std": 0.06070435213929338}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 22, "C": 0.005994842503189409, "split": "test", "acc": 0.676923076923077, "acc_std": 0.05792256818654922, "f1": 0.6690909090909091, "f1_std": 0.05927840706231435, "bacc": 0.6684362934362934, "bacc_std": 0.05876856765235649}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 23, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05426927383741408, "f1": 0.5411764705882354, "f1_std": 0.06215790303130002, "bacc": 0.5526061776061776, "bacc_std": 0.05567246337907471}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 24, "C": 0.005994842503189409, "split": "test", "acc": 0.676923076923077, "acc_std": 0.05266031370067268, "f1": 0.6500897205844656, "f1_std": 0.0600487333055939, "bacc": 0.6510617760617761, "bacc_std": 0.05519130375835446}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 25, "C": 0.000774263682681127, "split": "test", "acc": 0.6, "acc_std": 0.05597844893851678, "f1": 0.5626293995859213, "f1_std": 0.06369848285291128, "bacc": 0.5704633204633205, "bacc_std": 0.05755100879334684}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 26, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.058837688682233734, "f1": 0.5381034060279344, "f1_std": 0.06061713559113523, "bacc": 0.5386100386100386, "bacc_std": 0.0594176832945171}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 27, "C": 0.000774263682681127, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05810045575191198, "f1": 0.5644080416976918, "f1_std": 0.061223062076996924, "bacc": 0.5656370656370656, "bacc_std": 0.059142547955176164}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 28, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.056173391751152095, "f1": 0.545, "f1_std": 0.05990328108932862, "bacc": 0.5477799227799228, "bacc_std": 0.057244263596738974}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 29, "C": 0.000774263682681127, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05356550157125276, "f1": 0.5751633986928104, "f1_std": 0.06256352385242225, "bacc": 0.583976833976834, "bacc_std": 0.0557865532664454}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 30, "C": 0.000774263682681127, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05932626866613266, "f1": 0.5578231292517006, "f1_std": 0.06523901397147255, "bacc": 0.5612934362934363, "bacc_std": 0.06115253981183368}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 31, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.058603196794681886, "f1": 0.5905769715293525, "f1_std": 0.0635328202944105, "bacc": 0.5926640926640927, "bacc_std": 0.060246074437261904}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 32, "C": 0.046415888336127774, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06127886784994003, "f1": 0.5469838981014179, "f1_std": 0.06226163907065569, "bacc": 0.5472972972972974, "bacc_std": 0.062280828805024564}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 33, "C": 0.3593813663804626, "split": "test", "acc": 0.6, "acc_std": 0.061925513568877495, "f1": 0.5953065134099617, "f1_std": 0.06290342684129337, "bacc": 0.5965250965250966, "bacc_std": 0.06320772560998983}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 34, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.059881342235532715, "f1": 0.5810455956075435, "f1_std": 0.06020453016925911, "bacc": 0.583011583011583, "bacc_std": 0.06033649823807296}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 35, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.057396748562633146, "f1": 0.5512820512820513, "f1_std": 0.06075726221488239, "bacc": 0.5521235521235521, "bacc_std": 0.05890187492733762}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 36, "C": 0.005994842503189409, "split": "test", "acc": 0.7230769230769231, "acc_std": 0.05386394431381182, "f1": 0.7149122807017544, "f1_std": 0.056342636739836464, "bacc": 0.7133204633204633, "bacc_std": 0.05546971450442966}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 37, "C": 0.000774263682681127, "split": "test", "acc": 0.676923076923077, "acc_std": 0.04783300338696862, "f1": 0.6351242983159583, "f1_std": 0.05987041344292248, "bacc": 0.6423745173745173, "bacc_std": 0.051200731611956665}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 38, "C": 0.000774263682681127, "split": "test", "acc": 0.5230769230769231, "acc_std": 0.06064992377932448, "f1": 0.5115151515151515, "f1_std": 0.061714609204104924, "bacc": 0.5115830115830116, "bacc_std": 0.0610801540411007}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 39, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.056621036507184586, "f1": 0.6036585365853658, "f1_std": 0.06219801627552449, "bacc": 0.6061776061776062, "bacc_std": 0.05833223636472194}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 40, "C": 10000.0, "split": "test", "acc": 0.6, "acc_std": 0.058673453973525636, "f1": 0.5833333333333333, "f1_std": 0.06058733046494953, "bacc": 0.5834942084942085, "bacc_std": 0.059312206242286496}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 41, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06032153687242134, "f1": 0.5521501544309813, "f1_std": 0.06052284569814894, "bacc": 0.555984555984556, "bacc_std": 0.06046982366147405}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 42, "C": 0.005994842503189409, "split": "test", "acc": 0.5384615384615384, "acc_std": 0.06173270114398481, "f1": 0.5374762808349146, "f1_std": 0.0618262721673731, "bacc": 0.5424710424710424, "bacc_std": 0.06221652257362289}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 43, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.0595263036682303, "f1": 0.5966741126830479, "f1_std": 0.06296582066821696, "bacc": 0.597007722007722, "bacc_std": 0.06068776921401086}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 44, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.05573682029049267, "f1": 0.6425000000000001, "f1_std": 0.06073148828822711, "bacc": 0.6418918918918919, "bacc_std": 0.05809082562166728}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 45, "C": 0.005994842503189409, "split": "test", "acc": 0.5230769230769231, "acc_std": 0.06299233352432847, "f1": 0.5062484685126194, "f1_std": 0.06532016945212119, "bacc": 0.5072393822393823, "bacc_std": 0.06397253542919706}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 46, "C": 0.005994842503189409, "split": "test", "acc": 0.676923076923077, "acc_std": 0.05643008975347992, "f1": 0.656084656084656, "f1_std": 0.06234477923313667, "bacc": 0.6554054054054055, "bacc_std": 0.05861205218561572}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 47, "C": 0.046415888336127774, "split": "test", "acc": 0.46153846153846156, "acc_std": 0.06234833242527836, "f1": 0.4484848484848485, "f1_std": 0.062453538488762374, "bacc": 0.4488416988416989, "bacc_std": 0.06218658668910828}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 48, "C": 0.3593813663804626, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06280399489799936, "f1": 0.5666666666666667, "f1_std": 0.06318806739640898, "bacc": 0.5694980694980695, "bacc_std": 0.06343270481017102}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 49, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.0621158756827457, "f1": 0.5699583435432491, "f1_std": 0.06474627044543364, "bacc": 0.5699806949806949, "bacc_std": 0.06336489636056991}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 50, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.06033609221410216, "f1": 0.5921814671814671, "f1_std": 0.061513932952100905, "bacc": 0.5921814671814671, "bacc_std": 0.06099065790483331}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 51, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.06178359637529276, "f1": 0.578226387887527, "f1_std": 0.062920359380333, "bacc": 0.5786679536679536, "bacc_std": 0.0629802191214737}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 52, "C": 21.54434690031882, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.057219651525590914, "f1": 0.6474358974358974, "f1_std": 0.06013768427786563, "bacc": 0.6462355212355213, "bacc_std": 0.058820055283587996}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 53, "C": 0.005994842503189409, "split": "test", "acc": 0.676923076923077, "acc_std": 0.05364590456803835, "f1": 0.6431372549019607, "f1_std": 0.06376578705762419, "bacc": 0.6467181467181468, "bacc_std": 0.05706349375027341}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 54, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.058525179934713, "f1": 0.5512820512820513, "f1_std": 0.06211518037030336, "bacc": 0.5521235521235521, "bacc_std": 0.060094851279050426}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 55, "C": 0.046415888336127774, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06259961245319288, "f1": 0.5666666666666667, "f1_std": 0.06275244593889474, "bacc": 0.5694980694980695, "bacc_std": 0.0630294216601441}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 56, "C": 0.046415888336127774, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06114817777309814, "f1": 0.543030303030303, "f1_std": 0.06277857389640336, "bacc": 0.542953667953668, "bacc_std": 0.061992414940413895}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 57, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.05858267611528516, "f1": 0.5833333333333333, "f1_std": 0.06128363991719677, "bacc": 0.5834942084942085, "bacc_std": 0.059470954111315665}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 58, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05920079161472971, "f1": 0.61207925519217, "f1_std": 0.05991549883732456, "bacc": 0.6143822393822393, "bacc_std": 0.06017431376178625}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 59, "C": 0.000774263682681127, "split": "test", "acc": 0.6, "acc_std": 0.05949921982482892, "f1": 0.570630081300813, "f1_std": 0.06512401365319602, "bacc": 0.5748069498069498, "bacc_std": 0.0609204679930452}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 60, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.05391348757047567, "f1": 0.61, "f1_std": 0.057791572418128816, "bacc": 0.6105212355212355, "bacc_std": 0.055212965029909834}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 61, "C": 0.046415888336127774, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.05782168877080374, "f1": 0.5512820512820513, "f1_std": 0.06142783058378694, "bacc": 0.5521235521235521, "bacc_std": 0.0593999195454249}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 62, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.058634529479968354, "f1": 0.5833333333333333, "f1_std": 0.062153574966849055, "bacc": 0.5834942084942085, "bacc_std": 0.06036296920771857}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 63, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05859721907087602, "f1": 0.5644080416976918, "f1_std": 0.06171730498480744, "bacc": 0.5656370656370656, "bacc_std": 0.059677807764307776}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 64, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.0562814482766599, "f1": 0.6474358974358974, "f1_std": 0.05967887286854398, "bacc": 0.6462355212355213, "bacc_std": 0.05805761168703123}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 65, "C": 0.000774263682681127, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05821222141349264, "f1": 0.5644080416976918, "f1_std": 0.06191015276644697, "bacc": 0.5656370656370656, "bacc_std": 0.05957531362836289}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 66, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.05811557552277354, "f1": 0.5833333333333333, "f1_std": 0.061461961473551985, "bacc": 0.5834942084942085, "bacc_std": 0.0596798708976062}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 67, "C": 0.046415888336127774, "split": "test", "acc": 0.5230769230769231, "acc_std": 0.05411358643341235, "f1": 0.47320261437908495, "f1_std": 0.059891286755037826, "bacc": 0.48986486486486486, "bacc_std": 0.05437409992054961}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 68, "C": 0.000774263682681127, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.05406548735632486, "f1": 0.5834401435529352, "f1_std": 0.060554709999589944, "bacc": 0.5883204633204633, "bacc_std": 0.05568895352250174}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 69, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06026038961219067, "f1": 0.5512820512820513, "f1_std": 0.06256330745060376, "bacc": 0.5521235521235521, "bacc_std": 0.060852241961177354}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 70, "C": 0.000774263682681127, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.059221905430678574, "f1": 0.5321419707123356, "f1_std": 0.06277704698561554, "bacc": 0.5342664092664092, "bacc_std": 0.0603208493686423}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 71, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.0554689528811191, "f1": 0.61, "f1_std": 0.060005201893700184, "bacc": 0.6105212355212355, "bacc_std": 0.05734453590027393}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 72, "C": 0.046415888336127774, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.056923401246477316, "f1": 0.6549227799227799, "f1_std": 0.058000950235487084, "bacc": 0.6549227799227799, "bacc_std": 0.05761686566928672}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 73, "C": 0.000774263682681127, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.0615741127499759, "f1": 0.5966741126830479, "f1_std": 0.06519819896992515, "bacc": 0.597007722007722, "bacc_std": 0.06321107991353545}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 74, "C": 0.046415888336127774, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.06184462301970725, "f1": 0.6431129147767964, "f1_std": 0.062243731224664554, "bacc": 0.6457528957528957, "bacc_std": 0.062227302638423444}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 75, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.0629935659342478, "f1": 0.5830363506771205, "f1_std": 0.0631466371586398, "bacc": 0.5873552123552124, "bacc_std": 0.06282279914818478}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 76, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.05725392418271019, "f1": 0.5289855072463768, "f1_std": 0.06415230853343244, "bacc": 0.5390926640926641, "bacc_std": 0.0586922045248185}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 77, "C": 0.005994842503189409, "split": "test", "acc": 0.7230769230769231, "acc_std": 0.05086942095370932, "f1": 0.7027439024390244, "f1_std": 0.05787094737523379, "bacc": 0.7002895752895753, "bacc_std": 0.05402382190280039}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 78, "C": 0.046415888336127774, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.06143279389447035, "f1": 0.5966741126830479, "f1_std": 0.06496056282255665, "bacc": 0.597007722007722, "bacc_std": 0.06261017182595881}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 79, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06175202174300945, "f1": 0.5469838981014179, "f1_std": 0.06275857001581663, "bacc": 0.5472972972972974, "bacc_std": 0.06263433558866377}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 80, "C": 0.005994842503189409, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.057260736571080904, "f1": 0.5962732919254659, "f1_std": 0.06572039762313099, "bacc": 0.6018339768339769, "bacc_std": 0.05971898122035945}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 81, "C": 0.000774263682681127, "split": "test", "acc": 0.676923076923077, "acc_std": 0.04622558527187958, "f1": 0.6259249109345026, "f1_std": 0.06105849841764585, "bacc": 0.638030888030888, "bacc_std": 0.05003641377447831}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 82, "C": 0.005994842503189409, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.05731725475201767, "f1": 0.644808743169399, "f1_std": 0.05751726403773977, "bacc": 0.6500965250965252, "bacc_std": 0.057629030371378766}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 83, "C": 0.046415888336127774, "split": "test", "acc": 0.7076923076923077, "acc_std": 0.056523160671601624, "f1": 0.6973780936045086, "f1_std": 0.05901089982264924, "bacc": 0.6954633204633205, "bacc_std": 0.058281666753759793}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 84, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.06326329689265182, "f1": 0.5608108108108107, "f1_std": 0.06466085650148116, "bacc": 0.5608108108108107, "bacc_std": 0.06416369857028685}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 85, "C": 0.005994842503189409, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.05842074690244087, "f1": 0.6289401836684041, "f1_std": 0.06293966442834756, "bacc": 0.6283783783783784, "bacc_std": 0.06063115315592142}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 86, "C": 0.005994842503189409, "split": "test", "acc": 0.5692307692307692, "acc_std": 0.059017149247701246, "f1": 0.5289855072463768, "f1_std": 0.06665568052195199, "bacc": 0.5390926640926641, "bacc_std": 0.06069359024486291}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 87, "C": 0.000774263682681127, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.05484675584953895, "f1": 0.6299171842650104, "f1_std": 0.06414568418164646, "bacc": 0.6332046332046332, "bacc_std": 0.05796768037251713}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 88, "C": 0.005994842503189409, "split": "test", "acc": 0.6615384615384615, "acc_std": 0.05805115882926127, "f1": 0.6549227799227799, "f1_std": 0.059667220267746526, "bacc": 0.6549227799227799, "bacc_std": 0.05961911702342157}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 89, "C": 0.000774263682681127, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.056099143007303114, "f1": 0.5905769715293525, "f1_std": 0.06136589239568586, "bacc": 0.5926640926640927, "bacc_std": 0.057735868699971354}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 90, "C": 0.000774263682681127, "split": "test", "acc": 0.6, "acc_std": 0.05530143326090255, "f1": 0.5626293995859213, "f1_std": 0.06388592380212471, "bacc": 0.5704633204633205, "bacc_std": 0.05731430090760336}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 91, "C": 0.005994842503189409, "split": "test", "acc": 0.6153846153846154, "acc_std": 0.048978768495273634, "f1": 0.5656241646618552, "f1_std": 0.06027529294877572, "bacc": 0.5796332046332047, "bacc_std": 0.051299403091889045}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 92, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.05826529004191637, "f1": 0.570630081300813, "f1_std": 0.06320303222900987, "bacc": 0.5748069498069498, "bacc_std": 0.05936981050674195}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 93, "C": 0.000774263682681127, "split": "test", "acc": 0.6461538461538462, "acc_std": 0.044685260929960995, "f1": 0.5787545787545787, "f1_std": 0.06231896123216173, "bacc": 0.6023166023166023, "bacc_std": 0.04794562836932044}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 94, "C": 2.782559402207126, "split": "test", "acc": 0.5076923076923077, "acc_std": 0.06053623102136594, "f1": 0.4715447154471545, "f1_std": 0.06457680492433392, "bacc": 0.4806949806949807, "bacc_std": 0.06095776993353349}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 95, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.06043705904913702, "f1": 0.5250692869740489, "f1_std": 0.06469297773329087, "bacc": 0.5299227799227799, "bacc_std": 0.06129013832218233}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 96, "C": 0.005994842503189409, "split": "test", "acc": 0.6, "acc_std": 0.06206985185156913, "f1": 0.5921814671814671, "f1_std": 0.06370793698504675, "bacc": 0.5921814671814671, "bacc_std": 0.06340233085636056}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 97, "C": 0.005994842503189409, "split": "test", "acc": 0.5846153846153846, "acc_std": 0.05306973418428373, "f1": 0.5308740978348035, "f1_std": 0.06383967896359853, "bacc": 0.5482625482625483, "bacc_std": 0.05516302619925288}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 98, "C": 0.3593813663804626, "split": "test", "acc": 0.6307692307692307, "acc_std": 0.06178017146139056, "f1": 0.6235521235521235, "f1_std": 0.06267651123743039, "bacc": 0.6235521235521235, "bacc_std": 0.062255694841437746}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 99, "C": 0.005994842503189409, "split": "test", "acc": 0.5538461538461539, "acc_std": 0.05793553650854149, "f1": 0.5381034060279344, "f1_std": 0.05950087924458394, "bacc": 0.5386100386100386, "bacc_std": 0.05837661897131186}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adhd200_dx", "trial": 100, "C": 0.3593813663804626, "split": "test", "acc": 0.6, "acc_std": 0.057642029373503656, "f1": 0.5921814671814671, "f1_std": 0.058544257057188034, "bacc": 0.5921814671814671, "bacc_std": 0.058299775323032925}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset    | split   |   n_trials |      C |   C_std |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:-----------|:--------|-----------:|-------:|--------:|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | reg    | logistic | adhd200_dx | train   |        100 | 113.21 |    1007 | 0.7763  |  0.087812 | 0.76689 | 0.092902 | 0.76483 |   0.092879 |
+| flat_mae | reg    | logistic | adhd200_dx | test    |        100 | 113.21 |    1007 | 0.60554 |  0.051585 | 0.58546 | 0.053233 | 0.58827 |   0.051522 |
+
+
+done! total time: 0:04:50

decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__patch__logistic/config.yaml

@@ -0,0 +1,30 @@
+output_root: experiments/decoders/output
+name_prefix: eval_logistic
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (adni_ad_vs_cn patch logistic)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__patch__logistic
+model: flat_mae
+representation: patch
+dataset: adni_ad_vs_cn
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__patch__logistic
+remote_dir: null

decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__patch__logistic/eval_table.csv

@@ -0,0 +1,203 @@
+model,repr,clf,dataset,trial,C,split,acc,acc_std,f1,f1_std,bacc,bacc_std
+flat_mae,patch,logistic,adni_ad_vs_cn,,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,,166.81005372000556,test,0.6585365853658537,0.06721052012495501,0.5017361111111112,0.07684463515541452,0.5017361111111112,0.07654227112140105
+flat_mae,patch,logistic,adni_ad_vs_cn,1,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,1,166.81005372000556,test,0.7804878048780488,0.06175586637339688,0.7119437939110069,0.07982166095879796,0.7193548387096774,0.08516609223998574
+flat_mae,patch,logistic,adni_ad_vs_cn,2,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,2,2.782559402207126,test,0.7317073170731707,0.06428491014860176,0.6232247284878863,0.09162267407571494,0.6193548387096774,0.08786441042986236
+flat_mae,patch,logistic,adni_ad_vs_cn,3,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,3,21.54434690031882,test,0.6829268292682927,0.06948437611929524,0.5839188134270101,0.08497702446519155,0.5870967741935484,0.08792274414851722
+flat_mae,patch,logistic,adni_ad_vs_cn,4,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,4,166.81005372000556,test,0.8048780487804879,0.06127466177863133,0.783068783068783,0.05915914345251882,0.8709677419354839,0.040520340853611066
+flat_mae,patch,logistic,adni_ad_vs_cn,5,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,5,166.81005372000556,test,0.6829268292682927,0.0604363274265884,0.5547201336675021,0.0811416666945755,0.5532258064516129,0.07772808309192142
+flat_mae,patch,logistic,adni_ad_vs_cn,6,0.3593813663804626,train,0.975609756097561,0.007708574789036132,0.9648738695859115,0.011449808348987367,0.9517215876407263,0.015417822803499251
+flat_mae,patch,logistic,adni_ad_vs_cn,6,0.3593813663804626,test,0.7317073170731707,0.06479200235603146,0.6232247284878863,0.08930198197826686,0.6193548387096774,0.08677124800133185
+flat_mae,patch,logistic,adni_ad_vs_cn,7,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,7,166.81005372000556,test,0.7560975609756098,0.06372417884650156,0.6893939393939394,0.0784376993864998,0.7032258064516128,0.084067210612129
+flat_mae,patch,logistic,adni_ad_vs_cn,8,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,8,166.81005372000556,test,0.7317073170731707,0.06641316068338539,0.6479313036690086,0.08202671334913557,0.6532258064516129,0.0854468406709409
+flat_mae,patch,logistic,adni_ad_vs_cn,9,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,9,166.81005372000556,test,0.7804878048780488,0.0619757777327563,0.7119437939110069,0.08030073846778696,0.7193548387096774,0.08475320178801186
+flat_mae,patch,logistic,adni_ad_vs_cn,10,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,10,2.782559402207126,test,0.7317073170731707,0.06399665369003064,0.6479313036690086,0.08319998345571941,0.6532258064516129,0.08832318716066874
+flat_mae,patch,logistic,adni_ad_vs_cn,11,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,11,166.81005372000556,test,0.8048780487804879,0.05120557301614695,0.6893939393939394,0.09174013333260418,0.667741935483871,0.08241255189397681
+flat_mae,patch,logistic,adni_ad_vs_cn,12,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,12,166.81005372000556,test,0.7073170731707317,0.06106450106001445,0.5729166666666666,0.08751751006546477,0.5693548387096774,0.08186769382609813
+flat_mae,patch,logistic,adni_ad_vs_cn,13,0.046415888336127774,train,0.907859078590786,0.014137782041226739,0.8577551020408163,0.023699210872975725,0.8266085956118004,0.02589682684477505
+flat_mae,patch,logistic,adni_ad_vs_cn,13,0.046415888336127774,test,0.7560975609756098,0.05345814020626708,0.6117424242424243,0.09202468610571168,0.6016129032258064,0.0791058140916389
+flat_mae,patch,logistic,adni_ad_vs_cn,14,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,14,21.54434690031882,test,0.7560975609756098,0.06369536361170726,0.6693548387096775,0.08657520433598431,0.6693548387096775,0.08820304807605202
+flat_mae,patch,logistic,adni_ad_vs_cn,15,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,15,2.782559402207126,test,0.7560975609756098,0.06447931754782824,0.6693548387096775,0.08742264337378126,0.6693548387096775,0.08954270364040907
+flat_mae,patch,logistic,adni_ad_vs_cn,16,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,16,166.81005372000556,test,0.8292682926829268,0.04575690362803435,0.7144278606965174,0.09279895838312108,0.6838709677419355,0.07932248822621175
+flat_mae,patch,logistic,adni_ad_vs_cn,17,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,17,1291.5496650148827,test,0.6585365853658537,0.0711080986208361,0.5651515151515152,0.0852582525873165,0.5709677419354839,0.08977297363352293
+flat_mae,patch,logistic,adni_ad_vs_cn,18,0.3593813663804626,train,0.9701897018970189,0.008288273089307956,0.9566924513203521,0.012549269658976547,0.9400936806639822,0.017195936895224584
+flat_mae,patch,logistic,adni_ad_vs_cn,18,0.3593813663804626,test,0.8780487804878049,0.03814609194603692,0.7960199004975124,0.08301462416148489,0.75,0.07819948848937569
+flat_mae,patch,logistic,adni_ad_vs_cn,19,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,19,1291.5496650148827,test,0.7560975609756098,0.060457486500875714,0.6440972222222222,0.08494828399370564,0.635483870967742,0.08026925623399012
+flat_mae,patch,logistic,adni_ad_vs_cn,20,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,20,2.782559402207126,test,0.8048780487804879,0.06407793811820159,0.764367816091954,0.07146125035822799,0.8032258064516129,0.07456277628052574
+flat_mae,patch,logistic,adni_ad_vs_cn,21,10000.0,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,21,10000.0,test,0.7560975609756098,0.058075808423928135,0.6440972222222222,0.08780002288873556,0.635483870967742,0.08256684587327397
+flat_mae,patch,logistic,adni_ad_vs_cn,22,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,22,166.81005372000556,test,0.8292682926829268,0.051874648712642704,0.7402714932126697,0.08889597556100112,0.717741935483871,0.08333749263987081
+flat_mae,patch,logistic,adni_ad_vs_cn,23,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,23,1291.5496650148827,test,0.7073170731707317,0.06808896444446465,0.6272727272727273,0.08379512140986728,0.6370967741935484,0.0899293051054737
+flat_mae,patch,logistic,adni_ad_vs_cn,24,0.3593813663804626,train,0.983739837398374,0.0065384704279851,0.9766829555986183,0.00964648513664028,0.9651162790697674,0.014027299929805235
+flat_mae,patch,logistic,adni_ad_vs_cn,24,0.3593813663804626,test,0.6585365853658537,0.0730399481610212,0.5651515151515152,0.08570021624212107,0.5709677419354839,0.0912322758576622
+flat_mae,patch,logistic,adni_ad_vs_cn,25,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,25,21.54434690031882,test,0.7560975609756098,0.06502429875755188,0.6693548387096775,0.08661123117886423,0.6693548387096775,0.08749686576815639
+flat_mae,patch,logistic,adni_ad_vs_cn,26,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,26,2.782559402207126,test,0.7073170731707317,0.06437876837875586,0.603225806451613,0.08250327793555129,0.603225806451613,0.08280767299952495
+flat_mae,patch,logistic,adni_ad_vs_cn,27,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,27,166.81005372000556,test,0.7317073170731707,0.06309206892481466,0.6232247284878863,0.08547873759960081,0.6193548387096774,0.08351086534453545
+flat_mae,patch,logistic,adni_ad_vs_cn,28,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,28,2.782559402207126,test,0.7317073170731707,0.061738043058860034,0.6232247284878863,0.08457621269232801,0.6193548387096774,0.08213022222486381
+flat_mae,patch,logistic,adni_ad_vs_cn,29,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,29,1291.5496650148827,test,0.7073170731707317,0.07105567523251301,0.6272727272727273,0.08378732150900603,0.6370967741935484,0.08966387946055397
+flat_mae,patch,logistic,adni_ad_vs_cn,30,0.3593813663804626,train,0.983739837398374,0.006559159158815651,0.9768796992481203,0.009479445150930505,0.9691634481058427,0.012714895190838198
+flat_mae,patch,logistic,adni_ad_vs_cn,30,0.3593813663804626,test,0.7804878048780488,0.04778185109921527,0.6328358208955224,0.09319085105882638,0.6177419354838709,0.07593114487288175
+flat_mae,patch,logistic,adni_ad_vs_cn,31,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,31,21.54434690031882,test,0.7317073170731707,0.05097528299580182,0.5512437810945273,0.08895511919602032,0.5516129032258065,0.07090380837178853
+flat_mae,patch,logistic,adni_ad_vs_cn,32,0.046415888336127774,train,0.8943089430894309,0.013554222689814526,0.8308632543926662,0.025235333292744886,0.7934916591338648,0.0272427568828056
+flat_mae,patch,logistic,adni_ad_vs_cn,32,0.046415888336127774,test,0.6829268292682927,0.060572010624942316,0.5176470588235295,0.08190754554917075,0.5193548387096775,0.07278577578846016
+flat_mae,patch,logistic,adni_ad_vs_cn,33,0.3593813663804626,train,0.981029810298103,0.007124400959681112,0.9729123189697663,0.010422202997663934,0.9633494946174705,0.014179217743666121
+flat_mae,patch,logistic,adni_ad_vs_cn,33,0.3593813663804626,test,0.7560975609756098,0.06147417637105602,0.6893939393939394,0.07571128957091285,0.7032258064516128,0.08194897494976601
+flat_mae,patch,logistic,adni_ad_vs_cn,34,0.3593813663804626,train,0.9728997289972899,0.008204313324002887,0.9608013937282229,0.01230811473248259,0.9459076341523543,0.01656423681111059
+flat_mae,patch,logistic,adni_ad_vs_cn,34,0.3593813663804626,test,0.7804878048780488,0.062007195759818313,0.7119437939110069,0.07932457979132376,0.7193548387096774,0.08408095202384551
+flat_mae,patch,logistic,adni_ad_vs_cn,35,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,35,2.782559402207126,test,0.8292682926829268,0.05069360670594106,0.7402714932126697,0.08787541773146086,0.717741935483871,0.08333689079468769
+flat_mae,patch,logistic,adni_ad_vs_cn,36,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,36,166.81005372000556,test,0.6829268292682927,0.07224472587268083,0.5839188134270101,0.08441764447486973,0.5870967741935484,0.08659274535656608
+flat_mae,patch,logistic,adni_ad_vs_cn,37,0.3593813663804626,train,0.975609756097561,0.007619223721560962,0.9648738695859115,0.011305089118887153,0.9517215876407263,0.01532224775225614
+flat_mae,patch,logistic,adni_ad_vs_cn,37,0.3593813663804626,test,0.7073170731707317,0.05848117295950731,0.5729166666666666,0.0864804969138768,0.5693548387096774,0.07968586869134761
+flat_mae,patch,logistic,adni_ad_vs_cn,38,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,38,166.81005372000556,test,0.7317073170731707,0.06711712971906027,0.6479313036690086,0.08419349273094341,0.6532258064516129,0.08785933908891233
+flat_mae,patch,logistic,adni_ad_vs_cn,39,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,39,2.782559402207126,test,0.6829268292682927,0.06702252633471767,0.5547201336675021,0.08647402444083507,0.5532258064516129,0.08406115972553356
+flat_mae,patch,logistic,adni_ad_vs_cn,40,0.005994842503189409,train,0.8238482384823849,0.013158234002658847,0.6787307285318189,0.030644591915016183,0.6504232065083408,0.02491712736724843
+flat_mae,patch,logistic,adni_ad_vs_cn,40,0.005994842503189409,test,0.7317073170731707,0.04141989107375309,0.4972129319955407,0.07434953475605574,0.5177419354838709,0.05316787697999455
+flat_mae,patch,logistic,adni_ad_vs_cn,41,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,41,1291.5496650148827,test,0.7317073170731707,0.06787837743012205,0.6479313036690086,0.0851149835833768,0.6532258064516129,0.08764715534317276
+flat_mae,patch,logistic,adni_ad_vs_cn,42,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,42,166.81005372000556,test,0.6097560975609756,0.06637732172945972,0.47096774193548385,0.0727034266598651,0.47096774193548385,0.07333956454307135
+flat_mae,patch,logistic,adni_ad_vs_cn,43,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,43,166.81005372000556,test,0.7560975609756098,0.05879366313225619,0.6440972222222222,0.08787938032120693,0.635483870967742,0.08440252251237289
+flat_mae,patch,logistic,adni_ad_vs_cn,44,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,44,1291.5496650148827,test,0.7804878048780488,0.053685558606160026,0.6660633484162897,0.09088552881025565,0.6516129032258065,0.08400442304147263
+flat_mae,patch,logistic,adni_ad_vs_cn,45,0.005994842503189409,train,0.8238482384823849,0.012812434930479106,0.66522437015842,0.03335021052656857,0.638281699400115,0.025531601804187433
+flat_mae,patch,logistic,adni_ad_vs_cn,45,0.005994842503189409,test,0.8292682926829268,0.03504395132812059,0.6800445930880714,0.09352338117352493,0.65,0.07184010022264724
+flat_mae,patch,logistic,adni_ad_vs_cn,46,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,46,21.54434690031882,test,0.8048780487804879,0.053717838627345894,0.7152777777777778,0.08634601572777,0.7016129032258065,0.0851015784674308
+flat_mae,patch,logistic,adni_ad_vs_cn,47,10000.0,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,47,10000.0,test,0.7317073170731707,0.05180846067259221,0.5512437810945273,0.08654368574712812,0.5516129032258065,0.06887771184022792
+flat_mae,patch,logistic,adni_ad_vs_cn,48,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,48,21.54434690031882,test,0.7073170731707317,0.07177524574147591,0.646551724137931,0.08069310588351807,0.6709677419354838,0.09070532385276033
+flat_mae,patch,logistic,adni_ad_vs_cn,49,0.3593813663804626,train,0.981029810298103,0.006848965014229861,0.9729123189697663,0.010020162666062555,0.9633494946174705,0.013738646042460334
+flat_mae,patch,logistic,adni_ad_vs_cn,49,0.3593813663804626,test,0.6829268292682927,0.06505340322129365,0.5547201336675021,0.08618118501625521,0.5532258064516129,0.08281085000319596
+flat_mae,patch,logistic,adni_ad_vs_cn,50,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,50,21.54434690031882,test,0.7804878048780488,0.05796173017929806,0.6917293233082706,0.082836555047868,0.685483870967742,0.08079937068969135
+flat_mae,patch,logistic,adni_ad_vs_cn,51,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,51,1291.5496650148827,test,0.7073170731707317,0.06379389403478483,0.603225806451613,0.0842392921222764,0.603225806451613,0.08491606457492491
+flat_mae,patch,logistic,adni_ad_vs_cn,52,10000.0,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,52,10000.0,test,0.6097560975609756,0.0723268409477098,0.5287356321839081,0.07849295074799864,0.5387096774193548,0.08671881936254502
+flat_mae,patch,logistic,adni_ad_vs_cn,53,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,53,166.81005372000556,test,0.7560975609756098,0.05463414634146341,0.6117424242424243,0.09245079057319427,0.6016129032258064,0.07855340091304261
+flat_mae,patch,logistic,adni_ad_vs_cn,54,0.005994842503189409,train,0.8265582655826558,0.01307658590850382,0.6772006560962274,0.03129827302541176,0.6481428219245624,0.024816675539712823
+flat_mae,patch,logistic,adni_ad_vs_cn,54,0.005994842503189409,test,0.7560975609756098,0.043246701792655724,0.569327731092437,0.08661224222477408,0.567741935483871,0.06507710139450365
+flat_mae,patch,logistic,adni_ad_vs_cn,55,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,55,166.81005372000556,test,0.6829268292682927,0.07491260027493644,0.6259649122807017,0.07967230381546084,0.6548387096774193,0.08799429854531957
+flat_mae,patch,logistic,adni_ad_vs_cn,56,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,56,2.782559402207126,test,0.7560975609756098,0.06456399322146954,0.6693548387096775,0.08350107478969533,0.6693548387096775,0.08552273802962136
+flat_mae,patch,logistic,adni_ad_vs_cn,57,0.046415888336127774,train,0.8970189701897019,0.014490204467636533,0.8377684191040355,0.025952902301689142,0.8033527816583121,0.02804637318616214
+flat_mae,patch,logistic,adni_ad_vs_cn,57,0.046415888336127774,test,0.7073170731707317,0.06265693681558678,0.5729166666666666,0.08851241615812118,0.5693548387096774,0.08031695899487994
+flat_mae,patch,logistic,adni_ad_vs_cn,58,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,58,21.54434690031882,test,0.7804878048780488,0.06301145755703659,0.7119437939110069,0.08059125331526244,0.7193548387096774,0.08491596500908126
+flat_mae,patch,logistic,adni_ad_vs_cn,59,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,59,166.81005372000556,test,0.6829268292682927,0.06620653423560834,0.5839188134270101,0.08236688301326865,0.5870967741935484,0.08558086250473963
+flat_mae,patch,logistic,adni_ad_vs_cn,60,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,60,166.81005372000556,test,0.7317073170731707,0.06909857256497813,0.6676492262343405,0.08115322014834417,0.6870967741935483,0.08707174096641203
+flat_mae,patch,logistic,adni_ad_vs_cn,61,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,61,2.782559402207126,test,0.7073170731707317,0.06908674250022279,0.646551724137931,0.07872518273436985,0.6709677419354838,0.08707151581180167
+flat_mae,patch,logistic,adni_ad_vs_cn,62,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,62,166.81005372000556,test,0.6097560975609756,0.07572358072732534,0.5287356321839081,0.08277261956763161,0.5387096774193548,0.09205784562266699
+flat_mae,patch,logistic,adni_ad_vs_cn,63,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,63,166.81005372000556,test,0.6829268292682927,0.0734152661831638,0.6259649122807017,0.08025615812395835,0.6548387096774193,0.08996184785051084
+flat_mae,patch,logistic,adni_ad_vs_cn,64,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,64,21.54434690031882,test,0.7317073170731707,0.06738905659997667,0.6479313036690086,0.08728104520261751,0.6532258064516129,0.09135951418689889
+flat_mae,patch,logistic,adni_ad_vs_cn,65,0.046415888336127774,train,0.9051490514905149,0.013138443923635432,0.851341551849166,0.022812890984969258,0.8167474730873532,0.025038613147335433
+flat_mae,patch,logistic,adni_ad_vs_cn,65,0.046415888336127774,test,0.6585365853658537,0.05693606629763894,0.4564393939393939,0.0689541429669208,0.4693548387096774,0.06066977631889546
+flat_mae,patch,logistic,adni_ad_vs_cn,66,0.046415888336127774,train,0.8997289972899729,0.014385051554063727,0.8428467833834041,0.02517726875059078,0.8091667351466842,0.027314934931126078
+flat_mae,patch,logistic,adni_ad_vs_cn,66,0.046415888336127774,test,0.7560975609756098,0.05811496534376152,0.6440972222222222,0.0883637958972104,0.635483870967742,0.08271046870689876
+flat_mae,patch,logistic,adni_ad_vs_cn,67,0.3593813663804626,train,0.975609756097561,0.007574379819315634,0.9648738695859115,0.01124975992953432,0.9517215876407263,0.015042054484084388
+flat_mae,patch,logistic,adni_ad_vs_cn,67,0.3593813663804626,test,0.8536585365853658,0.04583027426562429,0.7670454545454546,0.08723058539747552,0.7338709677419355,0.08246891250438924
+flat_mae,patch,logistic,adni_ad_vs_cn,68,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,68,166.81005372000556,test,0.5853658536585366,0.07401906598765423,0.4863669859985261,0.08002417651154202,0.4887096774193548,0.08731729854014975
+flat_mae,patch,logistic,adni_ad_vs_cn,69,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,69,21.54434690031882,test,0.7560975609756098,0.06311846410995943,0.6893939393939394,0.07791947521483165,0.7032258064516128,0.08371906691078246
+flat_mae,patch,logistic,adni_ad_vs_cn,70,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,70,166.81005372000556,test,0.7073170731707317,0.054348031306029526,0.5340909090909092,0.08348737885347941,0.535483870967742,0.07056282217321602
+flat_mae,patch,logistic,adni_ad_vs_cn,71,0.046415888336127774,train,0.9051490514905149,0.013348935503183838,0.851341551849166,0.02344744741716075,0.8167474730873532,0.026260800659585477
+flat_mae,patch,logistic,adni_ad_vs_cn,71,0.046415888336127774,test,0.6829268292682927,0.050431947420682396,0.4696517412935323,0.06966787464617082,0.4854838709677419,0.05737906741775723
+flat_mae,patch,logistic,adni_ad_vs_cn,72,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,72,21.54434690031882,test,0.7317073170731707,0.039054866148776776,0.4972129319955407,0.07117770726827559,0.5177419354838709,0.0502263592071848
+flat_mae,patch,logistic,adni_ad_vs_cn,73,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,73,21.54434690031882,test,0.8048780487804879,0.06222086506928569,0.7354838709677419,0.08379770716347627,0.7354838709677419,0.08570259030759712
+flat_mae,patch,logistic,adni_ad_vs_cn,74,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,74,166.81005372000556,test,0.7560975609756098,0.0591854283720946,0.6440972222222222,0.0879698454449336,0.635483870967742,0.08222180832861987
+flat_mae,patch,logistic,adni_ad_vs_cn,75,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,75,2.782559402207126,test,0.8048780487804879,0.057072649570270084,0.7152777777777778,0.08873922328495465,0.7016129032258065,0.08743770525172663
+flat_mae,patch,logistic,adni_ad_vs_cn,76,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,76,21.54434690031882,test,0.7073170731707317,0.06581590607116995,0.603225806451613,0.08495157671395279,0.603225806451613,0.08502532376640316
+flat_mae,patch,logistic,adni_ad_vs_cn,77,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,77,166.81005372000556,test,0.7804878048780488,0.05596078766509243,0.6660633484162897,0.09223254185416252,0.6516129032258065,0.0852674496561477
+flat_mae,patch,logistic,adni_ad_vs_cn,78,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,78,166.81005372000556,test,0.7560975609756098,0.06717368964562526,0.7054597701149425,0.07507979001414222,0.7370967741935484,0.08133086616951436
+flat_mae,patch,logistic,adni_ad_vs_cn,79,0.3593813663804626,train,0.975609756097561,0.007465323414233344,0.9648738695859115,0.011107240241222735,0.9517215876407263,0.014977646220472533
+flat_mae,patch,logistic,adni_ad_vs_cn,79,0.3593813663804626,test,0.8536585365853658,0.05086874081777024,0.8016129032258064,0.07080787710040606,0.8016129032258064,0.0760470725436277
+flat_mae,patch,logistic,adni_ad_vs_cn,80,0.005994842503189409,train,0.8184281842818428,0.0138037235523689,0.6597299640777902,0.033433408262326596,0.6347481304955214,0.02601925090819029
+flat_mae,patch,logistic,adni_ad_vs_cn,80,0.005994842503189409,test,0.7560975609756098,0.04653287810564446,0.569327731092437,0.09082647728295093,0.567741935483871,0.06927126329752148
+flat_mae,patch,logistic,adni_ad_vs_cn,81,0.005994842503189409,train,0.8211382113821138,0.012468806121412535,0.6576417004048583,0.032476081067111535,0.6324677459117429,0.024563884736420162
+flat_mae,patch,logistic,adni_ad_vs_cn,81,0.005994842503189409,test,0.8048780487804879,0.03243440275164686,0.6095238095238095,0.09923652315029885,0.6,0.06649052564087607
+flat_mae,patch,logistic,adni_ad_vs_cn,82,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,82,2.782559402207126,test,0.7804878048780488,0.06339316709820725,0.7410526315789474,0.06983750868286032,0.7870967741935484,0.07512549583776422
+flat_mae,patch,logistic,adni_ad_vs_cn,83,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,83,1291.5496650148827,test,0.7560975609756098,0.05954359427574256,0.6440972222222222,0.08815069104446678,0.635483870967742,0.0839670799175816
+flat_mae,patch,logistic,adni_ad_vs_cn,84,0.046415888336127774,train,0.8888888888888888,0.01405966134057904,0.8240507065185788,0.02505437313369703,0.7899580902292711,0.026487319574584705
+flat_mae,patch,logistic,adni_ad_vs_cn,84,0.046415888336127774,test,0.7317073170731707,0.06004955320380165,0.5918552036199095,0.09122675202019988,0.5854838709677419,0.08151191635463105
+flat_mae,patch,logistic,adni_ad_vs_cn,85,0.3593813663804626,train,0.9728997289972899,0.007670853181249552,0.9608013937282229,0.011479427457058183,0.9459076341523543,0.015269017669144085
+flat_mae,patch,logistic,adni_ad_vs_cn,85,0.3593813663804626,test,0.8536585365853658,0.05588719836515769,0.8016129032258064,0.07605139267403929,0.8016129032258064,0.07898267042109322
+flat_mae,patch,logistic,adni_ad_vs_cn,86,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,86,2.782559402207126,test,0.6829268292682927,0.06103458611281899,0.5547201336675021,0.07988054022907731,0.5532258064516129,0.0777431176459443
+flat_mae,patch,logistic,adni_ad_vs_cn,87,0.005994842503189409,train,0.8211382113821138,0.013633525515117949,0.6671131765992345,0.032165534630711534,0.6405620839838935,0.025181131172370214
+flat_mae,patch,logistic,adni_ad_vs_cn,87,0.005994842503189409,test,0.8048780487804879,0.03149161149082581,0.6095238095238095,0.09722792853269492,0.6,0.06455780355619295
+flat_mae,patch,logistic,adni_ad_vs_cn,88,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,88,21.54434690031882,test,0.7073170731707317,0.07111024026170445,0.6620879120879121,0.0757924017478316,0.7048387096774194,0.08472392449535054
+flat_mae,patch,logistic,adni_ad_vs_cn,89,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,89,166.81005372000556,test,0.7560975609756098,0.06625776642881673,0.7054597701149425,0.07523037894605458,0.7370967741935484,0.08163465450295516
+flat_mae,patch,logistic,adni_ad_vs_cn,90,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,90,2.782559402207126,test,0.7560975609756098,0.06554580254154335,0.6893939393939394,0.07941124009353895,0.7032258064516128,0.08448739199944477
+flat_mae,patch,logistic,adni_ad_vs_cn,91,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,91,166.81005372000556,test,0.6341463414634146,0.0653331552725744,0.5199063231850116,0.07647618156375047,0.5209677419354839,0.0803545641077397
+flat_mae,patch,logistic,adni_ad_vs_cn,92,0.3593813663804626,train,0.975609756097561,0.007798346785477701,0.9648738695859115,0.011602399887716544,0.9517215876407263,0.01574706113489123
+flat_mae,patch,logistic,adni_ad_vs_cn,92,0.3593813663804626,test,0.6829268292682927,0.06160824979330037,0.5547201336675021,0.0821168795678255,0.5532258064516129,0.07936587020097206
+flat_mae,patch,logistic,adni_ad_vs_cn,93,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,93,2.782559402207126,test,0.6829268292682927,0.06310006406077198,0.5547201336675021,0.08275078503274196,0.5532258064516129,0.0796312035222479
+flat_mae,patch,logistic,adni_ad_vs_cn,94,0.046415888336127774,train,0.9024390243902439,0.014027112836810532,0.8493877551020408,0.02363886833492022,0.8190278576711316,0.02572259716768545
+flat_mae,patch,logistic,adni_ad_vs_cn,94,0.046415888336127774,test,0.6585365853658537,0.05667016354352588,0.4564393939393939,0.06808356982840677,0.4693548387096774,0.05991912204183492
+flat_mae,patch,logistic,adni_ad_vs_cn,95,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,patch,logistic,adni_ad_vs_cn,95,166.81005372000556,test,0.7804878048780488,0.05992460065145339,0.6917293233082706,0.08260908925667741,0.685483870967742,0.08349758892878209
+flat_mae,patch,logistic,adni_ad_vs_cn,96,0.3593813663804626,train,0.975609756097561,0.00805657648046619,0.9648738695859115,0.012032566686906486,0.9517215876407263,0.01648970949118496
+flat_mae,patch,logistic,adni_ad_vs_cn,96,0.3593813663804626,test,0.7560975609756098,0.0329243359489596,0.5119047619047619,0.07887369477154188,0.5338709677419355,0.05146099934604913
+flat_mae,patch,logistic,adni_ad_vs_cn,97,0.046415888336127774,train,0.8970189701897019,0.012783926019707197,0.8343179884677191,0.02369689063426384,0.7952584435861616,0.025257049816878374
+flat_mae,patch,logistic,adni_ad_vs_cn,97,0.046415888336127774,test,0.7073170731707317,0.034281405355135894,0.4142857142857143,0.01198068442848582,0.46774193548387094,0.022669961605815688
+flat_mae,patch,logistic,adni_ad_vs_cn,98,0.005994842503189409,train,0.8265582655826558,0.012387113816980542,0.6772006560962274,0.030512394849135333,0.6481428219245624,0.024218525055412284
+flat_mae,patch,logistic,adni_ad_vs_cn,98,0.005994842503189409,test,0.7317073170731707,0.05482840030560244,0.5918552036199095,0.08798511562641391,0.5854838709677419,0.07903013367783404
+flat_mae,patch,logistic,adni_ad_vs_cn,99,0.3593813663804626,train,0.983739837398374,0.006637997713035651,0.9766829555986183,0.009808844220459231,0.9651162790697674,0.014240820675059026
+flat_mae,patch,logistic,adni_ad_vs_cn,99,0.3593813663804626,test,0.8048780487804879,0.0430785566889253,0.6554621848739496,0.09396380328489887,0.6338709677419355,0.07389995881318914
+flat_mae,patch,logistic,adni_ad_vs_cn,100,0.046415888336127774,train,0.9024390243902439,0.01429182683271312,0.8463069233617179,0.02517205755009012,0.8109335195989811,0.02733477685839864
+flat_mae,patch,logistic,adni_ad_vs_cn,100,0.046415888336127774,test,0.7317073170731707,0.055358095513176074,0.5918552036199095,0.08881939160398047,0.5854838709677419,0.07944307152233746

decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__patch__logistic/log.txt

@@ -0,0 +1,240 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: aef99c83a386cf95c3d8ca503ecc968d8d5694af, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-03-07 21:47:01
+config:
+output_root: experiments/decoders/output
+name_prefix: eval_logistic
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (adni_ad_vs_cn patch logistic)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__patch__logistic
+model: flat_mae
+representation: patch
+dataset: adni_ad_vs_cn
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__patch__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=False, reg_tokens=4, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: adni_ad_vs_cn (flat)
+train (n=328):
+ADNIDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 525
+}),
+    labels=[0 1],
+    counts=[251  77]
+)
+
+validation (n=41):
+ADNIDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 66
+}),
+    labels=[0 1],
+    counts=[31 10]
+)
+
+test (n=41):
+ADNIDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 66
+}),
+    labels=[0 1],
+    counts=[32  9]
+)
+
+extracting features for all splits
+extract (train)  [  0/164]  eta: 0:11:10    time: 4.0894  data: 3.2760  max mem: 2698
+extract (train)  [ 20/164]  eta: 0:00:55    time: 0.1966  data: 0.0707  max mem: 2852
+extract (train)  [ 40/164]  eta: 0:00:33    time: 0.1567  data: 0.0475  max mem: 2852
+extract (train)  [ 60/164]  eta: 0:00:24    time: 0.1638  data: 0.0534  max mem: 2852
+extract (train)  [ 80/164]  eta: 0:00:18    time: 0.1691  data: 0.0549  max mem: 2852
+extract (train)  [100/164]  eta: 0:00:13    time: 0.1675  data: 0.0548  max mem: 2852
+extract (train)  [120/164]  eta: 0:00:08    time: 0.1714  data: 0.0559  max mem: 2852
+extract (train)  [140/164]  eta: 0:00:04    time: 0.1572  data: 0.0492  max mem: 2852
+extract (train)  [160/164]  eta: 0:00:00    time: 0.1522  data: 0.0480  max mem: 2852
+extract (train)  [163/164]  eta: 0:00:00    time: 0.1508  data: 0.0474  max mem: 2852
+extract (train) Total time: 0:00:31 (0.1928 s / it)
+extract (validation)  [ 0/21]  eta: 0:01:13    time: 3.4934  data: 3.3892  max mem: 2852
+extract (validation)  [20/21]  eta: 0:00:00    time: 0.1421  data: 0.0416  max mem: 2852
+extract (validation) Total time: 0:00:06 (0.3166 s / it)
+extract (test)  [ 0/21]  eta: 0:01:12    time: 3.4665  data: 3.3663  max mem: 2852
+extract (test)  [20/21]  eta: 0:00:00    time: 0.1396  data: 0.0420  max mem: 2852
+extract (test) Total time: 0:00:06 (0.3112 s / it)
+feature extraction time: 0:00:44
+train features: (328, 768)
+validation features: (41, 768)
+test features: (41, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset       | trial   |      C | split   |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:--------------|:--------|-------:|:--------|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | adni_ad_vs_cn |         | 166.81 | train   | 1       |  0        | 1       | 0        | 1       |   0        |
+| flat_mae | patch  | logistic | adni_ad_vs_cn |         | 166.81 | test    | 0.65854 |  0.067211 | 0.50174 | 0.076845 | 0.50174 |   0.076542 |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 1, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.06175586637339688, "f1": 0.7119437939110069, "f1_std": 0.07982166095879796, "bacc": 0.7193548387096774, "bacc_std": 0.08516609223998574}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 2, "C": 2.782559402207126, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06428491014860176, "f1": 0.6232247284878863, "f1_std": 0.09162267407571494, "bacc": 0.6193548387096774, "bacc_std": 0.08786441042986236}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 3, "C": 21.54434690031882, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06948437611929524, "f1": 0.5839188134270101, "f1_std": 0.08497702446519155, "bacc": 0.5870967741935484, "bacc_std": 0.08792274414851722}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 4, "C": 166.81005372000556, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.06127466177863133, "f1": 0.783068783068783, "f1_std": 0.05915914345251882, "bacc": 0.8709677419354839, "bacc_std": 0.040520340853611066}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 5, "C": 166.81005372000556, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.0604363274265884, "f1": 0.5547201336675021, "f1_std": 0.0811416666945755, "bacc": 0.5532258064516129, "bacc_std": 0.07772808309192142}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 6, "C": 0.3593813663804626, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06479200235603146, "f1": 0.6232247284878863, "f1_std": 0.08930198197826686, "bacc": 0.6193548387096774, "bacc_std": 0.08677124800133185}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 7, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06372417884650156, "f1": 0.6893939393939394, "f1_std": 0.0784376993864998, "bacc": 0.7032258064516128, "bacc_std": 0.084067210612129}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 8, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06641316068338539, "f1": 0.6479313036690086, "f1_std": 0.08202671334913557, "bacc": 0.6532258064516129, "bacc_std": 0.0854468406709409}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 9, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.0619757777327563, "f1": 0.7119437939110069, "f1_std": 0.08030073846778696, "bacc": 0.7193548387096774, "bacc_std": 0.08475320178801186}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 10, "C": 2.782559402207126, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06399665369003064, "f1": 0.6479313036690086, "f1_std": 0.08319998345571941, "bacc": 0.6532258064516129, "bacc_std": 0.08832318716066874}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 11, "C": 166.81005372000556, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05120557301614695, "f1": 0.6893939393939394, "f1_std": 0.09174013333260418, "bacc": 0.667741935483871, "bacc_std": 0.08241255189397681}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 12, "C": 166.81005372000556, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06106450106001445, "f1": 0.5729166666666666, "f1_std": 0.08751751006546477, "bacc": 0.5693548387096774, "bacc_std": 0.08186769382609813}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 13, "C": 0.046415888336127774, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.05345814020626708, "f1": 0.6117424242424243, "f1_std": 0.09202468610571168, "bacc": 0.6016129032258064, "bacc_std": 0.0791058140916389}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 14, "C": 21.54434690031882, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06369536361170726, "f1": 0.6693548387096775, "f1_std": 0.08657520433598431, "bacc": 0.6693548387096775, "bacc_std": 0.08820304807605202}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 15, "C": 2.782559402207126, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06447931754782824, "f1": 0.6693548387096775, "f1_std": 0.08742264337378126, "bacc": 0.6693548387096775, "bacc_std": 0.08954270364040907}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 16, "C": 166.81005372000556, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.04575690362803435, "f1": 0.7144278606965174, "f1_std": 0.09279895838312108, "bacc": 0.6838709677419355, "bacc_std": 0.07932248822621175}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 17, "C": 1291.5496650148827, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.0711080986208361, "f1": 0.5651515151515152, "f1_std": 0.0852582525873165, "bacc": 0.5709677419354839, "bacc_std": 0.08977297363352293}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 18, "C": 0.3593813663804626, "split": "test", "acc": 0.8780487804878049, "acc_std": 0.03814609194603692, "f1": 0.7960199004975124, "f1_std": 0.08301462416148489, "bacc": 0.75, "bacc_std": 0.07819948848937569}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 19, "C": 1291.5496650148827, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.060457486500875714, "f1": 0.6440972222222222, "f1_std": 0.08494828399370564, "bacc": 0.635483870967742, "bacc_std": 0.08026925623399012}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 20, "C": 2.782559402207126, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.06407793811820159, "f1": 0.764367816091954, "f1_std": 0.07146125035822799, "bacc": 0.8032258064516129, "bacc_std": 0.07456277628052574}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 21, "C": 10000.0, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.058075808423928135, "f1": 0.6440972222222222, "f1_std": 0.08780002288873556, "bacc": 0.635483870967742, "bacc_std": 0.08256684587327397}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 22, "C": 166.81005372000556, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.051874648712642704, "f1": 0.7402714932126697, "f1_std": 0.08889597556100112, "bacc": 0.717741935483871, "bacc_std": 0.08333749263987081}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 23, "C": 1291.5496650148827, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06808896444446465, "f1": 0.6272727272727273, "f1_std": 0.08379512140986728, "bacc": 0.6370967741935484, "bacc_std": 0.0899293051054737}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 24, "C": 0.3593813663804626, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.0730399481610212, "f1": 0.5651515151515152, "f1_std": 0.08570021624212107, "bacc": 0.5709677419354839, "bacc_std": 0.0912322758576622}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 25, "C": 21.54434690031882, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06502429875755188, "f1": 0.6693548387096775, "f1_std": 0.08661123117886423, "bacc": 0.6693548387096775, "bacc_std": 0.08749686576815639}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 26, "C": 2.782559402207126, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06437876837875586, "f1": 0.603225806451613, "f1_std": 0.08250327793555129, "bacc": 0.603225806451613, "bacc_std": 0.08280767299952495}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 27, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06309206892481466, "f1": 0.6232247284878863, "f1_std": 0.08547873759960081, "bacc": 0.6193548387096774, "bacc_std": 0.08351086534453545}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 28, "C": 2.782559402207126, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.061738043058860034, "f1": 0.6232247284878863, "f1_std": 0.08457621269232801, "bacc": 0.6193548387096774, "bacc_std": 0.08213022222486381}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 29, "C": 1291.5496650148827, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.07105567523251301, "f1": 0.6272727272727273, "f1_std": 0.08378732150900603, "bacc": 0.6370967741935484, "bacc_std": 0.08966387946055397}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 30, "C": 0.3593813663804626, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.04778185109921527, "f1": 0.6328358208955224, "f1_std": 0.09319085105882638, "bacc": 0.6177419354838709, "bacc_std": 0.07593114487288175}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 31, "C": 21.54434690031882, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05097528299580182, "f1": 0.5512437810945273, "f1_std": 0.08895511919602032, "bacc": 0.5516129032258065, "bacc_std": 0.07090380837178853}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 32, "C": 0.046415888336127774, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.060572010624942316, "f1": 0.5176470588235295, "f1_std": 0.08190754554917075, "bacc": 0.5193548387096775, "bacc_std": 0.07278577578846016}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 33, "C": 0.3593813663804626, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06147417637105602, "f1": 0.6893939393939394, "f1_std": 0.07571128957091285, "bacc": 0.7032258064516128, "bacc_std": 0.08194897494976601}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 34, "C": 0.3593813663804626, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.062007195759818313, "f1": 0.7119437939110069, "f1_std": 0.07932457979132376, "bacc": 0.7193548387096774, "bacc_std": 0.08408095202384551}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 35, "C": 2.782559402207126, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.05069360670594106, "f1": 0.7402714932126697, "f1_std": 0.08787541773146086, "bacc": 0.717741935483871, "bacc_std": 0.08333689079468769}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 36, "C": 166.81005372000556, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.07224472587268083, "f1": 0.5839188134270101, "f1_std": 0.08441764447486973, "bacc": 0.5870967741935484, "bacc_std": 0.08659274535656608}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 37, "C": 0.3593813663804626, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.05848117295950731, "f1": 0.5729166666666666, "f1_std": 0.0864804969138768, "bacc": 0.5693548387096774, "bacc_std": 0.07968586869134761}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 38, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06711712971906027, "f1": 0.6479313036690086, "f1_std": 0.08419349273094341, "bacc": 0.6532258064516129, "bacc_std": 0.08785933908891233}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 39, "C": 2.782559402207126, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06702252633471767, "f1": 0.5547201336675021, "f1_std": 0.08647402444083507, "bacc": 0.5532258064516129, "bacc_std": 0.08406115972553356}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 40, "C": 0.005994842503189409, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.04141989107375309, "f1": 0.4972129319955407, "f1_std": 0.07434953475605574, "bacc": 0.5177419354838709, "bacc_std": 0.05316787697999455}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 41, "C": 1291.5496650148827, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06787837743012205, "f1": 0.6479313036690086, "f1_std": 0.0851149835833768, "bacc": 0.6532258064516129, "bacc_std": 0.08764715534317276}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 42, "C": 166.81005372000556, "split": "test", "acc": 0.6097560975609756, "acc_std": 0.06637732172945972, "f1": 0.47096774193548385, "f1_std": 0.0727034266598651, "bacc": 0.47096774193548385, "bacc_std": 0.07333956454307135}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 43, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.05879366313225619, "f1": 0.6440972222222222, "f1_std": 0.08787938032120693, "bacc": 0.635483870967742, "bacc_std": 0.08440252251237289}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 44, "C": 1291.5496650148827, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.053685558606160026, "f1": 0.6660633484162897, "f1_std": 0.09088552881025565, "bacc": 0.6516129032258065, "bacc_std": 0.08400442304147263}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 45, "C": 0.005994842503189409, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.03504395132812059, "f1": 0.6800445930880714, "f1_std": 0.09352338117352493, "bacc": 0.65, "bacc_std": 0.07184010022264724}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 46, "C": 21.54434690031882, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.053717838627345894, "f1": 0.7152777777777778, "f1_std": 0.08634601572777, "bacc": 0.7016129032258065, "bacc_std": 0.0851015784674308}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 47, "C": 10000.0, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05180846067259221, "f1": 0.5512437810945273, "f1_std": 0.08654368574712812, "bacc": 0.5516129032258065, "bacc_std": 0.06887771184022792}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 48, "C": 21.54434690031882, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.07177524574147591, "f1": 0.646551724137931, "f1_std": 0.08069310588351807, "bacc": 0.6709677419354838, "bacc_std": 0.09070532385276033}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 49, "C": 0.3593813663804626, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06505340322129365, "f1": 0.5547201336675021, "f1_std": 0.08618118501625521, "bacc": 0.5532258064516129, "bacc_std": 0.08281085000319596}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 50, "C": 21.54434690031882, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05796173017929806, "f1": 0.6917293233082706, "f1_std": 0.082836555047868, "bacc": 0.685483870967742, "bacc_std": 0.08079937068969135}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 51, "C": 1291.5496650148827, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06379389403478483, "f1": 0.603225806451613, "f1_std": 0.0842392921222764, "bacc": 0.603225806451613, "bacc_std": 0.08491606457492491}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 52, "C": 10000.0, "split": "test", "acc": 0.6097560975609756, "acc_std": 0.0723268409477098, "f1": 0.5287356321839081, "f1_std": 0.07849295074799864, "bacc": 0.5387096774193548, "bacc_std": 0.08671881936254502}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 53, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.05463414634146341, "f1": 0.6117424242424243, "f1_std": 0.09245079057319427, "bacc": 0.6016129032258064, "bacc_std": 0.07855340091304261}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 54, "C": 0.005994842503189409, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.043246701792655724, "f1": 0.569327731092437, "f1_std": 0.08661224222477408, "bacc": 0.567741935483871, "bacc_std": 0.06507710139450365}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 55, "C": 166.81005372000556, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.07491260027493644, "f1": 0.6259649122807017, "f1_std": 0.07967230381546084, "bacc": 0.6548387096774193, "bacc_std": 0.08799429854531957}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 56, "C": 2.782559402207126, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06456399322146954, "f1": 0.6693548387096775, "f1_std": 0.08350107478969533, "bacc": 0.6693548387096775, "bacc_std": 0.08552273802962136}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 57, "C": 0.046415888336127774, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06265693681558678, "f1": 0.5729166666666666, "f1_std": 0.08851241615812118, "bacc": 0.5693548387096774, "bacc_std": 0.08031695899487994}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 58, "C": 21.54434690031882, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.06301145755703659, "f1": 0.7119437939110069, "f1_std": 0.08059125331526244, "bacc": 0.7193548387096774, "bacc_std": 0.08491596500908126}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 59, "C": 166.81005372000556, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06620653423560834, "f1": 0.5839188134270101, "f1_std": 0.08236688301326865, "bacc": 0.5870967741935484, "bacc_std": 0.08558086250473963}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 60, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06909857256497813, "f1": 0.6676492262343405, "f1_std": 0.08115322014834417, "bacc": 0.6870967741935483, "bacc_std": 0.08707174096641203}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 61, "C": 2.782559402207126, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06908674250022279, "f1": 0.646551724137931, "f1_std": 0.07872518273436985, "bacc": 0.6709677419354838, "bacc_std": 0.08707151581180167}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 62, "C": 166.81005372000556, "split": "test", "acc": 0.6097560975609756, "acc_std": 0.07572358072732534, "f1": 0.5287356321839081, "f1_std": 0.08277261956763161, "bacc": 0.5387096774193548, "bacc_std": 0.09205784562266699}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 63, "C": 166.81005372000556, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.0734152661831638, "f1": 0.6259649122807017, "f1_std": 0.08025615812395835, "bacc": 0.6548387096774193, "bacc_std": 0.08996184785051084}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 64, "C": 21.54434690031882, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06738905659997667, "f1": 0.6479313036690086, "f1_std": 0.08728104520261751, "bacc": 0.6532258064516129, "bacc_std": 0.09135951418689889}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 65, "C": 0.046415888336127774, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.05693606629763894, "f1": 0.4564393939393939, "f1_std": 0.0689541429669208, "bacc": 0.4693548387096774, "bacc_std": 0.06066977631889546}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 66, "C": 0.046415888336127774, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.05811496534376152, "f1": 0.6440972222222222, "f1_std": 0.0883637958972104, "bacc": 0.635483870967742, "bacc_std": 0.08271046870689876}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 67, "C": 0.3593813663804626, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.04583027426562429, "f1": 0.7670454545454546, "f1_std": 0.08723058539747552, "bacc": 0.7338709677419355, "bacc_std": 0.08246891250438924}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 68, "C": 166.81005372000556, "split": "test", "acc": 0.5853658536585366, "acc_std": 0.07401906598765423, "f1": 0.4863669859985261, "f1_std": 0.08002417651154202, "bacc": 0.4887096774193548, "bacc_std": 0.08731729854014975}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 69, "C": 21.54434690031882, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06311846410995943, "f1": 0.6893939393939394, "f1_std": 0.07791947521483165, "bacc": 0.7032258064516128, "bacc_std": 0.08371906691078246}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 70, "C": 166.81005372000556, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.054348031306029526, "f1": 0.5340909090909092, "f1_std": 0.08348737885347941, "bacc": 0.535483870967742, "bacc_std": 0.07056282217321602}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 71, "C": 0.046415888336127774, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.050431947420682396, "f1": 0.4696517412935323, "f1_std": 0.06966787464617082, "bacc": 0.4854838709677419, "bacc_std": 0.05737906741775723}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 72, "C": 21.54434690031882, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.039054866148776776, "f1": 0.4972129319955407, "f1_std": 0.07117770726827559, "bacc": 0.5177419354838709, "bacc_std": 0.0502263592071848}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 73, "C": 21.54434690031882, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.06222086506928569, "f1": 0.7354838709677419, "f1_std": 0.08379770716347627, "bacc": 0.7354838709677419, "bacc_std": 0.08570259030759712}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 74, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.0591854283720946, "f1": 0.6440972222222222, "f1_std": 0.0879698454449336, "bacc": 0.635483870967742, "bacc_std": 0.08222180832861987}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 75, "C": 2.782559402207126, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.057072649570270084, "f1": 0.7152777777777778, "f1_std": 0.08873922328495465, "bacc": 0.7016129032258065, "bacc_std": 0.08743770525172663}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 76, "C": 21.54434690031882, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06581590607116995, "f1": 0.603225806451613, "f1_std": 0.08495157671395279, "bacc": 0.603225806451613, "bacc_std": 0.08502532376640316}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 77, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05596078766509243, "f1": 0.6660633484162897, "f1_std": 0.09223254185416252, "bacc": 0.6516129032258065, "bacc_std": 0.0852674496561477}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 78, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06717368964562526, "f1": 0.7054597701149425, "f1_std": 0.07507979001414222, "bacc": 0.7370967741935484, "bacc_std": 0.08133086616951436}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 79, "C": 0.3593813663804626, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.05086874081777024, "f1": 0.8016129032258064, "f1_std": 0.07080787710040606, "bacc": 0.8016129032258064, "bacc_std": 0.0760470725436277}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 80, "C": 0.005994842503189409, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.04653287810564446, "f1": 0.569327731092437, "f1_std": 0.09082647728295093, "bacc": 0.567741935483871, "bacc_std": 0.06927126329752148}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 81, "C": 0.005994842503189409, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.03243440275164686, "f1": 0.6095238095238095, "f1_std": 0.09923652315029885, "bacc": 0.6, "bacc_std": 0.06649052564087607}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 82, "C": 2.782559402207126, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.06339316709820725, "f1": 0.7410526315789474, "f1_std": 0.06983750868286032, "bacc": 0.7870967741935484, "bacc_std": 0.07512549583776422}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 83, "C": 1291.5496650148827, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.05954359427574256, "f1": 0.6440972222222222, "f1_std": 0.08815069104446678, "bacc": 0.635483870967742, "bacc_std": 0.0839670799175816}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 84, "C": 0.046415888336127774, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06004955320380165, "f1": 0.5918552036199095, "f1_std": 0.09122675202019988, "bacc": 0.5854838709677419, "bacc_std": 0.08151191635463105}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 85, "C": 0.3593813663804626, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.05588719836515769, "f1": 0.8016129032258064, "f1_std": 0.07605139267403929, "bacc": 0.8016129032258064, "bacc_std": 0.07898267042109322}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 86, "C": 2.782559402207126, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06103458611281899, "f1": 0.5547201336675021, "f1_std": 0.07988054022907731, "bacc": 0.5532258064516129, "bacc_std": 0.0777431176459443}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 87, "C": 0.005994842503189409, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.03149161149082581, "f1": 0.6095238095238095, "f1_std": 0.09722792853269492, "bacc": 0.6, "bacc_std": 0.06455780355619295}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 88, "C": 21.54434690031882, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.07111024026170445, "f1": 0.6620879120879121, "f1_std": 0.0757924017478316, "bacc": 0.7048387096774194, "bacc_std": 0.08472392449535054}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 89, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06625776642881673, "f1": 0.7054597701149425, "f1_std": 0.07523037894605458, "bacc": 0.7370967741935484, "bacc_std": 0.08163465450295516}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 90, "C": 2.782559402207126, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06554580254154335, "f1": 0.6893939393939394, "f1_std": 0.07941124009353895, "bacc": 0.7032258064516128, "bacc_std": 0.08448739199944477}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 91, "C": 166.81005372000556, "split": "test", "acc": 0.6341463414634146, "acc_std": 0.0653331552725744, "f1": 0.5199063231850116, "f1_std": 0.07647618156375047, "bacc": 0.5209677419354839, "bacc_std": 0.0803545641077397}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 92, "C": 0.3593813663804626, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06160824979330037, "f1": 0.5547201336675021, "f1_std": 0.0821168795678255, "bacc": 0.5532258064516129, "bacc_std": 0.07936587020097206}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 93, "C": 2.782559402207126, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06310006406077198, "f1": 0.5547201336675021, "f1_std": 0.08275078503274196, "bacc": 0.5532258064516129, "bacc_std": 0.0796312035222479}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 94, "C": 0.046415888336127774, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.05667016354352588, "f1": 0.4564393939393939, "f1_std": 0.06808356982840677, "bacc": 0.4693548387096774, "bacc_std": 0.05991912204183492}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 95, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05992460065145339, "f1": 0.6917293233082706, "f1_std": 0.08260908925667741, "bacc": 0.685483870967742, "bacc_std": 0.08349758892878209}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 96, "C": 0.3593813663804626, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.0329243359489596, "f1": 0.5119047619047619, "f1_std": 0.07887369477154188, "bacc": 0.5338709677419355, "bacc_std": 0.05146099934604913}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 97, "C": 0.046415888336127774, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.034281405355135894, "f1": 0.4142857142857143, "f1_std": 0.01198068442848582, "bacc": 0.46774193548387094, "bacc_std": 0.022669961605815688}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 98, "C": 0.005994842503189409, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05482840030560244, "f1": 0.5918552036199095, "f1_std": 0.08798511562641391, "bacc": 0.5854838709677419, "bacc_std": 0.07903013367783404}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 99, "C": 0.3593813663804626, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.0430785566889253, "f1": 0.6554621848739496, "f1_std": 0.09396380328489887, "bacc": 0.6338709677419355, "bacc_std": 0.07389995881318914}
+{"model": "flat_mae", "repr": "patch", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 100, "C": 0.046415888336127774, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.055358095513176074, "f1": 0.5918552036199095, "f1_std": 0.08881939160398047, "bacc": 0.5854838709677419, "bacc_std": 0.07944307152233746}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset       | split   |   n_trials |      C |   C_std |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:--------------|:--------|-----------:|-------:|--------:|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | patch  | logistic | adni_ad_vs_cn | train   |        100 | 455.19 |  1721.6 | 0.97444 |  0.051405 | 0.95652 | 0.092261 | 0.94942 |   0.10257  |
+| flat_mae | patch  | logistic | adni_ad_vs_cn | test    |        100 | 455.19 |  1721.6 | 0.73878 |  0.056783 | 0.62697 | 0.081969 | 0.62979 |   0.081878 |
+
+
+done! total time: 0:04:29

decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__reg__logistic/config.yaml

@@ -0,0 +1,30 @@
+output_root: experiments/decoders/output
+name_prefix: eval_logistic
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (adni_ad_vs_cn reg logistic)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__reg__logistic
+model: flat_mae
+representation: reg
+dataset: adni_ad_vs_cn
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__reg__logistic
+remote_dir: null

decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__reg__logistic/eval_table.csv

@@ -0,0 +1,203 @@
+model,repr,clf,dataset,trial,C,split,acc,acc_std,f1,f1_std,bacc,bacc_std
+flat_mae,reg,logistic,adni_ad_vs_cn,,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,,21.54434690031882,test,0.7317073170731707,0.061503103716848896,0.5918552036199095,0.08573707634043229,0.5885416666666666,0.08348856664257005
+flat_mae,reg,logistic,adni_ad_vs_cn,1,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,1,1291.5496650148827,test,0.7804878048780488,0.05965197498965619,0.6917293233082706,0.08421119542537488,0.685483870967742,0.08393276104532095
+flat_mae,reg,logistic,adni_ad_vs_cn,2,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,2,166.81005372000556,test,0.7804878048780488,0.060952795732066516,0.6917293233082706,0.08639343869179594,0.685483870967742,0.08576363132386826
+flat_mae,reg,logistic,adni_ad_vs_cn,3,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,3,21.54434690031882,test,0.6829268292682927,0.06567274155042077,0.5547201336675021,0.087446297612882,0.5532258064516129,0.08359551859045979
+flat_mae,reg,logistic,adni_ad_vs_cn,4,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,4,2.782559402207126,test,0.8780487804878049,0.04999700169475366,0.8287385129490392,0.07348973685417326,0.8177419354838709,0.07690595220053496
+flat_mae,reg,logistic,adni_ad_vs_cn,5,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,5,21.54434690031882,test,0.7560975609756098,0.04297104153345551,0.569327731092437,0.08611760164225671,0.567741935483871,0.06503548219094353
+flat_mae,reg,logistic,adni_ad_vs_cn,6,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,6,1291.5496650148827,test,0.6829268292682927,0.06638268088317434,0.5839188134270101,0.08277286962308099,0.5870967741935484,0.08582816236372913
+flat_mae,reg,logistic,adni_ad_vs_cn,7,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,7,1291.5496650148827,test,0.7317073170731707,0.06694010747894036,0.6479313036690086,0.08462423368279873,0.6532258064516129,0.087403401174901
+flat_mae,reg,logistic,adni_ad_vs_cn,8,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,8,21.54434690031882,test,0.7804878048780488,0.05315549121475575,0.6660633484162897,0.08475383202565291,0.6516129032258065,0.07784234253517273
+flat_mae,reg,logistic,adni_ad_vs_cn,9,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,9,166.81005372000556,test,0.7804878048780488,0.06612365608547857,0.7280766396462786,0.07837400086085841,0.7532258064516129,0.08378285793692576
+flat_mae,reg,logistic,adni_ad_vs_cn,10,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,10,166.81005372000556,test,0.7560975609756098,0.06383834059499562,0.6693548387096775,0.08516590831766294,0.6693548387096775,0.0883888507007518
+flat_mae,reg,logistic,adni_ad_vs_cn,11,0.3593813663804626,train,0.997289972899729,0.0026130283104582475,0.9961941891766453,0.0036965715345039585,0.9941860465116279,0.005605857247436579
+flat_mae,reg,logistic,adni_ad_vs_cn,11,0.3593813663804626,test,0.8048780487804879,0.047657812611521855,0.6893939393939394,0.08683037374702807,0.667741935483871,0.07779467064289844
+flat_mae,reg,logistic,adni_ad_vs_cn,12,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,12,1291.5496650148827,test,0.7317073170731707,0.06135900927358825,0.6232247284878863,0.0847216698151674,0.6193548387096774,0.08315368943466846
+flat_mae,reg,logistic,adni_ad_vs_cn,13,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,13,166.81005372000556,test,0.8048780487804879,0.05510293218186004,0.7152777777777778,0.0854177483880803,0.7016129032258065,0.08431675443257824
+flat_mae,reg,logistic,adni_ad_vs_cn,14,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,14,2.782559402207126,test,0.8048780487804879,0.056206871231079474,0.7152777777777778,0.08764975410828699,0.7016129032258065,0.08555144371272728
+flat_mae,reg,logistic,adni_ad_vs_cn,15,0.3593813663804626,train,0.991869918699187,0.005010253072471148,0.9884880564885973,0.007238768213460641,0.9825581395348837,0.010748740603150363
+flat_mae,reg,logistic,adni_ad_vs_cn,15,0.3593813663804626,test,0.7804878048780488,0.06187635191920471,0.7119437939110069,0.08057680751526211,0.7193548387096774,0.0849466671759876
+flat_mae,reg,logistic,adni_ad_vs_cn,16,0.046415888336127774,train,0.9186991869918699,0.01293309887559064,0.8719224361347648,0.02318690969873031,0.8336757334209877,0.026417961547204792
+flat_mae,reg,logistic,adni_ad_vs_cn,16,0.046415888336127774,test,0.7560975609756098,0.04553796358320358,0.569327731092437,0.0882871089406134,0.567741935483871,0.0665605567188106
+flat_mae,reg,logistic,adni_ad_vs_cn,17,0.3593813663804626,train,0.997289972899729,0.0028031428328590575,0.9961941891766453,0.003970995437953729,0.9941860465116279,0.0060137192170057665
+flat_mae,reg,logistic,adni_ad_vs_cn,17,0.3593813663804626,test,0.8048780487804879,0.05758403837273237,0.7152777777777778,0.0890620732247795,0.7016129032258065,0.08537799286076218
+flat_mae,reg,logistic,adni_ad_vs_cn,18,0.3593813663804626,train,0.997289972899729,0.0025586789432252076,0.9961941891766453,0.0036181555244458206,0.9941860465116279,0.005489258895640111
+flat_mae,reg,logistic,adni_ad_vs_cn,18,0.3593813663804626,test,0.8536585365853658,0.04516901961671339,0.7670454545454546,0.0846177698850047,0.7338709677419355,0.07993524538484041
+flat_mae,reg,logistic,adni_ad_vs_cn,19,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,19,166.81005372000556,test,0.7317073170731707,0.05950747682215606,0.5918552036199095,0.0884153461176964,0.5854838709677419,0.07815662887328642
+flat_mae,reg,logistic,adni_ad_vs_cn,20,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,20,166.81005372000556,test,0.8536585365853658,0.05659327110465348,0.8136363636363637,0.06858781424200065,0.8354838709677419,0.07111708121383084
+flat_mae,reg,logistic,adni_ad_vs_cn,21,0.3593813663804626,train,0.991869918699187,0.004525907531615195,0.9884880564885973,0.00650970253559408,0.9825581395348837,0.009709650460267538
+flat_mae,reg,logistic,adni_ad_vs_cn,21,0.3593813663804626,test,0.8292682926829268,0.04505614185763996,0.7144278606965174,0.0975771118188164,0.6838709677419355,0.08241832907637679
+flat_mae,reg,logistic,adni_ad_vs_cn,22,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,22,21.54434690031882,test,0.6829268292682927,0.06592906252139064,0.5547201336675021,0.08585452757079653,0.5532258064516129,0.08427651307195332
+flat_mae,reg,logistic,adni_ad_vs_cn,23,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,23,166.81005372000556,test,0.7804878048780488,0.059625222535461715,0.6660633484162897,0.09619101759136134,0.6516129032258065,0.08917984868248081
+flat_mae,reg,logistic,adni_ad_vs_cn,24,0.3593813663804626,train,0.994579945799458,0.0037225317364751646,0.9923570836785418,0.005310895002197235,0.9883720930232558,0.007986129132321749
+flat_mae,reg,logistic,adni_ad_vs_cn,24,0.3593813663804626,test,0.7804878048780488,0.05419973197226676,0.6660633484162897,0.09112898780124525,0.6516129032258065,0.08338953256565988
+flat_mae,reg,logistic,adni_ad_vs_cn,25,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,25,166.81005372000556,test,0.7317073170731707,0.05691516590029985,0.5918552036199095,0.08838742606970534,0.5854838709677419,0.07953571380877775
+flat_mae,reg,logistic,adni_ad_vs_cn,26,0.046415888336127774,train,0.9132791327913279,0.012600683008105858,0.861952861952862,0.02278275646182898,0.8220478264442436,0.025635743595965443
+flat_mae,reg,logistic,adni_ad_vs_cn,26,0.046415888336127774,test,0.8048780487804879,0.05156489820554526,0.6893939393939394,0.0923113828438389,0.667741935483871,0.08182558568335797
+flat_mae,reg,logistic,adni_ad_vs_cn,27,0.046415888336127774,train,0.9051490514905149,0.01405086303946133,0.851341551849166,0.024467684791434408,0.8167474730873532,0.02667977666821483
+flat_mae,reg,logistic,adni_ad_vs_cn,27,0.046415888336127774,test,0.8048780487804879,0.049591214128501086,0.6893939393939394,0.08944368214047783,0.667741935483871,0.07884346449681522
+flat_mae,reg,logistic,adni_ad_vs_cn,28,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,28,21.54434690031882,test,0.7804878048780488,0.04781980851961939,0.6328358208955224,0.08942107078387601,0.6177419354838709,0.07419629525069428
+flat_mae,reg,logistic,adni_ad_vs_cn,29,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,29,2.782559402207126,test,0.6829268292682927,0.072968484718037,0.5839188134270101,0.08879885560928677,0.5870967741935484,0.09131133099201774
+flat_mae,reg,logistic,adni_ad_vs_cn,30,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,30,2.782559402207126,test,0.7073170731707317,0.061711057439902794,0.5729166666666666,0.08717477286339652,0.5693548387096774,0.08016214700652118
+flat_mae,reg,logistic,adni_ad_vs_cn,31,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,31,21.54434690031882,test,0.7560975609756098,0.059906014034895344,0.6440972222222222,0.09155689710746522,0.635483870967742,0.08651491778276979
+flat_mae,reg,logistic,adni_ad_vs_cn,32,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,32,1291.5496650148827,test,0.6341463414634146,0.07794651837223623,0.5858585858585859,0.08061991945745695,0.6225806451612903,0.09214856455774029
+flat_mae,reg,logistic,adni_ad_vs_cn,33,0.3593813663804626,train,0.997289972899729,0.003059000035483591,0.9961941891766453,0.004338876103463684,0.9941860465116279,0.006562622169147942
+flat_mae,reg,logistic,adni_ad_vs_cn,33,0.3593813663804626,test,0.6585365853658537,0.07007493227434483,0.5876436781609196,0.08178028971133995,0.6048387096774194,0.09067111756713982
+flat_mae,reg,logistic,adni_ad_vs_cn,34,0.046415888336127774,train,0.9186991869918699,0.012458415195933334,0.8719224361347648,0.02242585939264016,0.8336757334209877,0.025662137996899034
+flat_mae,reg,logistic,adni_ad_vs_cn,34,0.046415888336127774,test,0.7560975609756098,0.06018055324352879,0.6440972222222222,0.0899098719239102,0.635483870967742,0.08489291356876348
+flat_mae,reg,logistic,adni_ad_vs_cn,35,0.3593813663804626,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,35,0.3593813663804626,test,0.7560975609756098,0.059202934956267585,0.6440972222222222,0.08909329156291687,0.635483870967742,0.08434189476955956
+flat_mae,reg,logistic,adni_ad_vs_cn,36,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,36,166.81005372000556,test,0.5853658536585366,0.07539519517241305,0.5108771929824562,0.07668583182122167,0.5225806451612903,0.08555055530507844
+flat_mae,reg,logistic,adni_ad_vs_cn,37,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,37,166.81005372000556,test,0.7560975609756098,0.06205850130941747,0.6693548387096775,0.08402566855862001,0.6693548387096775,0.0869573466434881
+flat_mae,reg,logistic,adni_ad_vs_cn,38,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,38,21.54434690031882,test,0.7317073170731707,0.0706569663986105,0.6676492262343405,0.08392506774076239,0.6870967741935483,0.0901794286677328
+flat_mae,reg,logistic,adni_ad_vs_cn,39,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,39,21.54434690031882,test,0.6829268292682927,0.06415411248486724,0.5547201336675021,0.08445878953795607,0.5532258064516129,0.08042621918609777
+flat_mae,reg,logistic,adni_ad_vs_cn,40,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,40,21.54434690031882,test,0.7317073170731707,0.06689544524766987,0.6479313036690086,0.08348594721613219,0.6532258064516129,0.0866521325829056
+flat_mae,reg,logistic,adni_ad_vs_cn,41,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,41,2.782559402207126,test,0.6341463414634146,0.07105311332898664,0.5684210526315789,0.07619345303127505,0.5887096774193548,0.08778585066810832
+flat_mae,reg,logistic,adni_ad_vs_cn,42,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,42,1291.5496650148827,test,0.8292682926829268,0.051537133652849725,0.7402714932126697,0.08541743756073045,0.717741935483871,0.08257709821011713
+flat_mae,reg,logistic,adni_ad_vs_cn,43,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,43,2.782559402207126,test,0.8048780487804879,0.05821219057439226,0.7152777777777778,0.08983117682872478,0.7016129032258065,0.08668411465551701
+flat_mae,reg,logistic,adni_ad_vs_cn,44,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,44,166.81005372000556,test,0.8292682926829268,0.045681982723045474,0.7144278606965174,0.09295168511434643,0.6838709677419355,0.08154485343507693
+flat_mae,reg,logistic,adni_ad_vs_cn,45,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,45,166.81005372000556,test,0.7804878048780488,0.06406367668062286,0.7280766396462786,0.07657020332749781,0.7532258064516129,0.08305472187794433
+flat_mae,reg,logistic,adni_ad_vs_cn,46,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,46,21.54434690031882,test,0.8048780487804879,0.054857101693279356,0.7354838709677419,0.07788234853373478,0.7354838709677419,0.08207292733066335
+flat_mae,reg,logistic,adni_ad_vs_cn,47,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,47,21.54434690031882,test,0.7073170731707317,0.062241379259203776,0.5729166666666666,0.08683763851162027,0.5693548387096774,0.08206348111286849
+flat_mae,reg,logistic,adni_ad_vs_cn,48,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,48,21.54434690031882,test,0.6829268292682927,0.07145652176863818,0.5839188134270101,0.0870668817515905,0.5870967741935484,0.09073478154889847
+flat_mae,reg,logistic,adni_ad_vs_cn,49,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,49,2.782559402207126,test,0.7317073170731707,0.05199862717265304,0.5512437810945273,0.09039742888627253,0.5516129032258065,0.07205512075237205
+flat_mae,reg,logistic,adni_ad_vs_cn,50,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,50,21.54434690031882,test,0.8292682926829268,0.05878200585024896,0.7759562841530054,0.0766716918114474,0.7854838709677419,0.08059190008953519
+flat_mae,reg,logistic,adni_ad_vs_cn,51,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,51,21.54434690031882,test,0.6829268292682927,0.07209569620348251,0.6259649122807017,0.07865515971011627,0.6548387096774193,0.08860048153245872
+flat_mae,reg,logistic,adni_ad_vs_cn,52,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,52,166.81005372000556,test,0.7560975609756098,0.05413260647593688,0.6117424242424243,0.09191297635254492,0.6016129032258064,0.07879840609637905
+flat_mae,reg,logistic,adni_ad_vs_cn,53,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,53,166.81005372000556,test,0.7073170731707317,0.05852184773826586,0.5729166666666666,0.08533728565060335,0.5693548387096774,0.07863301157837932
+flat_mae,reg,logistic,adni_ad_vs_cn,54,0.046415888336127774,train,0.9186991869918699,0.012219307543989796,0.8719224361347648,0.021768864080769633,0.8336757334209877,0.024718358947797506
+flat_mae,reg,logistic,adni_ad_vs_cn,54,0.046415888336127774,test,0.7073170731707317,0.059563732182869866,0.5729166666666666,0.08375471949209458,0.5693548387096774,0.07763501655854649
+flat_mae,reg,logistic,adni_ad_vs_cn,55,0.3593813663804626,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,55,0.3593813663804626,test,0.7560975609756098,0.059685872125239035,0.6440972222222222,0.09056736965829243,0.635483870967742,0.08535591220401186
+flat_mae,reg,logistic,adni_ad_vs_cn,56,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,56,166.81005372000556,test,0.7317073170731707,0.05257441399144382,0.5512437810945273,0.0882881544227209,0.5516129032258065,0.07113187025279717
+flat_mae,reg,logistic,adni_ad_vs_cn,57,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,57,2.782559402207126,test,0.7317073170731707,0.06456856313739374,0.6479313036690086,0.08245623609374042,0.6532258064516129,0.08598236539668559
+flat_mae,reg,logistic,adni_ad_vs_cn,58,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,58,21.54434690031882,test,0.8048780487804879,0.05875720619796659,0.7152777777777778,0.08769526628535422,0.7016129032258065,0.08484627367237413
+flat_mae,reg,logistic,adni_ad_vs_cn,59,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,59,2.782559402207126,test,0.7804878048780488,0.05655749944618306,0.6660633484162897,0.09321834066404755,0.6516129032258065,0.08649535701180368
+flat_mae,reg,logistic,adni_ad_vs_cn,60,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,60,2.782559402207126,test,0.8780487804878049,0.04847033105175277,0.8287385129490392,0.07333204559007016,0.8177419354838709,0.07820235258642817
+flat_mae,reg,logistic,adni_ad_vs_cn,61,0.046415888336127774,train,0.924119241192412,0.01281720317057954,0.8828571428571428,0.022009622409507105,0.8493508094338073,0.025891173853371845
+flat_mae,reg,logistic,adni_ad_vs_cn,61,0.046415888336127774,test,0.7317073170731707,0.04893874329081201,0.5512437810945273,0.08810481853024506,0.5516129032258065,0.06974457777513789
+flat_mae,reg,logistic,adni_ad_vs_cn,62,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,62,2.782559402207126,test,0.8536585365853658,0.05534459675744239,0.8016129032258064,0.07580692980956008,0.8016129032258064,0.08022866714746085
+flat_mae,reg,logistic,adni_ad_vs_cn,63,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,63,21.54434690031882,test,0.7804878048780488,0.060204272538465095,0.7119437939110069,0.07803112756648595,0.7193548387096774,0.08221201114359575
+flat_mae,reg,logistic,adni_ad_vs_cn,64,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,64,21.54434690031882,test,0.7073170731707317,0.06578200256899232,0.5729166666666666,0.09125109978236295,0.5693548387096774,0.08493307373177764
+flat_mae,reg,logistic,adni_ad_vs_cn,65,0.046415888336127774,train,0.9132791327913279,0.012815745377767435,0.8647732478240953,0.022223798272055943,0.8301421645163941,0.02493051215023349
+flat_mae,reg,logistic,adni_ad_vs_cn,65,0.046415888336127774,test,0.7804878048780488,0.05620416170516529,0.6660633484162897,0.09020495655759604,0.6516129032258065,0.0828187233451739
+flat_mae,reg,logistic,adni_ad_vs_cn,66,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,66,166.81005372000556,test,0.8292682926829268,0.0480918217522762,0.7402714932126697,0.08265546278534666,0.717741935483871,0.07863693756786329
+flat_mae,reg,logistic,adni_ad_vs_cn,67,0.046415888336127774,train,0.9132791327913279,0.01301833836122437,0.861952861952862,0.023584640443634203,0.8220478264442436,0.026463870722474605
+flat_mae,reg,logistic,adni_ad_vs_cn,67,0.046415888336127774,test,0.8292682926829268,0.04484672820105256,0.7144278606965174,0.09250793822282488,0.6838709677419355,0.08073356567395498
+flat_mae,reg,logistic,adni_ad_vs_cn,68,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,68,2.782559402207126,test,0.6829268292682927,0.0628885495214838,0.5547201336675021,0.0833917942345513,0.5532258064516129,0.08051352320793594
+flat_mae,reg,logistic,adni_ad_vs_cn,69,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,69,166.81005372000556,test,0.7804878048780488,0.06387962728192434,0.7119437939110069,0.08194811343093335,0.7193548387096774,0.08543176567816362
+flat_mae,reg,logistic,adni_ad_vs_cn,70,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,70,21.54434690031882,test,0.8048780487804879,0.054270983764463314,0.7152777777777778,0.08579718952741154,0.7016129032258065,0.08392429173353842
+flat_mae,reg,logistic,adni_ad_vs_cn,71,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,71,166.81005372000556,test,0.6585365853658537,0.06914492621255675,0.5651515151515152,0.08211122033999584,0.5709677419354839,0.08671599943350904
+flat_mae,reg,logistic,adni_ad_vs_cn,72,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,72,2.782559402207126,test,0.6829268292682927,0.0389172614930417,0.4057971014492754,0.013931014571946607,0.45161290322580644,0.02573560840668887
+flat_mae,reg,logistic,adni_ad_vs_cn,73,0.046415888336127774,train,0.9105691056910569,0.012475004496142717,0.8568842921784099,0.022953237454800768,0.8162338729558715,0.025369187275369486
+flat_mae,reg,logistic,adni_ad_vs_cn,73,0.046415888336127774,test,0.8536585365853658,0.050436854225144684,0.7864583333333333,0.07938752523206534,0.7677419354838709,0.08293850234780074
+flat_mae,reg,logistic,adni_ad_vs_cn,74,0.046415888336127774,train,0.924119241192412,0.011860292343370915,0.8816765918460834,0.020708989660686794,0.8453036403977319,0.024104130835493856
+flat_mae,reg,logistic,adni_ad_vs_cn,74,0.046415888336127774,test,0.7073170731707317,0.033799070348968036,0.4142857142857143,0.011755025975526823,0.46774193548387094,0.022350998133995006
+flat_mae,reg,logistic,adni_ad_vs_cn,75,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,75,166.81005372000556,test,0.6585365853658537,0.06499851277299176,0.5370967741935484,0.07952403284750914,0.5370967741935484,0.07833145536158176
+flat_mae,reg,logistic,adni_ad_vs_cn,76,0.3593813663804626,train,0.994579945799458,0.003836276687532417,0.9923570836785418,0.005476299414689379,0.9883720930232558,0.00823015173081086
+flat_mae,reg,logistic,adni_ad_vs_cn,76,0.3593813663804626,test,0.7073170731707317,0.05996983179698337,0.5729166666666666,0.08345813132615776,0.5693548387096774,0.0783107529465472
+flat_mae,reg,logistic,adni_ad_vs_cn,77,10000.0,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,77,10000.0,test,0.7560975609756098,0.06449132863698401,0.6893939393939394,0.07994717234162295,0.7032258064516128,0.08617110244729557
+flat_mae,reg,logistic,adni_ad_vs_cn,78,0.3593813663804626,train,0.991869918699187,0.0046970203461552396,0.9884880564885973,0.0067715079588420365,0.9825581395348837,0.01007674713797263
+flat_mae,reg,logistic,adni_ad_vs_cn,78,0.3593813663804626,test,0.8048780487804879,0.05647632471067588,0.7152777777777778,0.08684678224415845,0.7016129032258065,0.0841359872446731
+flat_mae,reg,logistic,adni_ad_vs_cn,79,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,79,21.54434690031882,test,0.9024390243902439,0.04799613309825579,0.8757575757575757,0.05817015566432631,0.9016129032258065,0.05613435820115493
+flat_mae,reg,logistic,adni_ad_vs_cn,80,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,80,21.54434690031882,test,0.6829268292682927,0.06847326299651309,0.5839188134270101,0.08383691342986946,0.5870967741935484,0.08590296754773238
+flat_mae,reg,logistic,adni_ad_vs_cn,81,0.046415888336127774,train,0.9105691056910569,0.012976224421871204,0.859836320314928,0.02294412090442928,0.824328211028022,0.025956560139363236
+flat_mae,reg,logistic,adni_ad_vs_cn,81,0.046415888336127774,test,0.8048780487804879,0.03243440275164686,0.6095238095238095,0.09923652315029885,0.6,0.06649052564087607
+flat_mae,reg,logistic,adni_ad_vs_cn,82,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,82,166.81005372000556,test,0.7560975609756098,0.06358646513593161,0.6693548387096775,0.08301033633603491,0.6693548387096775,0.08329853505150835
+flat_mae,reg,logistic,adni_ad_vs_cn,83,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,83,21.54434690031882,test,0.7317073170731707,0.06351097747137437,0.6479313036690086,0.07913038193372654,0.6532258064516129,0.08292518066262645
+flat_mae,reg,logistic,adni_ad_vs_cn,84,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,84,166.81005372000556,test,0.7804878048780488,0.05483030985904191,0.6660633484162897,0.0902142186547496,0.6516129032258065,0.08328298136108757
+flat_mae,reg,logistic,adni_ad_vs_cn,85,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,85,21.54434690031882,test,0.8292682926829268,0.05048165392011647,0.7402714932126697,0.08745252243279043,0.717741935483871,0.08344078629721811
+flat_mae,reg,logistic,adni_ad_vs_cn,86,1291.5496650148827,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,86,1291.5496650148827,test,0.6097560975609756,0.06851634107748343,0.5030303030303029,0.07481504565310759,0.5048387096774194,0.08021132105241745
+flat_mae,reg,logistic,adni_ad_vs_cn,87,0.046415888336127774,train,0.907859078590786,0.012782138079552238,0.8548454276194002,0.02274282412910148,0.81851425753965,0.025430776500154444
+flat_mae,reg,logistic,adni_ad_vs_cn,87,0.046415888336127774,test,0.8292682926829268,0.04671928140720097,0.7144278606965174,0.09910780874344727,0.6838709677419355,0.08430175372779433
+flat_mae,reg,logistic,adni_ad_vs_cn,88,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,88,2.782559402207126,test,0.7317073170731707,0.06125485327200916,0.6479313036690086,0.07993610002148978,0.6532258064516129,0.08593059773719361
+flat_mae,reg,logistic,adni_ad_vs_cn,89,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,89,2.782559402207126,test,0.8048780487804879,0.06160052457729155,0.7515151515151515,0.07495851243586206,0.7693548387096774,0.08066976263469468
+flat_mae,reg,logistic,adni_ad_vs_cn,90,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,90,166.81005372000556,test,0.7073170731707317,0.06670303192193607,0.603225806451613,0.08847945773979611,0.603225806451613,0.08773534723842394
+flat_mae,reg,logistic,adni_ad_vs_cn,91,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,91,2.782559402207126,test,0.7317073170731707,0.06135949402852056,0.6232247284878863,0.08261580552971312,0.6193548387096774,0.08075232744074175
+flat_mae,reg,logistic,adni_ad_vs_cn,92,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,92,21.54434690031882,test,0.6585365853658537,0.06914828147147303,0.5370967741935484,0.08847396521713383,0.5370967741935484,0.0875774532181361
+flat_mae,reg,logistic,adni_ad_vs_cn,93,0.046415888336127774,train,0.924119241192412,0.01213931827394878,0.8828571428571428,0.020512909281761492,0.8493508094338073,0.02370005916281869
+flat_mae,reg,logistic,adni_ad_vs_cn,93,0.046415888336127774,test,0.7317073170731707,0.06862877230334495,0.6479313036690086,0.08529350536364876,0.6532258064516129,0.08989633635650889
+flat_mae,reg,logistic,adni_ad_vs_cn,94,0.046415888336127774,train,0.9214092140921409,0.011979501836242357,0.8780665671539749,0.020658386205241637,0.8435368559454351,0.023667072950013735
+flat_mae,reg,logistic,adni_ad_vs_cn,94,0.046415888336127774,test,0.7317073170731707,0.04362797676870197,0.4972129319955407,0.07627781221575902,0.5177419354838709,0.05503232084101854
+flat_mae,reg,logistic,adni_ad_vs_cn,95,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,95,166.81005372000556,test,0.7317073170731707,0.059721722317740165,0.5918552036199095,0.09019059005179225,0.5854838709677419,0.08060899731078766
+flat_mae,reg,logistic,adni_ad_vs_cn,96,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,96,21.54434690031882,test,0.7317073170731707,0.05803711709021023,0.5918552036199095,0.0865822025837848,0.5854838709677419,0.07840331466373182
+flat_mae,reg,logistic,adni_ad_vs_cn,97,166.81005372000556,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,97,166.81005372000556,test,0.7560975609756098,0.05908273673327025,0.6440972222222222,0.08791019771882562,0.635483870967742,0.0844447270412904
+flat_mae,reg,logistic,adni_ad_vs_cn,98,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,98,2.782559402207126,test,0.7073170731707317,0.06181663891797104,0.5729166666666666,0.08532454535139765,0.5693548387096774,0.07922126779843966
+flat_mae,reg,logistic,adni_ad_vs_cn,99,21.54434690031882,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,99,21.54434690031882,test,0.7073170731707317,0.05934875598028067,0.5729166666666666,0.08340847340715268,0.5693548387096774,0.07742467582521038
+flat_mae,reg,logistic,adni_ad_vs_cn,100,2.782559402207126,train,1.0,0.0,1.0,0.0,1.0,0.0
+flat_mae,reg,logistic,adni_ad_vs_cn,100,2.782559402207126,test,0.7317073170731707,0.06114082981765291,0.6232247284878863,0.08670443589806903,0.6193548387096774,0.08472381745729438

decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__reg__logistic/log.txt

@@ -0,0 +1,240 @@
+fMRI foundation model logistic probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: aef99c83a386cf95c3d8ca503ecc968d8d5694af, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-03-07 21:19:26
+config:
+output_root: experiments/decoders/output
+name_prefix: eval_logistic
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (adni_ad_vs_cn reg logistic)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+num_workers: 16
+batch_size: 2
+cv_folds: 5
+max_iter: 1000
+Cs: 10
+balanced_sampling: false
+metrics:
+- acc
+- f1
+- bacc
+cv_metric: bacc
+n_trials: 100
+amp: true
+device: cuda
+seed: 4466
+debug: false
+name: decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__reg__logistic
+model: flat_mae
+representation: reg
+dataset: adni_ad_vs_cn
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/adni_ad_vs_cn__reg__logistic
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=False, reg_tokens=4, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: adni_ad_vs_cn (flat)
+train (n=328):
+ADNIDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 525
+}),
+    labels=[0 1],
+    counts=[251  77]
+)
+
+validation (n=41):
+ADNIDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 66
+}),
+    labels=[0 1],
+    counts=[31 10]
+)
+
+test (n=41):
+ADNIDataset(
+    dataset=Dataset({
+    features: ['sub', 'visit', 'mod', 'task', 'path', 'start', 'end', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 66
+}),
+    labels=[0 1],
+    counts=[32  9]
+)
+
+extracting features for all splits
+extract (train)  [  0/164]  eta: 0:12:27    time: 4.5608  data: 3.7108  max mem: 2698
+extract (train)  [ 20/164]  eta: 0:00:57    time: 0.1878  data: 0.0668  max mem: 2852
+extract (train)  [ 40/164]  eta: 0:00:34    time: 0.1627  data: 0.0509  max mem: 2852
+extract (train)  [ 60/164]  eta: 0:00:25    time: 0.1627  data: 0.0532  max mem: 2852
+extract (train)  [ 80/164]  eta: 0:00:18    time: 0.1618  data: 0.0522  max mem: 2852
+extract (train)  [100/164]  eta: 0:00:13    time: 0.1761  data: 0.0607  max mem: 2852
+extract (train)  [120/164]  eta: 0:00:09    time: 0.1717  data: 0.0566  max mem: 2852
+extract (train)  [140/164]  eta: 0:00:04    time: 0.1746  data: 0.0569  max mem: 2852
+extract (train)  [160/164]  eta: 0:00:00    time: 0.1652  data: 0.0530  max mem: 2852
+extract (train)  [163/164]  eta: 0:00:00    time: 0.1640  data: 0.0527  max mem: 2852
+extract (train) Total time: 0:00:33 (0.2024 s / it)
+extract (validation)  [ 0/21]  eta: 0:01:24    time: 4.0446  data: 3.9344  max mem: 2852
+extract (validation)  [20/21]  eta: 0:00:00    time: 0.1479  data: 0.0418  max mem: 2852
+extract (validation) Total time: 0:00:07 (0.3533 s / it)
+extract (test)  [ 0/21]  eta: 0:01:21    time: 3.9047  data: 3.7845  max mem: 2852
+extract (test)  [20/21]  eta: 0:00:00    time: 0.1374  data: 0.0368  max mem: 2852
+extract (test) Total time: 0:00:06 (0.3303 s / it)
+feature extraction time: 0:00:47
+train features: (328, 768)
+validation features: (41, 768)
+test features: (41, 768)
+evaluating fixed splits
+eval results (fixed splits):
+
+| model    | repr   | clf      | dataset       | trial   |      C | split   |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:--------------|:--------|-------:|:--------|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | reg    | logistic | adni_ad_vs_cn |         | 21.544 | train   | 1       |  0        | 1       | 0        | 1       |   0        |
+| flat_mae | reg    | logistic | adni_ad_vs_cn |         | 21.544 | test    | 0.73171 |  0.061503 | 0.59186 | 0.085737 | 0.58854 |   0.083489 |
+
+
+evaluating random splits (n=100)
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 1, "C": 1291.5496650148827, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05965197498965619, "f1": 0.6917293233082706, "f1_std": 0.08421119542537488, "bacc": 0.685483870967742, "bacc_std": 0.08393276104532095}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 2, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.060952795732066516, "f1": 0.6917293233082706, "f1_std": 0.08639343869179594, "bacc": 0.685483870967742, "bacc_std": 0.08576363132386826}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 3, "C": 21.54434690031882, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06567274155042077, "f1": 0.5547201336675021, "f1_std": 0.087446297612882, "bacc": 0.5532258064516129, "bacc_std": 0.08359551859045979}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 4, "C": 2.782559402207126, "split": "test", "acc": 0.8780487804878049, "acc_std": 0.04999700169475366, "f1": 0.8287385129490392, "f1_std": 0.07348973685417326, "bacc": 0.8177419354838709, "bacc_std": 0.07690595220053496}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 5, "C": 21.54434690031882, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.04297104153345551, "f1": 0.569327731092437, "f1_std": 0.08611760164225671, "bacc": 0.567741935483871, "bacc_std": 0.06503548219094353}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 6, "C": 1291.5496650148827, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06638268088317434, "f1": 0.5839188134270101, "f1_std": 0.08277286962308099, "bacc": 0.5870967741935484, "bacc_std": 0.08582816236372913}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 7, "C": 1291.5496650148827, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06694010747894036, "f1": 0.6479313036690086, "f1_std": 0.08462423368279873, "bacc": 0.6532258064516129, "bacc_std": 0.087403401174901}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 8, "C": 21.54434690031882, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05315549121475575, "f1": 0.6660633484162897, "f1_std": 0.08475383202565291, "bacc": 0.6516129032258065, "bacc_std": 0.07784234253517273}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 9, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.06612365608547857, "f1": 0.7280766396462786, "f1_std": 0.07837400086085841, "bacc": 0.7532258064516129, "bacc_std": 0.08378285793692576}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 10, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06383834059499562, "f1": 0.6693548387096775, "f1_std": 0.08516590831766294, "bacc": 0.6693548387096775, "bacc_std": 0.0883888507007518}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 11, "C": 0.3593813663804626, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.047657812611521855, "f1": 0.6893939393939394, "f1_std": 0.08683037374702807, "bacc": 0.667741935483871, "bacc_std": 0.07779467064289844}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 12, "C": 1291.5496650148827, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06135900927358825, "f1": 0.6232247284878863, "f1_std": 0.0847216698151674, "bacc": 0.6193548387096774, "bacc_std": 0.08315368943466846}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 13, "C": 166.81005372000556, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05510293218186004, "f1": 0.7152777777777778, "f1_std": 0.0854177483880803, "bacc": 0.7016129032258065, "bacc_std": 0.08431675443257824}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 14, "C": 2.782559402207126, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.056206871231079474, "f1": 0.7152777777777778, "f1_std": 0.08764975410828699, "bacc": 0.7016129032258065, "bacc_std": 0.08555144371272728}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 15, "C": 0.3593813663804626, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.06187635191920471, "f1": 0.7119437939110069, "f1_std": 0.08057680751526211, "bacc": 0.7193548387096774, "bacc_std": 0.0849466671759876}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 16, "C": 0.046415888336127774, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.04553796358320358, "f1": 0.569327731092437, "f1_std": 0.0882871089406134, "bacc": 0.567741935483871, "bacc_std": 0.0665605567188106}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 17, "C": 0.3593813663804626, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05758403837273237, "f1": 0.7152777777777778, "f1_std": 0.0890620732247795, "bacc": 0.7016129032258065, "bacc_std": 0.08537799286076218}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 18, "C": 0.3593813663804626, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.04516901961671339, "f1": 0.7670454545454546, "f1_std": 0.0846177698850047, "bacc": 0.7338709677419355, "bacc_std": 0.07993524538484041}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 19, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05950747682215606, "f1": 0.5918552036199095, "f1_std": 0.0884153461176964, "bacc": 0.5854838709677419, "bacc_std": 0.07815662887328642}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 20, "C": 166.81005372000556, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.05659327110465348, "f1": 0.8136363636363637, "f1_std": 0.06858781424200065, "bacc": 0.8354838709677419, "bacc_std": 0.07111708121383084}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 21, "C": 0.3593813663804626, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.04505614185763996, "f1": 0.7144278606965174, "f1_std": 0.0975771118188164, "bacc": 0.6838709677419355, "bacc_std": 0.08241832907637679}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 22, "C": 21.54434690031882, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06592906252139064, "f1": 0.5547201336675021, "f1_std": 0.08585452757079653, "bacc": 0.5532258064516129, "bacc_std": 0.08427651307195332}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 23, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.059625222535461715, "f1": 0.6660633484162897, "f1_std": 0.09619101759136134, "bacc": 0.6516129032258065, "bacc_std": 0.08917984868248081}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 24, "C": 0.3593813663804626, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05419973197226676, "f1": 0.6660633484162897, "f1_std": 0.09112898780124525, "bacc": 0.6516129032258065, "bacc_std": 0.08338953256565988}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 25, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05691516590029985, "f1": 0.5918552036199095, "f1_std": 0.08838742606970534, "bacc": 0.5854838709677419, "bacc_std": 0.07953571380877775}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 26, "C": 0.046415888336127774, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05156489820554526, "f1": 0.6893939393939394, "f1_std": 0.0923113828438389, "bacc": 0.667741935483871, "bacc_std": 0.08182558568335797}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 27, "C": 0.046415888336127774, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.049591214128501086, "f1": 0.6893939393939394, "f1_std": 0.08944368214047783, "bacc": 0.667741935483871, "bacc_std": 0.07884346449681522}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 28, "C": 21.54434690031882, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.04781980851961939, "f1": 0.6328358208955224, "f1_std": 0.08942107078387601, "bacc": 0.6177419354838709, "bacc_std": 0.07419629525069428}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 29, "C": 2.782559402207126, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.072968484718037, "f1": 0.5839188134270101, "f1_std": 0.08879885560928677, "bacc": 0.5870967741935484, "bacc_std": 0.09131133099201774}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 30, "C": 2.782559402207126, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.061711057439902794, "f1": 0.5729166666666666, "f1_std": 0.08717477286339652, "bacc": 0.5693548387096774, "bacc_std": 0.08016214700652118}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 31, "C": 21.54434690031882, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.059906014034895344, "f1": 0.6440972222222222, "f1_std": 0.09155689710746522, "bacc": 0.635483870967742, "bacc_std": 0.08651491778276979}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 32, "C": 1291.5496650148827, "split": "test", "acc": 0.6341463414634146, "acc_std": 0.07794651837223623, "f1": 0.5858585858585859, "f1_std": 0.08061991945745695, "bacc": 0.6225806451612903, "bacc_std": 0.09214856455774029}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 33, "C": 0.3593813663804626, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.07007493227434483, "f1": 0.5876436781609196, "f1_std": 0.08178028971133995, "bacc": 0.6048387096774194, "bacc_std": 0.09067111756713982}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 34, "C": 0.046415888336127774, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06018055324352879, "f1": 0.6440972222222222, "f1_std": 0.0899098719239102, "bacc": 0.635483870967742, "bacc_std": 0.08489291356876348}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 35, "C": 0.3593813663804626, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.059202934956267585, "f1": 0.6440972222222222, "f1_std": 0.08909329156291687, "bacc": 0.635483870967742, "bacc_std": 0.08434189476955956}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 36, "C": 166.81005372000556, "split": "test", "acc": 0.5853658536585366, "acc_std": 0.07539519517241305, "f1": 0.5108771929824562, "f1_std": 0.07668583182122167, "bacc": 0.5225806451612903, "bacc_std": 0.08555055530507844}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 37, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06205850130941747, "f1": 0.6693548387096775, "f1_std": 0.08402566855862001, "bacc": 0.6693548387096775, "bacc_std": 0.0869573466434881}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 38, "C": 21.54434690031882, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.0706569663986105, "f1": 0.6676492262343405, "f1_std": 0.08392506774076239, "bacc": 0.6870967741935483, "bacc_std": 0.0901794286677328}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 39, "C": 21.54434690031882, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06415411248486724, "f1": 0.5547201336675021, "f1_std": 0.08445878953795607, "bacc": 0.5532258064516129, "bacc_std": 0.08042621918609777}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 40, "C": 21.54434690031882, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06689544524766987, "f1": 0.6479313036690086, "f1_std": 0.08348594721613219, "bacc": 0.6532258064516129, "bacc_std": 0.0866521325829056}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 41, "C": 2.782559402207126, "split": "test", "acc": 0.6341463414634146, "acc_std": 0.07105311332898664, "f1": 0.5684210526315789, "f1_std": 0.07619345303127505, "bacc": 0.5887096774193548, "bacc_std": 0.08778585066810832}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 42, "C": 1291.5496650148827, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.051537133652849725, "f1": 0.7402714932126697, "f1_std": 0.08541743756073045, "bacc": 0.717741935483871, "bacc_std": 0.08257709821011713}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 43, "C": 2.782559402207126, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05821219057439226, "f1": 0.7152777777777778, "f1_std": 0.08983117682872478, "bacc": 0.7016129032258065, "bacc_std": 0.08668411465551701}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 44, "C": 166.81005372000556, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.045681982723045474, "f1": 0.7144278606965174, "f1_std": 0.09295168511434643, "bacc": 0.6838709677419355, "bacc_std": 0.08154485343507693}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 45, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.06406367668062286, "f1": 0.7280766396462786, "f1_std": 0.07657020332749781, "bacc": 0.7532258064516129, "bacc_std": 0.08305472187794433}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 46, "C": 21.54434690031882, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.054857101693279356, "f1": 0.7354838709677419, "f1_std": 0.07788234853373478, "bacc": 0.7354838709677419, "bacc_std": 0.08207292733066335}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 47, "C": 21.54434690031882, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.062241379259203776, "f1": 0.5729166666666666, "f1_std": 0.08683763851162027, "bacc": 0.5693548387096774, "bacc_std": 0.08206348111286849}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 48, "C": 21.54434690031882, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.07145652176863818, "f1": 0.5839188134270101, "f1_std": 0.0870668817515905, "bacc": 0.5870967741935484, "bacc_std": 0.09073478154889847}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 49, "C": 2.782559402207126, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05199862717265304, "f1": 0.5512437810945273, "f1_std": 0.09039742888627253, "bacc": 0.5516129032258065, "bacc_std": 0.07205512075237205}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 50, "C": 21.54434690031882, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.05878200585024896, "f1": 0.7759562841530054, "f1_std": 0.0766716918114474, "bacc": 0.7854838709677419, "bacc_std": 0.08059190008953519}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 51, "C": 21.54434690031882, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.07209569620348251, "f1": 0.6259649122807017, "f1_std": 0.07865515971011627, "bacc": 0.6548387096774193, "bacc_std": 0.08860048153245872}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 52, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.05413260647593688, "f1": 0.6117424242424243, "f1_std": 0.09191297635254492, "bacc": 0.6016129032258064, "bacc_std": 0.07879840609637905}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 53, "C": 166.81005372000556, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.05852184773826586, "f1": 0.5729166666666666, "f1_std": 0.08533728565060335, "bacc": 0.5693548387096774, "bacc_std": 0.07863301157837932}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 54, "C": 0.046415888336127774, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.059563732182869866, "f1": 0.5729166666666666, "f1_std": 0.08375471949209458, "bacc": 0.5693548387096774, "bacc_std": 0.07763501655854649}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 55, "C": 0.3593813663804626, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.059685872125239035, "f1": 0.6440972222222222, "f1_std": 0.09056736965829243, "bacc": 0.635483870967742, "bacc_std": 0.08535591220401186}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 56, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05257441399144382, "f1": 0.5512437810945273, "f1_std": 0.0882881544227209, "bacc": 0.5516129032258065, "bacc_std": 0.07113187025279717}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 57, "C": 2.782559402207126, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06456856313739374, "f1": 0.6479313036690086, "f1_std": 0.08245623609374042, "bacc": 0.6532258064516129, "bacc_std": 0.08598236539668559}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 58, "C": 21.54434690031882, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05875720619796659, "f1": 0.7152777777777778, "f1_std": 0.08769526628535422, "bacc": 0.7016129032258065, "bacc_std": 0.08484627367237413}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 59, "C": 2.782559402207126, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05655749944618306, "f1": 0.6660633484162897, "f1_std": 0.09321834066404755, "bacc": 0.6516129032258065, "bacc_std": 0.08649535701180368}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 60, "C": 2.782559402207126, "split": "test", "acc": 0.8780487804878049, "acc_std": 0.04847033105175277, "f1": 0.8287385129490392, "f1_std": 0.07333204559007016, "bacc": 0.8177419354838709, "bacc_std": 0.07820235258642817}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 61, "C": 0.046415888336127774, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.04893874329081201, "f1": 0.5512437810945273, "f1_std": 0.08810481853024506, "bacc": 0.5516129032258065, "bacc_std": 0.06974457777513789}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 62, "C": 2.782559402207126, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.05534459675744239, "f1": 0.8016129032258064, "f1_std": 0.07580692980956008, "bacc": 0.8016129032258064, "bacc_std": 0.08022866714746085}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 63, "C": 21.54434690031882, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.060204272538465095, "f1": 0.7119437939110069, "f1_std": 0.07803112756648595, "bacc": 0.7193548387096774, "bacc_std": 0.08221201114359575}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 64, "C": 21.54434690031882, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06578200256899232, "f1": 0.5729166666666666, "f1_std": 0.09125109978236295, "bacc": 0.5693548387096774, "bacc_std": 0.08493307373177764}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 65, "C": 0.046415888336127774, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05620416170516529, "f1": 0.6660633484162897, "f1_std": 0.09020495655759604, "bacc": 0.6516129032258065, "bacc_std": 0.0828187233451739}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 66, "C": 166.81005372000556, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.0480918217522762, "f1": 0.7402714932126697, "f1_std": 0.08265546278534666, "bacc": 0.717741935483871, "bacc_std": 0.07863693756786329}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 67, "C": 0.046415888336127774, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.04484672820105256, "f1": 0.7144278606965174, "f1_std": 0.09250793822282488, "bacc": 0.6838709677419355, "bacc_std": 0.08073356567395498}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 68, "C": 2.782559402207126, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.0628885495214838, "f1": 0.5547201336675021, "f1_std": 0.0833917942345513, "bacc": 0.5532258064516129, "bacc_std": 0.08051352320793594}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 69, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.06387962728192434, "f1": 0.7119437939110069, "f1_std": 0.08194811343093335, "bacc": 0.7193548387096774, "bacc_std": 0.08543176567816362}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 70, "C": 21.54434690031882, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.054270983764463314, "f1": 0.7152777777777778, "f1_std": 0.08579718952741154, "bacc": 0.7016129032258065, "bacc_std": 0.08392429173353842}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 71, "C": 166.81005372000556, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.06914492621255675, "f1": 0.5651515151515152, "f1_std": 0.08211122033999584, "bacc": 0.5709677419354839, "bacc_std": 0.08671599943350904}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 72, "C": 2.782559402207126, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.0389172614930417, "f1": 0.4057971014492754, "f1_std": 0.013931014571946607, "bacc": 0.45161290322580644, "bacc_std": 0.02573560840668887}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 73, "C": 0.046415888336127774, "split": "test", "acc": 0.8536585365853658, "acc_std": 0.050436854225144684, "f1": 0.7864583333333333, "f1_std": 0.07938752523206534, "bacc": 0.7677419354838709, "bacc_std": 0.08293850234780074}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 74, "C": 0.046415888336127774, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.033799070348968036, "f1": 0.4142857142857143, "f1_std": 0.011755025975526823, "bacc": 0.46774193548387094, "bacc_std": 0.022350998133995006}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 75, "C": 166.81005372000556, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.06499851277299176, "f1": 0.5370967741935484, "f1_std": 0.07952403284750914, "bacc": 0.5370967741935484, "bacc_std": 0.07833145536158176}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 76, "C": 0.3593813663804626, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.05996983179698337, "f1": 0.5729166666666666, "f1_std": 0.08345813132615776, "bacc": 0.5693548387096774, "bacc_std": 0.0783107529465472}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 77, "C": 10000.0, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06449132863698401, "f1": 0.6893939393939394, "f1_std": 0.07994717234162295, "bacc": 0.7032258064516128, "bacc_std": 0.08617110244729557}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 78, "C": 0.3593813663804626, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.05647632471067588, "f1": 0.7152777777777778, "f1_std": 0.08684678224415845, "bacc": 0.7016129032258065, "bacc_std": 0.0841359872446731}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 79, "C": 21.54434690031882, "split": "test", "acc": 0.9024390243902439, "acc_std": 0.04799613309825579, "f1": 0.8757575757575757, "f1_std": 0.05817015566432631, "bacc": 0.9016129032258065, "bacc_std": 0.05613435820115493}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 80, "C": 21.54434690031882, "split": "test", "acc": 0.6829268292682927, "acc_std": 0.06847326299651309, "f1": 0.5839188134270101, "f1_std": 0.08383691342986946, "bacc": 0.5870967741935484, "bacc_std": 0.08590296754773238}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 81, "C": 0.046415888336127774, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.03243440275164686, "f1": 0.6095238095238095, "f1_std": 0.09923652315029885, "bacc": 0.6, "bacc_std": 0.06649052564087607}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 82, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.06358646513593161, "f1": 0.6693548387096775, "f1_std": 0.08301033633603491, "bacc": 0.6693548387096775, "bacc_std": 0.08329853505150835}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 83, "C": 21.54434690031882, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06351097747137437, "f1": 0.6479313036690086, "f1_std": 0.07913038193372654, "bacc": 0.6532258064516129, "bacc_std": 0.08292518066262645}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 84, "C": 166.81005372000556, "split": "test", "acc": 0.7804878048780488, "acc_std": 0.05483030985904191, "f1": 0.6660633484162897, "f1_std": 0.0902142186547496, "bacc": 0.6516129032258065, "bacc_std": 0.08328298136108757}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 85, "C": 21.54434690031882, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.05048165392011647, "f1": 0.7402714932126697, "f1_std": 0.08745252243279043, "bacc": 0.717741935483871, "bacc_std": 0.08344078629721811}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 86, "C": 1291.5496650148827, "split": "test", "acc": 0.6097560975609756, "acc_std": 0.06851634107748343, "f1": 0.5030303030303029, "f1_std": 0.07481504565310759, "bacc": 0.5048387096774194, "bacc_std": 0.08021132105241745}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 87, "C": 0.046415888336127774, "split": "test", "acc": 0.8292682926829268, "acc_std": 0.04671928140720097, "f1": 0.7144278606965174, "f1_std": 0.09910780874344727, "bacc": 0.6838709677419355, "bacc_std": 0.08430175372779433}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 88, "C": 2.782559402207126, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06125485327200916, "f1": 0.6479313036690086, "f1_std": 0.07993610002148978, "bacc": 0.6532258064516129, "bacc_std": 0.08593059773719361}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 89, "C": 2.782559402207126, "split": "test", "acc": 0.8048780487804879, "acc_std": 0.06160052457729155, "f1": 0.7515151515151515, "f1_std": 0.07495851243586206, "bacc": 0.7693548387096774, "bacc_std": 0.08066976263469468}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 90, "C": 166.81005372000556, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06670303192193607, "f1": 0.603225806451613, "f1_std": 0.08847945773979611, "bacc": 0.603225806451613, "bacc_std": 0.08773534723842394}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 91, "C": 2.782559402207126, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06135949402852056, "f1": 0.6232247284878863, "f1_std": 0.08261580552971312, "bacc": 0.6193548387096774, "bacc_std": 0.08075232744074175}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 92, "C": 21.54434690031882, "split": "test", "acc": 0.6585365853658537, "acc_std": 0.06914828147147303, "f1": 0.5370967741935484, "f1_std": 0.08847396521713383, "bacc": 0.5370967741935484, "bacc_std": 0.0875774532181361}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 93, "C": 0.046415888336127774, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06862877230334495, "f1": 0.6479313036690086, "f1_std": 0.08529350536364876, "bacc": 0.6532258064516129, "bacc_std": 0.08989633635650889}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 94, "C": 0.046415888336127774, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.04362797676870197, "f1": 0.4972129319955407, "f1_std": 0.07627781221575902, "bacc": 0.5177419354838709, "bacc_std": 0.05503232084101854}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 95, "C": 166.81005372000556, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.059721722317740165, "f1": 0.5918552036199095, "f1_std": 0.09019059005179225, "bacc": 0.5854838709677419, "bacc_std": 0.08060899731078766}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 96, "C": 21.54434690031882, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.05803711709021023, "f1": 0.5918552036199095, "f1_std": 0.0865822025837848, "bacc": 0.5854838709677419, "bacc_std": 0.07840331466373182}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 97, "C": 166.81005372000556, "split": "test", "acc": 0.7560975609756098, "acc_std": 0.05908273673327025, "f1": 0.6440972222222222, "f1_std": 0.08791019771882562, "bacc": 0.635483870967742, "bacc_std": 0.0844447270412904}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 98, "C": 2.782559402207126, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.06181663891797104, "f1": 0.5729166666666666, "f1_std": 0.08532454535139765, "bacc": 0.5693548387096774, "bacc_std": 0.07922126779843966}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 99, "C": 21.54434690031882, "split": "test", "acc": 0.7073170731707317, "acc_std": 0.05934875598028067, "f1": 0.5729166666666666, "f1_std": 0.08340847340715268, "bacc": 0.5693548387096774, "bacc_std": 0.07742467582521038}
+{"model": "flat_mae", "repr": "reg", "clf": "logistic", "dataset": "adni_ad_vs_cn", "trial": 100, "C": 2.782559402207126, "split": "test", "acc": 0.7317073170731707, "acc_std": 0.06114082981765291, "f1": 0.6232247284878863, "f1_std": 0.08670443589806903, "bacc": 0.6193548387096774, "bacc_std": 0.08472381745729438}
+eval results (random splits):
+
+| model    | repr   | clf      | dataset       | split   |   n_trials |      C |   C_std |     acc |   acc_std |      f1 |   f1_std |    bacc |   bacc_std |
+|:---------|:-------|:---------|:--------------|:--------|-----------:|-------:|--------:|--------:|----------:|--------:|---------:|--------:|-----------:|
+| flat_mae | reg    | logistic | adni_ad_vs_cn | train   |        100 | 236.38 |  1038.5 | 0.98778 |  0.029245 | 0.98088 | 0.045972 | 0.9754  |   0.058683 |
+| flat_mae | reg    | logistic | adni_ad_vs_cn | test    |        100 | 236.38 |  1038.5 | 0.75341 |  0.061712 | 0.64527 | 0.085747 | 0.64319 |   0.08146  |
+
+
+done! total time: 0:04:45

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__attn/config.yaml

@@ -0,0 +1,96 @@
+output_root: experiments/decoders/output
+name_prefix: eval_probe
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (hcpya_task21 patch attn)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: false
+  norm: false
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__attn
+model: flat_mae
+representation: patch
+classifier: attn
+dataset: hcpya_task21
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__attn
+remote_dir: null

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__attn/eval_log.json

@@ -0,0 +1 @@
+{"eval/epoch": 17, "eval/id_best": 43, "eval/lr_best": 0.006599999999999999, "eval/wd_best": 0.05, "eval/train/loss": 0.0001459822669858113, "eval/train/acc": 1.0, "eval/train/acc_std": 0.0, "eval/train/f1": 1.0, "eval/train/f1_std": 0.0, "eval/validation/loss": 0.0525173656642437, "eval/validation/acc": 0.9880952380952381, "eval/validation/acc_std": 0.0016299318625606525, "eval/validation/f1": 0.9857562615366776, "eval/validation/f1_std": 0.0022326693782531346, "eval/test/loss": 0.07207445800304413, "eval/test/acc": 0.9861111111111112, "eval/test/acc_std": 0.0015754859836431907, "eval/test/f1": 0.983203683100717, "eval/test/f1_std": 0.002095390680912686}

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__attn/eval_log_best.json

@@ -0,0 +1 @@
+{"eval/best/epoch": 17, "eval/best/id_best": 43, "eval/best/lr_best": 0.006599999999999999, "eval/best/wd_best": 0.05, "eval/best/train/loss": 0.0001459822669858113, "eval/best/train/acc": 1.0, "eval/best/train/acc_std": 0.0, "eval/best/train/f1": 1.0, "eval/best/train/f1_std": 0.0, "eval/best/validation/loss": 0.0525173656642437, "eval/best/validation/acc": 0.9880952380952381, "eval/best/validation/acc_std": 0.0016299318625606525, "eval/best/validation/f1": 0.9857562615366776, "eval/best/validation/f1_std": 0.0022326693782531346, "eval/best/test/loss": 0.07207445800304413, "eval/best/test/acc": 0.9861111111111112, "eval/best/test/acc_std": 0.0015754859836431907, "eval/best/test/f1": 0.983203683100717, "eval/best/test/f1_std": 0.002095390680912686}

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__attn/eval_log_last.json

@@ -0,0 +1 @@
+{"eval/last/epoch": 19, "eval/last/id_best": 43, "eval/last/lr_best": 0.006599999999999999, "eval/last/wd_best": 0.05, "eval/last/train/loss": 0.00014662370085716248, "eval/last/train/acc": 1.0, "eval/last/train/acc_std": 0.0, "eval/last/train/f1": 1.0, "eval/last/train/f1_std": 0.0, "eval/last/validation/loss": 0.05252598226070404, "eval/last/validation/acc": 0.9878472222222222, "eval/last/validation/acc_std": 0.0016466534949246495, "eval/last/validation/f1": 0.9855294027870821, "eval/last/validation/f1_std": 0.002239532859910713, "eval/last/test/loss": 0.07176008820533752, "eval/last/test/acc": 0.9861111111111112, "eval/last/test/acc_std": 0.0015754859836431907, "eval/last/test/f1": 0.9832299793100557, "eval/last/test/f1_std": 0.0020954409398735822}

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__attn/eval_table.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,hcpya_task21,best,17,0.006599999999999999,0.05,43,"[22, 1.0]",train,0.0001459822669858113,1.0,0.0,1.0,0.0
+flat_mae,patch,attn,hcpya_task21,best,17,0.006599999999999999,0.05,43,"[22, 1.0]",validation,0.0525173656642437,0.9880952380952381,0.0016299318625606525,0.9857562615366776,0.0022326693782531346
+flat_mae,patch,attn,hcpya_task21,best,17,0.006599999999999999,0.05,43,"[22, 1.0]",test,0.07207445800304413,0.9861111111111112,0.0015754859836431907,0.983203683100717,0.002095390680912686

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__attn/eval_table_best.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,hcpya_task21,best,17,0.006599999999999999,0.05,43,"[22, 1.0]",train,0.0001459822669858113,1.0,0.0,1.0,0.0
+flat_mae,patch,attn,hcpya_task21,best,17,0.006599999999999999,0.05,43,"[22, 1.0]",validation,0.0525173656642437,0.9880952380952381,0.0016299318625606525,0.9857562615366776,0.0022326693782531346
+flat_mae,patch,attn,hcpya_task21,best,17,0.006599999999999999,0.05,43,"[22, 1.0]",test,0.07207445800304413,0.9861111111111112,0.0015754859836431907,0.983203683100717,0.002095390680912686

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__attn/eval_table_last.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,attn,hcpya_task21,last,19,0.006599999999999999,0.05,43,"[22, 1.0]",train,0.00014662370085716248,1.0,0.0,1.0,0.0
+flat_mae,patch,attn,hcpya_task21,last,19,0.006599999999999999,0.05,43,"[22, 1.0]",validation,0.05252598226070404,0.9878472222222222,0.0016466534949246495,0.9855294027870821,0.002239532859910713
+flat_mae,patch,attn,hcpya_task21,last,19,0.006599999999999999,0.05,43,"[22, 1.0]",test,0.07176008820533752,0.9861111111111112,0.0015754859836431907,0.9832299793100557,0.0020954409398735822

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/config.yaml

@@ -0,0 +1,96 @@
+output_root: experiments/decoders/output
+name_prefix: eval_probe
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (hcpya_task21 patch linear)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: false
+  norm: false
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear
+model: flat_mae
+representation: patch
+classifier: linear
+dataset: hcpya_task21
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear
+remote_dir: null

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/eval_log.json

@@ -0,0 +1 @@
+{"eval/epoch": 18, "eval/id_best": 48, "eval/lr_best": 0.015, "eval/wd_best": 0.05, "eval/train/loss": 0.30549120903015137, "eval/train/acc": 0.9457866203484394, "eval/train/acc_std": 0.0015117116549057134, "eval/train/f1": 0.9424528099876407, "eval/train/f1_std": 0.0018296477406085077, "eval/validation/loss": 0.339051216840744, "eval/validation/acc": 0.9320436507936508, "eval/validation/acc_std": 0.0038815927504984933, "eval/validation/f1": 0.9251783585297395, "eval/validation/f1_std": 0.0047207324832681035, "eval/test/loss": 0.34095045924186707, "eval/test/acc": 0.9327380952380953, "eval/test/acc_std": 0.003278294931146475, "eval/test/f1": 0.9250129678240288, "eval/test/f1_std": 0.004080640658782675}

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/eval_log_best.json

@@ -0,0 +1 @@
+{"eval/best/epoch": 18, "eval/best/id_best": 48, "eval/best/lr_best": 0.015, "eval/best/wd_best": 0.05, "eval/best/train/loss": 0.30549120903015137, "eval/best/train/acc": 0.9457866203484394, "eval/best/train/acc_std": 0.0015117116549057134, "eval/best/train/f1": 0.9424528099876407, "eval/best/train/f1_std": 0.0018296477406085077, "eval/best/validation/loss": 0.339051216840744, "eval/best/validation/acc": 0.9320436507936508, "eval/best/validation/acc_std": 0.0038815927504984933, "eval/best/validation/f1": 0.9251783585297395, "eval/best/validation/f1_std": 0.0047207324832681035, "eval/best/test/loss": 0.34095045924186707, "eval/best/test/acc": 0.9327380952380953, "eval/best/test/acc_std": 0.003278294931146475, "eval/best/test/f1": 0.9250129678240288, "eval/best/test/f1_std": 0.004080640658782675}

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/eval_log_last.json

@@ -0,0 +1 @@
+{"eval/last/epoch": 19, "eval/last/id_best": 48, "eval/last/lr_best": 0.015, "eval/last/wd_best": 0.05, "eval/last/train/loss": 0.30505988001823425, "eval/last/train/acc": 0.9450497394599716, "eval/last/train/acc_std": 0.0015442731153987098, "eval/last/train/f1": 0.94144574298719, "eval/last/train/f1_std": 0.001868007935164643, "eval/last/validation/loss": 0.3388846814632416, "eval/last/validation/acc": 0.9317956349206349, "eval/last/validation/acc_std": 0.0038951688379079744, "eval/last/validation/f1": 0.9241029616907658, "eval/last/validation/f1_std": 0.0047618741708175955, "eval/last/test/loss": 0.3405951261520386, "eval/last/test/acc": 0.9331349206349207, "eval/last/test/acc_std": 0.003325990211103103, "eval/last/test/f1": 0.9250078020719377, "eval/last/test/f1_std": 0.0041453204889727235}

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/eval_table.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,linear,hcpya_task21,best,18,0.015,0.05,48,"[50, 1.0]",train,0.30549120903015137,0.9457866203484394,0.0015117116549057134,0.9424528099876407,0.0018296477406085077
+flat_mae,patch,linear,hcpya_task21,best,18,0.015,0.05,48,"[50, 1.0]",validation,0.339051216840744,0.9320436507936508,0.0038815927504984933,0.9251783585297395,0.0047207324832681035
+flat_mae,patch,linear,hcpya_task21,best,18,0.015,0.05,48,"[50, 1.0]",test,0.34095045924186707,0.9327380952380953,0.003278294931146475,0.9250129678240288,0.004080640658782675

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/eval_table_best.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,linear,hcpya_task21,best,18,0.015,0.05,48,"[50, 1.0]",train,0.30549120903015137,0.9457866203484394,0.0015117116549057134,0.9424528099876407,0.0018296477406085077
+flat_mae,patch,linear,hcpya_task21,best,18,0.015,0.05,48,"[50, 1.0]",validation,0.339051216840744,0.9320436507936508,0.0038815927504984933,0.9251783585297395,0.0047207324832681035
+flat_mae,patch,linear,hcpya_task21,best,18,0.015,0.05,48,"[50, 1.0]",test,0.34095045924186707,0.9327380952380953,0.003278294931146475,0.9250129678240288,0.004080640658782675

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/eval_table_last.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,patch,linear,hcpya_task21,last,19,0.015,0.05,48,"[50, 1.0]",train,0.30505988001823425,0.9450497394599716,0.0015442731153987098,0.94144574298719,0.001868007935164643
+flat_mae,patch,linear,hcpya_task21,last,19,0.015,0.05,48,"[50, 1.0]",validation,0.3388846814632416,0.9317956349206349,0.0038951688379079744,0.9241029616907658,0.0047618741708175955
+flat_mae,patch,linear,hcpya_task21,last,19,0.015,0.05,48,"[50, 1.0]",test,0.3405951261520386,0.9331349206349207,0.003325990211103103,0.9250078020719377,0.0041453204889727235

decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/log.txt

@@ -0,0 +1,892 @@
+fMRI foundation model probe eval
+version: 0.1.dev66+g7ddd3aa04
+sha: aef99c83a386cf95c3d8ca503ecc968d8d5694af, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-03-07 23:04:47
+config:
+output_root: experiments/decoders/output
+name_prefix: eval_probe
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (hcpya_task21 patch linear)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: false
+  norm: false
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear
+model: flat_mae
+representation: patch
+classifier: linear
+dataset: hcpya_task21
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=False, reg_tokens=4, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: hcpya_task21 (flat)
+train (n=18999):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 18999
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[ 832 1248 3201 1660  832  832  832  832  832 1248 1247 1243  832  416
+  416  416  416  416  416  416  416]
+)
+
+validation (n=4032):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 4032
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[176 264 688 352 176 176 176 176 176 264 264 264 176  88  88  88  88  88
+  88  88  88]
+)
+
+test (n=5040):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 5040
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[220 330 860 440 220 220 220 220 220 330 330 330 220 110 110 110 110 110
+ 110 110 110]
+)
+
+running backbone on example batch to get embedding dim
+embedding feature dim (patch): 768
+initializing sweep of classifier heads
+classifiers:
+ModuleList(
+  (0-48): 49 x LinearClassifier(
+    (linear): Linear(in_features=768, out_features=21, bias=True)
+  )
+)
+classifier params (train): 0.8M (0.8M)
+setting up optimizer
+total batch size: 128 = 64 bs per gpu x 2 accum
+lr: 3.00e-04
+full schedule: epochs = 20 (steps = 4000) (decay = True)
+warmup: epochs = 5 (steps = 1000)
+start training for 20 epochs
+train: [0]  [  0/400]  eta: 0:22:27  lr: nan  time: 3.3682  data: 2.9601  max mem: 3914
+train: [0]  [ 20/400]  eta: 0:03:06  lr: 0.000003  loss: 3.0382 (3.0385)  grad: 0.1210 (0.1172)  time: 0.3477  data: 0.0039  max mem: 3956
+train: [0]  [ 40/400]  eta: 0:02:24  lr: 0.000006  loss: 3.0382 (3.0376)  grad: 0.1146 (0.1120)  time: 0.3072  data: 0.0029  max mem: 3956
+train: [0]  [ 60/400]  eta: 0:02:06  lr: 0.000009  loss: 3.0277 (3.0325)  grad: 0.1057 (0.1130)  time: 0.3108  data: 0.0031  max mem: 3956
+train: [0]  [ 80/400]  eta: 0:01:54  lr: 0.000012  loss: 3.0120 (3.0273)  grad: 0.1127 (0.1136)  time: 0.3147  data: 0.0033  max mem: 3956
+train: [0]  [100/400]  eta: 0:01:45  lr: 0.000015  loss: 3.0005 (3.0205)  grad: 0.1145 (0.1152)  time: 0.3284  data: 0.0033  max mem: 3956
+train: [0]  [120/400]  eta: 0:01:37  lr: 0.000018  loss: 2.9888 (3.0151)  grad: 0.1130 (0.1140)  time: 0.3351  data: 0.0033  max mem: 3956
+train: [0]  [140/400]  eta: 0:01:31  lr: 0.000021  loss: 2.9867 (3.0109)  grad: 0.1098 (0.1140)  time: 0.3632  data: 0.0033  max mem: 3956
+train: [0]  [160/400]  eta: 0:01:24  lr: 0.000024  loss: 2.9737 (3.0065)  grad: 0.1036 (0.1125)  time: 0.3520  data: 0.0033  max mem: 3956
+train: [0]  [180/400]  eta: 0:01:17  lr: 0.000027  loss: 2.9626 (2.9992)  grad: 0.0998 (0.1116)  time: 0.3405  data: 0.0032  max mem: 3956
+train: [0]  [200/400]  eta: 0:01:09  lr: 0.000030  loss: 2.9351 (2.9934)  grad: 0.0984 (0.1105)  time: 0.3477  data: 0.0032  max mem: 3956
+train: [0]  [220/400]  eta: 0:01:03  lr: 0.000033  loss: 2.9427 (2.9891)  grad: 0.0978 (0.1096)  time: 0.3579  data: 0.0033  max mem: 3956
+train: [0]  [240/400]  eta: 0:00:56  lr: 0.000036  loss: 2.9285 (2.9828)  grad: 0.1017 (0.1092)  time: 0.3547  data: 0.0032  max mem: 3956
+train: [0]  [260/400]  eta: 0:00:49  lr: 0.000039  loss: 2.8943 (2.9757)  grad: 0.1074 (0.1090)  time: 0.3451  data: 0.0032  max mem: 3956
+train: [0]  [280/400]  eta: 0:00:42  lr: 0.000042  loss: 2.8881 (2.9712)  grad: 0.0976 (0.1080)  time: 0.3691  data: 0.0032  max mem: 3956
+train: [0]  [300/400]  eta: 0:00:36  lr: 0.000045  loss: 2.8880 (2.9660)  grad: 0.0909 (0.1073)  time: 0.5296  data: 0.1932  max mem: 3956
+train: [0]  [320/400]  eta: 0:00:29  lr: 0.000048  loss: 2.8672 (2.9606)  grad: 0.0961 (0.1069)  time: 0.3469  data: 0.0037  max mem: 3956
+train: [0]  [340/400]  eta: 0:00:21  lr: 0.000051  loss: 2.8573 (2.9548)  grad: 0.1065 (0.1067)  time: 0.3597  data: 0.0026  max mem: 3956
+train: [0]  [360/400]  eta: 0:00:14  lr: 0.000054  loss: 2.8433 (2.9493)  grad: 0.0923 (0.1058)  time: 0.3593  data: 0.0034  max mem: 3956
+train: [0]  [380/400]  eta: 0:00:07  lr: 0.000057  loss: 2.8414 (2.9428)  grad: 0.0894 (0.1052)  time: 0.4274  data: 0.0038  max mem: 3956
+train: [0]  [399/400]  eta: 0:00:00  lr: 0.000060  loss: 2.8296 (2.9365)  grad: 0.0926 (0.1049)  time: 0.3535  data: 0.0034  max mem: 3956
+train: [0] Total time: 0:02:26 (0.3652 s / it)
+train: [0] Summary: lr: 0.000060  loss: 2.8296 (2.9365)  grad: 0.0926 (0.1049)
+eval (validation): [0]  [ 0/63]  eta: 0:03:38    time: 3.4642  data: 3.2269  max mem: 3956
+eval (validation): [0]  [20/63]  eta: 0:00:23    time: 0.3987  data: 0.0363  max mem: 3956
+eval (validation): [0]  [40/63]  eta: 0:00:10    time: 0.3591  data: 0.0026  max mem: 3956
+eval (validation): [0]  [60/63]  eta: 0:00:01    time: 0.3728  data: 0.0033  max mem: 3956
+eval (validation): [0]  [62/63]  eta: 0:00:00    time: 0.3727  data: 0.0032  max mem: 3956
+eval (validation): [0] Total time: 0:00:27 (0.4301 s / it)
+cv: [0]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 2.103  acc: 0.407  f1: 0.202
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [1]  [  0/400]  eta: 0:25:30  lr: nan  time: 3.8252  data: 3.5139  max mem: 3956
+train: [1]  [ 20/400]  eta: 0:03:22  lr: 0.000063  loss: 2.7878 (2.8015)  grad: 0.0811 (0.0858)  time: 0.3670  data: 0.0107  max mem: 3956
+train: [1]  [ 40/400]  eta: 0:02:40  lr: 0.000066  loss: 2.8068 (2.8159)  grad: 0.0824 (0.0865)  time: 0.3579  data: 0.0030  max mem: 3956
+train: [1]  [ 60/400]  eta: 0:02:23  lr: 0.000069  loss: 2.8002 (2.8045)  grad: 0.0891 (0.0893)  time: 0.3673  data: 0.0030  max mem: 3956
+train: [1]  [ 80/400]  eta: 0:02:13  lr: 0.000072  loss: 2.7775 (2.8027)  grad: 0.0902 (0.0881)  time: 0.4020  data: 0.0036  max mem: 3956
+train: [1]  [100/400]  eta: 0:02:01  lr: 0.000075  loss: 2.7775 (2.7961)  grad: 0.0902 (0.0892)  time: 0.3610  data: 0.0032  max mem: 3956
+train: [1]  [120/400]  eta: 0:01:52  lr: 0.000078  loss: 2.7386 (2.7856)  grad: 0.0897 (0.0896)  time: 0.3913  data: 0.0035  max mem: 3956
+train: [1]  [140/400]  eta: 0:01:43  lr: 0.000081  loss: 2.7294 (2.7810)  grad: 0.0877 (0.0894)  time: 0.3599  data: 0.0035  max mem: 3956
+train: [1]  [160/400]  eta: 0:01:35  lr: 0.000084  loss: 2.7393 (2.7744)  grad: 0.0871 (0.0894)  time: 0.4032  data: 0.0034  max mem: 3956
+train: [1]  [180/400]  eta: 0:01:27  lr: 0.000087  loss: 2.7183 (2.7665)  grad: 0.0912 (0.0895)  time: 0.3860  data: 0.0033  max mem: 3956
+train: [1]  [200/400]  eta: 0:01:19  lr: 0.000090  loss: 2.6855 (2.7566)  grad: 0.0917 (0.0899)  time: 0.3870  data: 0.0034  max mem: 3956
+train: [1]  [220/400]  eta: 0:01:11  lr: 0.000093  loss: 2.6808 (2.7491)  grad: 0.0917 (0.0902)  time: 0.3992  data: 0.0035  max mem: 3956
+train: [1]  [240/400]  eta: 0:01:03  lr: 0.000096  loss: 2.6750 (2.7424)  grad: 0.0885 (0.0899)  time: 0.3723  data: 0.0032  max mem: 3956
+train: [1]  [260/400]  eta: 0:00:54  lr: 0.000099  loss: 2.6692 (2.7380)  grad: 0.0833 (0.0894)  time: 0.3627  data: 0.0033  max mem: 3956
+train: [1]  [280/400]  eta: 0:00:46  lr: 0.000102  loss: 2.6618 (2.7317)  grad: 0.0846 (0.0895)  time: 0.3907  data: 0.0035  max mem: 3956
+train: [1]  [300/400]  eta: 0:00:40  lr: 0.000105  loss: 2.6550 (2.7250)  grad: 0.0838 (0.0892)  time: 0.5890  data: 0.2012  max mem: 3956
+train: [1]  [320/400]  eta: 0:00:32  lr: 0.000108  loss: 2.6159 (2.7168)  grad: 0.0836 (0.0889)  time: 0.3803  data: 0.0037  max mem: 3956
+train: [1]  [340/400]  eta: 0:00:24  lr: 0.000111  loss: 2.5978 (2.7115)  grad: 0.0793 (0.0882)  time: 0.3767  data: 0.0030  max mem: 3956
+train: [1]  [360/400]  eta: 0:00:15  lr: 0.000114  loss: 2.6142 (2.7054)  grad: 0.0782 (0.0879)  time: 0.3673  data: 0.0034  max mem: 3956
+train: [1]  [380/400]  eta: 0:00:07  lr: 0.000117  loss: 2.5867 (2.6997)  grad: 0.0811 (0.0875)  time: 0.3737  data: 0.0033  max mem: 3956
+train: [1]  [399/400]  eta: 0:00:00  lr: 0.000120  loss: 2.5848 (2.6938)  grad: 0.0837 (0.0874)  time: 0.3651  data: 0.0033  max mem: 3956
+train: [1] Total time: 0:02:38 (0.3969 s / it)
+train: [1] Summary: lr: 0.000120  loss: 2.5848 (2.6938)  grad: 0.0837 (0.0874)
+eval (validation): [1]  [ 0/63]  eta: 0:03:48    time: 3.6260  data: 3.3944  max mem: 3956
+eval (validation): [1]  [20/63]  eta: 0:00:23    time: 0.3875  data: 0.0035  max mem: 3956
+eval (validation): [1]  [40/63]  eta: 0:00:10    time: 0.3481  data: 0.0037  max mem: 3956
+eval (validation): [1]  [60/63]  eta: 0:00:01    time: 0.3285  data: 0.0032  max mem: 3956
+eval (validation): [1]  [62/63]  eta: 0:00:00    time: 0.3265  data: 0.0031  max mem: 3956
+eval (validation): [1] Total time: 0:00:25 (0.4102 s / it)
+cv: [1]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 1.237  acc: 0.736  f1: 0.661
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [2]  [  0/400]  eta: 0:23:20  lr: nan  time: 3.5017  data: 3.2263  max mem: 3956
+train: [2]  [ 20/400]  eta: 0:03:14  lr: 0.000123  loss: 2.5009 (2.5217)  grad: 0.0809 (0.0843)  time: 0.3631  data: 0.0137  max mem: 3956
+train: [2]  [ 40/400]  eta: 0:02:38  lr: 0.000126  loss: 2.5347 (2.5388)  grad: 0.0840 (0.0853)  time: 0.3639  data: 0.0034  max mem: 3956
+train: [2]  [ 60/400]  eta: 0:02:19  lr: 0.000129  loss: 2.5432 (2.5391)  grad: 0.0801 (0.0824)  time: 0.3498  data: 0.0025  max mem: 3956
+train: [2]  [ 80/400]  eta: 0:02:08  lr: 0.000132  loss: 2.5429 (2.5420)  grad: 0.0732 (0.0807)  time: 0.3718  data: 0.0035  max mem: 3956
+train: [2]  [100/400]  eta: 0:01:57  lr: 0.000135  loss: 2.5269 (2.5358)  grad: 0.0800 (0.0815)  time: 0.3554  data: 0.0034  max mem: 3956
+train: [2]  [120/400]  eta: 0:01:47  lr: 0.000138  loss: 2.5052 (2.5308)  grad: 0.0819 (0.0812)  time: 0.3430  data: 0.0033  max mem: 3956
+train: [2]  [140/400]  eta: 0:01:38  lr: 0.000141  loss: 2.5027 (2.5271)  grad: 0.0780 (0.0808)  time: 0.3581  data: 0.0036  max mem: 3956
+train: [2]  [160/400]  eta: 0:01:30  lr: 0.000144  loss: 2.4808 (2.5219)  grad: 0.0780 (0.0810)  time: 0.3540  data: 0.0034  max mem: 3956
+train: [2]  [180/400]  eta: 0:01:22  lr: 0.000147  loss: 2.4620 (2.5141)  grad: 0.0842 (0.0817)  time: 0.3564  data: 0.0033  max mem: 3956
+train: [2]  [200/400]  eta: 0:01:14  lr: 0.000150  loss: 2.4639 (2.5108)  grad: 0.0821 (0.0813)  time: 0.3642  data: 0.0035  max mem: 3956
+train: [2]  [220/400]  eta: 0:01:07  lr: 0.000153  loss: 2.4606 (2.5035)  grad: 0.0755 (0.0814)  time: 0.3649  data: 0.0035  max mem: 3956
+train: [2]  [240/400]  eta: 0:00:59  lr: 0.000156  loss: 2.4371 (2.4980)  grad: 0.0770 (0.0808)  time: 0.3570  data: 0.0031  max mem: 3956
+train: [2]  [260/400]  eta: 0:00:51  lr: 0.000159  loss: 2.4487 (2.4941)  grad: 0.0760 (0.0804)  time: 0.3589  data: 0.0033  max mem: 3956
+train: [2]  [280/400]  eta: 0:00:44  lr: 0.000162  loss: 2.4487 (2.4899)  grad: 0.0760 (0.0801)  time: 0.3506  data: 0.0036  max mem: 3956
+train: [2]  [300/400]  eta: 0:00:37  lr: 0.000165  loss: 2.4291 (2.4853)  grad: 0.0750 (0.0800)  time: 0.5270  data: 0.1889  max mem: 3956
+train: [2]  [320/400]  eta: 0:00:30  lr: 0.000168  loss: 2.4186 (2.4806)  grad: 0.0785 (0.0801)  time: 0.3652  data: 0.0035  max mem: 3956
+train: [2]  [340/400]  eta: 0:00:22  lr: 0.000171  loss: 2.4186 (2.4771)  grad: 0.0738 (0.0796)  time: 0.3510  data: 0.0030  max mem: 3956
+train: [2]  [360/400]  eta: 0:00:15  lr: 0.000174  loss: 2.3784 (2.4702)  grad: 0.0695 (0.0794)  time: 0.3616  data: 0.0034  max mem: 3956
+train: [2]  [380/400]  eta: 0:00:07  lr: 0.000177  loss: 2.3712 (2.4656)  grad: 0.0697 (0.0790)  time: 0.3434  data: 0.0035  max mem: 3956
+train: [2]  [399/400]  eta: 0:00:00  lr: 0.000180  loss: 2.3904 (2.4613)  grad: 0.0749 (0.0788)  time: 0.3404  data: 0.0033  max mem: 3956
+train: [2] Total time: 0:02:29 (0.3732 s / it)
+train: [2] Summary: lr: 0.000180  loss: 2.3904 (2.4613)  grad: 0.0749 (0.0788)
+eval (validation): [2]  [ 0/63]  eta: 0:04:27    time: 4.2458  data: 4.0183  max mem: 3956
+eval (validation): [2]  [20/63]  eta: 0:00:22    time: 0.3456  data: 0.0033  max mem: 3956
+eval (validation): [2]  [40/63]  eta: 0:00:10    time: 0.3342  data: 0.0034  max mem: 3956
+eval (validation): [2]  [60/63]  eta: 0:00:01    time: 0.3273  data: 0.0033  max mem: 3956
+eval (validation): [2]  [62/63]  eta: 0:00:00    time: 0.3270  data: 0.0033  max mem: 3956
+eval (validation): [2] Total time: 0:00:25 (0.4026 s / it)
+cv: [2]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.835  acc: 0.832  f1: 0.785
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [3]  [  0/400]  eta: 0:23:05  lr: nan  time: 3.4640  data: 3.1862  max mem: 3956
+train: [3]  [ 20/400]  eta: 0:03:07  lr: 0.000183  loss: 2.3519 (2.3505)  grad: 0.0732 (0.0740)  time: 0.3452  data: 0.0041  max mem: 3956
+train: [3]  [ 40/400]  eta: 0:02:37  lr: 0.000186  loss: 2.3421 (2.3411)  grad: 0.0744 (0.0749)  time: 0.3763  data: 0.0033  max mem: 3956
+train: [3]  [ 60/400]  eta: 0:02:19  lr: 0.000189  loss: 2.3062 (2.3301)  grad: 0.0730 (0.0740)  time: 0.3587  data: 0.0031  max mem: 3956
+train: [3]  [ 80/400]  eta: 0:02:08  lr: 0.000192  loss: 2.3246 (2.3326)  grad: 0.0717 (0.0748)  time: 0.3717  data: 0.0034  max mem: 3956
+train: [3]  [100/400]  eta: 0:01:57  lr: 0.000195  loss: 2.3183 (2.3243)  grad: 0.0722 (0.0738)  time: 0.3479  data: 0.0032  max mem: 3956
+train: [3]  [120/400]  eta: 0:01:47  lr: 0.000198  loss: 2.2664 (2.3162)  grad: 0.0719 (0.0736)  time: 0.3585  data: 0.0033  max mem: 3956
+train: [3]  [140/400]  eta: 0:01:39  lr: 0.000201  loss: 2.2821 (2.3162)  grad: 0.0726 (0.0736)  time: 0.3761  data: 0.0035  max mem: 3956
+train: [3]  [160/400]  eta: 0:01:31  lr: 0.000204  loss: 2.2821 (2.3120)  grad: 0.0704 (0.0733)  time: 0.3535  data: 0.0033  max mem: 3956
+train: [3]  [180/400]  eta: 0:01:23  lr: 0.000207  loss: 2.2835 (2.3095)  grad: 0.0717 (0.0735)  time: 0.3699  data: 0.0033  max mem: 3956
+train: [3]  [200/400]  eta: 0:01:15  lr: 0.000210  loss: 2.2832 (2.3059)  grad: 0.0725 (0.0734)  time: 0.3579  data: 0.0033  max mem: 3956
+train: [3]  [220/400]  eta: 0:01:07  lr: 0.000213  loss: 2.2792 (2.3052)  grad: 0.0697 (0.0732)  time: 0.3586  data: 0.0034  max mem: 3956
+train: [3]  [240/400]  eta: 0:00:59  lr: 0.000216  loss: 2.2792 (2.3027)  grad: 0.0697 (0.0732)  time: 0.3567  data: 0.0033  max mem: 3956
+train: [3]  [260/400]  eta: 0:00:52  lr: 0.000219  loss: 2.2596 (2.3003)  grad: 0.0759 (0.0736)  time: 0.3631  data: 0.0033  max mem: 3956
+train: [3]  [280/400]  eta: 0:00:44  lr: 0.000222  loss: 2.2382 (2.2959)  grad: 0.0727 (0.0731)  time: 0.3644  data: 0.0035  max mem: 3956
+train: [3]  [300/400]  eta: 0:00:38  lr: 0.000225  loss: 2.2275 (2.2907)  grad: 0.0648 (0.0730)  time: 0.5245  data: 0.1859  max mem: 3956
+train: [3]  [320/400]  eta: 0:00:30  lr: 0.000228  loss: 2.2097 (2.2874)  grad: 0.0648 (0.0726)  time: 0.3995  data: 0.0043  max mem: 3956
+train: [3]  [340/400]  eta: 0:00:22  lr: 0.000231  loss: 2.2049 (2.2819)  grad: 0.0663 (0.0723)  time: 0.3439  data: 0.0027  max mem: 3956
+train: [3]  [360/400]  eta: 0:00:15  lr: 0.000234  loss: 2.2022 (2.2783)  grad: 0.0646 (0.0718)  time: 0.3949  data: 0.0035  max mem: 3956
+train: [3]  [380/400]  eta: 0:00:07  lr: 0.000237  loss: 2.2240 (2.2765)  grad: 0.0645 (0.0715)  time: 0.3790  data: 0.0035  max mem: 3956
+train: [3]  [399/400]  eta: 0:00:00  lr: 0.000240  loss: 2.2224 (2.2716)  grad: 0.0678 (0.0714)  time: 0.3647  data: 0.0035  max mem: 3956
+train: [3] Total time: 0:02:32 (0.3813 s / it)
+train: [3] Summary: lr: 0.000240  loss: 2.2224 (2.2716)  grad: 0.0678 (0.0714)
+eval (validation): [3]  [ 0/63]  eta: 0:03:52    time: 3.6954  data: 3.4600  max mem: 3956
+eval (validation): [3]  [20/63]  eta: 0:00:23    time: 0.3887  data: 0.0030  max mem: 3956
+eval (validation): [3]  [40/63]  eta: 0:00:10    time: 0.3527  data: 0.0034  max mem: 3956
+eval (validation): [3]  [60/63]  eta: 0:00:01    time: 0.3287  data: 0.0033  max mem: 3956
+eval (validation): [3]  [62/63]  eta: 0:00:00    time: 0.3289  data: 0.0033  max mem: 3956
+eval (validation): [3] Total time: 0:00:26 (0.4144 s / it)
+cv: [3]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.638  acc: 0.872  f1: 0.852
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [4]  [  0/400]  eta: 0:22:52  lr: nan  time: 3.4305  data: 3.1927  max mem: 3956
+train: [4]  [ 20/400]  eta: 0:03:12  lr: 0.000243  loss: 2.1948 (2.2009)  grad: 0.0624 (0.0686)  time: 0.3595  data: 0.0060  max mem: 3956
+train: [4]  [ 40/400]  eta: 0:02:37  lr: 0.000246  loss: 2.1636 (2.1912)  grad: 0.0641 (0.0676)  time: 0.3648  data: 0.0032  max mem: 3956
+train: [4]  [ 60/400]  eta: 0:02:20  lr: 0.000249  loss: 2.1632 (2.1759)  grad: 0.0652 (0.0687)  time: 0.3615  data: 0.0029  max mem: 3956
+train: [4]  [ 80/400]  eta: 0:02:07  lr: 0.000252  loss: 2.1584 (2.1672)  grad: 0.0675 (0.0682)  time: 0.3533  data: 0.0035  max mem: 3956
+train: [4]  [100/400]  eta: 0:01:56  lr: 0.000255  loss: 2.1609 (2.1658)  grad: 0.0649 (0.0675)  time: 0.3477  data: 0.0031  max mem: 3956
+train: [4]  [120/400]  eta: 0:01:47  lr: 0.000258  loss: 2.1711 (2.1662)  grad: 0.0649 (0.0671)  time: 0.3651  data: 0.0033  max mem: 3956
+train: [4]  [140/400]  eta: 0:01:39  lr: 0.000261  loss: 2.1459 (2.1594)  grad: 0.0678 (0.0673)  time: 0.3662  data: 0.0034  max mem: 3956
+train: [4]  [160/400]  eta: 0:01:30  lr: 0.000264  loss: 2.1264 (2.1559)  grad: 0.0668 (0.0669)  time: 0.3579  data: 0.0032  max mem: 3956
+train: [4]  [180/400]  eta: 0:01:22  lr: 0.000267  loss: 2.1317 (2.1523)  grad: 0.0663 (0.0668)  time: 0.3576  data: 0.0034  max mem: 3956
+train: [4]  [200/400]  eta: 0:01:14  lr: 0.000270  loss: 2.1037 (2.1457)  grad: 0.0644 (0.0668)  time: 0.3508  data: 0.0035  max mem: 3956
+train: [4]  [220/400]  eta: 0:01:06  lr: 0.000273  loss: 2.1036 (2.1444)  grad: 0.0622 (0.0661)  time: 0.3510  data: 0.0033  max mem: 3956
+train: [4]  [240/400]  eta: 0:00:59  lr: 0.000276  loss: 2.1010 (2.1392)  grad: 0.0611 (0.0662)  time: 0.3734  data: 0.0035  max mem: 3956
+train: [4]  [260/400]  eta: 0:00:52  lr: 0.000279  loss: 2.0573 (2.1336)  grad: 0.0643 (0.0659)  time: 0.3687  data: 0.0035  max mem: 3956
+train: [4]  [280/400]  eta: 0:00:44  lr: 0.000282  loss: 2.0792 (2.1316)  grad: 0.0609 (0.0656)  time: 0.3554  data: 0.0035  max mem: 3956
+train: [4]  [300/400]  eta: 0:00:38  lr: 0.000285  loss: 2.1022 (2.1283)  grad: 0.0606 (0.0654)  time: 0.5426  data: 0.2144  max mem: 3956
+train: [4]  [320/400]  eta: 0:00:30  lr: 0.000288  loss: 2.0829 (2.1243)  grad: 0.0655 (0.0657)  time: 0.3732  data: 0.0123  max mem: 3956
+train: [4]  [340/400]  eta: 0:00:22  lr: 0.000291  loss: 2.0829 (2.1223)  grad: 0.0631 (0.0654)  time: 0.3461  data: 0.0043  max mem: 3956
+train: [4]  [360/400]  eta: 0:00:15  lr: 0.000294  loss: 2.0690 (2.1195)  grad: 0.0616 (0.0653)  time: 0.3493  data: 0.0034  max mem: 3956
+train: [4]  [380/400]  eta: 0:00:07  lr: 0.000297  loss: 2.0690 (2.1168)  grad: 0.0616 (0.0651)  time: 0.3754  data: 0.0034  max mem: 3956
+train: [4]  [399/400]  eta: 0:00:00  lr: 0.000300  loss: 2.0624 (2.1131)  grad: 0.0589 (0.0649)  time: 0.3757  data: 0.0033  max mem: 3956
+train: [4] Total time: 0:02:31 (0.3776 s / it)
+train: [4] Summary: lr: 0.000300  loss: 2.0624 (2.1131)  grad: 0.0589 (0.0649)
+eval (validation): [4]  [ 0/63]  eta: 0:03:44    time: 3.5711  data: 3.3217  max mem: 3956
+eval (validation): [4]  [20/63]  eta: 0:00:23    time: 0.3861  data: 0.0039  max mem: 3956
+eval (validation): [4]  [40/63]  eta: 0:00:10    time: 0.3536  data: 0.0032  max mem: 3956
+eval (validation): [4]  [60/63]  eta: 0:00:01    time: 0.3682  data: 0.0033  max mem: 3956
+eval (validation): [4]  [62/63]  eta: 0:00:00    time: 0.3739  data: 0.0033  max mem: 3956
+eval (validation): [4] Total time: 0:00:26 (0.4256 s / it)
+cv: [4]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.533  acc: 0.890  f1: 0.872
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [5]  [  0/400]  eta: 0:24:13  lr: nan  time: 3.6329  data: 3.3237  max mem: 3956
+train: [5]  [ 20/400]  eta: 0:03:21  lr: 0.000300  loss: 2.0223 (2.0412)  grad: 0.0590 (0.0602)  time: 0.3751  data: 0.0042  max mem: 3956
+train: [5]  [ 40/400]  eta: 0:02:42  lr: 0.000300  loss: 2.0567 (2.0572)  grad: 0.0577 (0.0597)  time: 0.3683  data: 0.0032  max mem: 3956
+train: [5]  [ 60/400]  eta: 0:02:23  lr: 0.000300  loss: 2.0382 (2.0416)  grad: 0.0645 (0.0621)  time: 0.3647  data: 0.0035  max mem: 3956
+train: [5]  [ 80/400]  eta: 0:02:09  lr: 0.000300  loss: 1.9948 (2.0353)  grad: 0.0645 (0.0617)  time: 0.3515  data: 0.0035  max mem: 3956
+train: [5]  [100/400]  eta: 0:01:58  lr: 0.000300  loss: 1.9896 (2.0257)  grad: 0.0646 (0.0629)  time: 0.3610  data: 0.0036  max mem: 3956
+train: [5]  [120/400]  eta: 0:01:49  lr: 0.000300  loss: 2.0153 (2.0236)  grad: 0.0665 (0.0635)  time: 0.3731  data: 0.0035  max mem: 3956
+train: [5]  [140/400]  eta: 0:01:40  lr: 0.000300  loss: 2.0186 (2.0206)  grad: 0.0637 (0.0633)  time: 0.3521  data: 0.0034  max mem: 3956
+train: [5]  [160/400]  eta: 0:01:32  lr: 0.000299  loss: 1.9782 (2.0163)  grad: 0.0609 (0.0632)  time: 0.3629  data: 0.0034  max mem: 3956
+train: [5]  [180/400]  eta: 0:01:23  lr: 0.000299  loss: 1.9904 (2.0164)  grad: 0.0603 (0.0630)  time: 0.3619  data: 0.0038  max mem: 3956
+train: [5]  [200/400]  eta: 0:01:15  lr: 0.000299  loss: 2.0021 (2.0171)  grad: 0.0616 (0.0630)  time: 0.3594  data: 0.0036  max mem: 3956
+train: [5]  [220/400]  eta: 0:01:07  lr: 0.000299  loss: 1.9935 (2.0134)  grad: 0.0597 (0.0628)  time: 0.3573  data: 0.0036  max mem: 3956
+train: [5]  [240/400]  eta: 0:01:00  lr: 0.000299  loss: 1.9692 (2.0098)  grad: 0.0590 (0.0627)  time: 0.3567  data: 0.0033  max mem: 3956
+train: [5]  [260/400]  eta: 0:00:52  lr: 0.000299  loss: 1.9796 (2.0085)  grad: 0.0563 (0.0624)  time: 0.3510  data: 0.0033  max mem: 3956
+train: [5]  [280/400]  eta: 0:00:44  lr: 0.000298  loss: 1.9796 (2.0062)  grad: 0.0559 (0.0621)  time: 0.3635  data: 0.0033  max mem: 3956
+train: [5]  [300/400]  eta: 0:00:38  lr: 0.000298  loss: 1.9642 (2.0042)  grad: 0.0560 (0.0616)  time: 0.5563  data: 0.1949  max mem: 3956
+train: [5]  [320/400]  eta: 0:00:30  lr: 0.000298  loss: 1.9673 (2.0031)  grad: 0.0577 (0.0612)  time: 0.3599  data: 0.0034  max mem: 3956
+train: [5]  [340/400]  eta: 0:00:22  lr: 0.000298  loss: 1.9637 (2.0012)  grad: 0.0591 (0.0612)  time: 0.3456  data: 0.0032  max mem: 3956
+train: [5]  [360/400]  eta: 0:00:15  lr: 0.000297  loss: 1.9784 (1.9999)  grad: 0.0591 (0.0611)  time: 0.3400  data: 0.0034  max mem: 3956
+train: [5]  [380/400]  eta: 0:00:07  lr: 0.000297  loss: 1.9973 (1.9994)  grad: 0.0614 (0.0612)  time: 0.3531  data: 0.0034  max mem: 3956
+train: [5]  [399/400]  eta: 0:00:00  lr: 0.000297  loss: 1.9507 (1.9964)  grad: 0.0617 (0.0611)  time: 0.3680  data: 0.0034  max mem: 3956
+train: [5] Total time: 0:02:31 (0.3776 s / it)
+train: [5] Summary: lr: 0.000297  loss: 1.9507 (1.9964)  grad: 0.0617 (0.0611)
+eval (validation): [5]  [ 0/63]  eta: 0:03:43    time: 3.5461  data: 3.2695  max mem: 3956
+eval (validation): [5]  [20/63]  eta: 0:00:21    time: 0.3439  data: 0.0127  max mem: 3956
+eval (validation): [5]  [40/63]  eta: 0:00:09    time: 0.3437  data: 0.0035  max mem: 3956
+eval (validation): [5]  [60/63]  eta: 0:00:01    time: 0.3300  data: 0.0030  max mem: 3956
+eval (validation): [5]  [62/63]  eta: 0:00:00    time: 0.3291  data: 0.0033  max mem: 3956
+eval (validation): [5] Total time: 0:00:24 (0.3947 s / it)
+cv: [5]  best hparam: (36, 1.0) (046) ('046_lr3.6e+01_wd1.0e+00')  loss: 0.524  acc: 0.895  f1: 0.879
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [6]  [  0/400]  eta: 0:23:25  lr: nan  time: 3.5137  data: 3.2326  max mem: 3956
+train: [6]  [ 20/400]  eta: 0:03:12  lr: 0.000296  loss: 1.9619 (1.9624)  grad: 0.0543 (0.0545)  time: 0.3557  data: 0.0035  max mem: 3956
+train: [6]  [ 40/400]  eta: 0:02:38  lr: 0.000296  loss: 1.9481 (1.9541)  grad: 0.0543 (0.0566)  time: 0.3728  data: 0.0032  max mem: 3956
+train: [6]  [ 60/400]  eta: 0:02:22  lr: 0.000296  loss: 1.9322 (1.9467)  grad: 0.0596 (0.0573)  time: 0.3741  data: 0.0033  max mem: 3956
+train: [6]  [ 80/400]  eta: 0:02:09  lr: 0.000295  loss: 1.9021 (1.9356)  grad: 0.0585 (0.0573)  time: 0.3599  data: 0.0034  max mem: 3956
+train: [6]  [100/400]  eta: 0:01:57  lr: 0.000295  loss: 1.9092 (1.9318)  grad: 0.0559 (0.0572)  time: 0.3440  data: 0.0036  max mem: 3956
+train: [6]  [120/400]  eta: 0:01:48  lr: 0.000295  loss: 1.9093 (1.9263)  grad: 0.0572 (0.0574)  time: 0.3686  data: 0.0035  max mem: 3956
+train: [6]  [140/400]  eta: 0:01:40  lr: 0.000294  loss: 1.8930 (1.9193)  grad: 0.0602 (0.0585)  time: 0.3657  data: 0.0035  max mem: 3956
+train: [6]  [160/400]  eta: 0:01:32  lr: 0.000294  loss: 1.8864 (1.9177)  grad: 0.0602 (0.0589)  time: 0.3752  data: 0.0035  max mem: 3956
+train: [6]  [180/400]  eta: 0:01:24  lr: 0.000293  loss: 1.9140 (1.9183)  grad: 0.0572 (0.0591)  time: 0.3815  data: 0.0034  max mem: 3956
+train: [6]  [200/400]  eta: 0:01:16  lr: 0.000293  loss: 1.9310 (1.9183)  grad: 0.0568 (0.0590)  time: 0.3819  data: 0.0036  max mem: 3956
+train: [6]  [220/400]  eta: 0:01:09  lr: 0.000292  loss: 1.8974 (1.9150)  grad: 0.0554 (0.0590)  time: 0.3876  data: 0.0032  max mem: 3956
+train: [6]  [240/400]  eta: 0:01:01  lr: 0.000292  loss: 1.8893 (1.9136)  grad: 0.0567 (0.0591)  time: 0.3686  data: 0.0034  max mem: 3956
+train: [6]  [260/400]  eta: 0:00:53  lr: 0.000291  loss: 1.8855 (1.9128)  grad: 0.0578 (0.0593)  time: 0.3810  data: 0.0036  max mem: 3956
+train: [6]  [280/400]  eta: 0:00:45  lr: 0.000291  loss: 1.8645 (1.9088)  grad: 0.0553 (0.0590)  time: 0.3846  data: 0.0032  max mem: 3956
+train: [6]  [300/400]  eta: 0:00:39  lr: 0.000290  loss: 1.8635 (1.9061)  grad: 0.0532 (0.0587)  time: 0.5316  data: 0.1928  max mem: 3956
+train: [6]  [320/400]  eta: 0:00:31  lr: 0.000290  loss: 1.8621 (1.9023)  grad: 0.0551 (0.0585)  time: 0.4282  data: 0.0034  max mem: 3956
+train: [6]  [340/400]  eta: 0:00:23  lr: 0.000289  loss: 1.8558 (1.9006)  grad: 0.0601 (0.0585)  time: 0.3646  data: 0.0029  max mem: 3956
+train: [6]  [360/400]  eta: 0:00:15  lr: 0.000288  loss: 1.8489 (1.8973)  grad: 0.0603 (0.0586)  time: 0.3973  data: 0.0031  max mem: 3956
+train: [6]  [380/400]  eta: 0:00:07  lr: 0.000288  loss: 1.8426 (1.8967)  grad: 0.0569 (0.0585)  time: 0.3686  data: 0.0033  max mem: 3956
+train: [6]  [399/400]  eta: 0:00:00  lr: 0.000287  loss: 1.8567 (1.8950)  grad: 0.0575 (0.0586)  time: 0.3775  data: 0.0031  max mem: 3956
+train: [6] Total time: 0:02:36 (0.3915 s / it)
+train: [6] Summary: lr: 0.000287  loss: 1.8567 (1.8950)  grad: 0.0575 (0.0586)
+eval (validation): [6]  [ 0/63]  eta: 0:03:53    time: 3.7134  data: 3.4028  max mem: 3956
+eval (validation): [6]  [20/63]  eta: 0:00:22    time: 0.3718  data: 0.0041  max mem: 3956
+eval (validation): [6]  [40/63]  eta: 0:00:10    time: 0.3732  data: 0.0028  max mem: 3956
+eval (validation): [6]  [60/63]  eta: 0:00:01    time: 0.3564  data: 0.0034  max mem: 3956
+eval (validation): [6]  [62/63]  eta: 0:00:00    time: 0.3569  data: 0.0034  max mem: 3956
+eval (validation): [6] Total time: 0:00:26 (0.4245 s / it)
+cv: [6]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.443  acc: 0.909  f1: 0.889
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [7]  [  0/400]  eta: 0:25:15  lr: nan  time: 3.7887  data: 3.4544  max mem: 3956
+train: [7]  [ 20/400]  eta: 0:03:49  lr: 0.000286  loss: 1.8169 (1.8531)  grad: 0.0581 (0.0566)  time: 0.4443  data: 0.0030  max mem: 3956
+train: [7]  [ 40/400]  eta: 0:02:57  lr: 0.000286  loss: 1.8248 (1.8516)  grad: 0.0555 (0.0569)  time: 0.3762  data: 0.0033  max mem: 3956
+train: [7]  [ 60/400]  eta: 0:02:35  lr: 0.000285  loss: 1.8248 (1.8404)  grad: 0.0600 (0.0591)  time: 0.3892  data: 0.0032  max mem: 3956
+train: [7]  [ 80/400]  eta: 0:02:22  lr: 0.000284  loss: 1.8725 (1.8525)  grad: 0.0557 (0.0577)  time: 0.4000  data: 0.0036  max mem: 3956
+train: [7]  [100/400]  eta: 0:02:08  lr: 0.000284  loss: 1.8551 (1.8452)  grad: 0.0508 (0.0565)  time: 0.3695  data: 0.0033  max mem: 3956
+train: [7]  [120/400]  eta: 0:01:58  lr: 0.000283  loss: 1.8241 (1.8420)  grad: 0.0546 (0.0564)  time: 0.3889  data: 0.0038  max mem: 3956
+train: [7]  [140/400]  eta: 0:01:48  lr: 0.000282  loss: 1.8455 (1.8436)  grad: 0.0538 (0.0558)  time: 0.3974  data: 0.0036  max mem: 3956
+train: [7]  [160/400]  eta: 0:01:39  lr: 0.000282  loss: 1.8286 (1.8417)  grad: 0.0526 (0.0555)  time: 0.3967  data: 0.0036  max mem: 3956
+train: [7]  [180/400]  eta: 0:01:31  lr: 0.000281  loss: 1.8136 (1.8390)  grad: 0.0526 (0.0554)  time: 0.4090  data: 0.0036  max mem: 3956
+train: [7]  [200/400]  eta: 0:01:22  lr: 0.000280  loss: 1.8132 (1.8390)  grad: 0.0525 (0.0550)  time: 0.3772  data: 0.0033  max mem: 3956
+train: [7]  [220/400]  eta: 0:01:13  lr: 0.000279  loss: 1.8443 (1.8397)  grad: 0.0532 (0.0552)  time: 0.3710  data: 0.0034  max mem: 3956
+train: [7]  [240/400]  eta: 0:01:04  lr: 0.000278  loss: 1.8495 (1.8395)  grad: 0.0554 (0.0552)  time: 0.3829  data: 0.0035  max mem: 3956
+train: [7]  [260/400]  eta: 0:00:56  lr: 0.000278  loss: 1.8149 (1.8382)  grad: 0.0543 (0.0550)  time: 0.3748  data: 0.0034  max mem: 3956
+train: [7]  [280/400]  eta: 0:00:48  lr: 0.000277  loss: 1.8303 (1.8380)  grad: 0.0513 (0.0548)  time: 0.3690  data: 0.0034  max mem: 3956
+train: [7]  [300/400]  eta: 0:00:41  lr: 0.000276  loss: 1.8279 (1.8375)  grad: 0.0507 (0.0545)  time: 0.5592  data: 0.2241  max mem: 3956
+train: [7]  [320/400]  eta: 0:00:32  lr: 0.000275  loss: 1.8255 (1.8342)  grad: 0.0515 (0.0546)  time: 0.3877  data: 0.0032  max mem: 3956
+train: [7]  [340/400]  eta: 0:00:24  lr: 0.000274  loss: 1.7815 (1.8305)  grad: 0.0562 (0.0548)  time: 0.3856  data: 0.0031  max mem: 3956
+train: [7]  [360/400]  eta: 0:00:16  lr: 0.000273  loss: 1.8128 (1.8317)  grad: 0.0557 (0.0549)  time: 0.3672  data: 0.0033  max mem: 3956
+train: [7]  [380/400]  eta: 0:00:08  lr: 0.000272  loss: 1.8286 (1.8304)  grad: 0.0557 (0.0552)  time: 0.3615  data: 0.0034  max mem: 3956
+train: [7]  [399/400]  eta: 0:00:00  lr: 0.000271  loss: 1.8149 (1.8296)  grad: 0.0565 (0.0553)  time: 0.3770  data: 0.0033  max mem: 3956
+train: [7] Total time: 0:02:41 (0.4030 s / it)
+train: [7] Summary: lr: 0.000271  loss: 1.8149 (1.8296)  grad: 0.0565 (0.0553)
+eval (validation): [7]  [ 0/63]  eta: 0:03:55    time: 3.7348  data: 3.4152  max mem: 3956
+eval (validation): [7]  [20/63]  eta: 0:00:23    time: 0.3908  data: 0.0041  max mem: 3956
+eval (validation): [7]  [40/63]  eta: 0:00:10    time: 0.3740  data: 0.0030  max mem: 3956
+eval (validation): [7]  [60/63]  eta: 0:00:01    time: 0.3654  data: 0.0032  max mem: 3956
+eval (validation): [7]  [62/63]  eta: 0:00:00    time: 0.3618  data: 0.0032  max mem: 3956
+eval (validation): [7] Total time: 0:00:27 (0.4332 s / it)
+cv: [7]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.439  acc: 0.897  f1: 0.872
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+train: [8]  [  0/400]  eta: 0:24:20  lr: nan  time: 3.6522  data: 3.3708  max mem: 3956
+train: [8]  [ 20/400]  eta: 0:03:25  lr: 0.000270  loss: 1.7738 (1.7676)  grad: 0.0509 (0.0520)  time: 0.3863  data: 0.0036  max mem: 3956
+train: [8]  [ 40/400]  eta: 0:02:44  lr: 0.000270  loss: 1.7958 (1.8000)  grad: 0.0517 (0.0550)  time: 0.3695  data: 0.0030  max mem: 3956
+train: [8]  [ 60/400]  eta: 0:02:25  lr: 0.000269  loss: 1.8185 (1.8046)  grad: 0.0536 (0.0555)  time: 0.3704  data: 0.0035  max mem: 3956
+train: [8]  [ 80/400]  eta: 0:02:09  lr: 0.000268  loss: 1.8310 (1.8119)  grad: 0.0527 (0.0545)  time: 0.3355  data: 0.0033  max mem: 3956
+train: [8]  [100/400]  eta: 0:01:58  lr: 0.000267  loss: 1.8067 (1.8043)  grad: 0.0508 (0.0540)  time: 0.3518  data: 0.0035  max mem: 3956
+train: [8]  [120/400]  eta: 0:01:49  lr: 0.000266  loss: 1.7793 (1.8020)  grad: 0.0522 (0.0537)  time: 0.3608  data: 0.0035  max mem: 3956
+train: [8]  [140/400]  eta: 0:01:40  lr: 0.000265  loss: 1.7958 (1.8009)  grad: 0.0534 (0.0538)  time: 0.3616  data: 0.0033  max mem: 3956
+train: [8]  [160/400]  eta: 0:01:31  lr: 0.000264  loss: 1.7710 (1.7930)  grad: 0.0515 (0.0536)  time: 0.3580  data: 0.0034  max mem: 3956
+train: [8]  [180/400]  eta: 0:01:23  lr: 0.000263  loss: 1.7448 (1.7881)  grad: 0.0546 (0.0543)  time: 0.3601  data: 0.0033  max mem: 3956
+train: [8]  [200/400]  eta: 0:01:15  lr: 0.000262  loss: 1.7708 (1.7881)  grad: 0.0568 (0.0543)  time: 0.3575  data: 0.0032  max mem: 3956
+train: [8]  [220/400]  eta: 0:01:07  lr: 0.000260  loss: 1.7915 (1.7889)  grad: 0.0543 (0.0546)  time: 0.3557  data: 0.0035  max mem: 3956
+train: [8]  [240/400]  eta: 0:00:59  lr: 0.000259  loss: 1.7638 (1.7841)  grad: 0.0543 (0.0547)  time: 0.3610  data: 0.0037  max mem: 3956
+train: [8]  [260/400]  eta: 0:00:52  lr: 0.000258  loss: 1.7432 (1.7819)  grad: 0.0533 (0.0546)  time: 0.3658  data: 0.0035  max mem: 3956
+train: [8]  [280/400]  eta: 0:00:44  lr: 0.000257  loss: 1.7435 (1.7795)  grad: 0.0518 (0.0547)  time: 0.3709  data: 0.0033  max mem: 3956
+train: [8]  [300/400]  eta: 0:00:38  lr: 0.000256  loss: 1.7452 (1.7785)  grad: 0.0516 (0.0544)  time: 0.5332  data: 0.1957  max mem: 3956
+train: [8]  [320/400]  eta: 0:00:30  lr: 0.000255  loss: 1.7569 (1.7766)  grad: 0.0495 (0.0541)  time: 0.3653  data: 0.0037  max mem: 3956
+train: [8]  [340/400]  eta: 0:00:22  lr: 0.000254  loss: 1.7569 (1.7754)  grad: 0.0498 (0.0539)  time: 0.3599  data: 0.0033  max mem: 3956
+train: [8]  [360/400]  eta: 0:00:15  lr: 0.000253  loss: 1.7267 (1.7734)  grad: 0.0509 (0.0538)  time: 0.3526  data: 0.0033  max mem: 3956
+train: [8]  [380/400]  eta: 0:00:07  lr: 0.000252  loss: 1.7350 (1.7725)  grad: 0.0520 (0.0537)  time: 0.3380  data: 0.0032  max mem: 3956
+train: [8]  [399/400]  eta: 0:00:00  lr: 0.000250  loss: 1.7686 (1.7727)  grad: 0.0528 (0.0538)  time: 0.3618  data: 0.0036  max mem: 3956
+train: [8] Total time: 0:02:30 (0.3773 s / it)
+train: [8] Summary: lr: 0.000250  loss: 1.7686 (1.7727)  grad: 0.0528 (0.0538)
+eval (validation): [8]  [ 0/63]  eta: 0:03:37    time: 3.4582  data: 3.2299  max mem: 3956
+eval (validation): [8]  [20/63]  eta: 0:00:21    time: 0.3473  data: 0.0044  max mem: 3956
+eval (validation): [8]  [40/63]  eta: 0:00:09    time: 0.3691  data: 0.0034  max mem: 3956
+eval (validation): [8]  [60/63]  eta: 0:00:01    time: 0.3460  data: 0.0030  max mem: 3956
+eval (validation): [8]  [62/63]  eta: 0:00:00    time: 0.3434  data: 0.0032  max mem: 3956
+eval (validation): [8] Total time: 0:00:25 (0.4074 s / it)
+cv: [8]  best hparam: (26, 1.0) (044) ('044_lr2.6e+01_wd1.0e+00')  loss: 0.469  acc: 0.907  f1: 0.896
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+train: [9]  [  0/400]  eta: 0:24:00  lr: nan  time: 3.6003  data: 3.2898  max mem: 3956
+train: [9]  [ 20/400]  eta: 0:03:12  lr: 0.000249  loss: 1.7352 (1.7469)  grad: 0.0531 (0.0530)  time: 0.3515  data: 0.0028  max mem: 3956
+train: [9]  [ 40/400]  eta: 0:02:39  lr: 0.000248  loss: 1.7352 (1.7467)  grad: 0.0509 (0.0518)  time: 0.3785  data: 0.0033  max mem: 3956
+train: [9]  [ 60/400]  eta: 0:02:24  lr: 0.000247  loss: 1.7195 (1.7392)  grad: 0.0489 (0.0516)  time: 0.3859  data: 0.0035  max mem: 3956
+train: [9]  [ 80/400]  eta: 0:02:11  lr: 0.000246  loss: 1.7152 (1.7361)  grad: 0.0541 (0.0528)  time: 0.3668  data: 0.0034  max mem: 3956
+train: [9]  [100/400]  eta: 0:01:59  lr: 0.000244  loss: 1.7063 (1.7281)  grad: 0.0541 (0.0526)  time: 0.3447  data: 0.0033  max mem: 3956
+train: [9]  [120/400]  eta: 0:01:50  lr: 0.000243  loss: 1.7063 (1.7312)  grad: 0.0494 (0.0521)  time: 0.3719  data: 0.0033  max mem: 3956
+train: [9]  [140/400]  eta: 0:01:42  lr: 0.000242  loss: 1.7598 (1.7350)  grad: 0.0496 (0.0523)  time: 0.3935  data: 0.0033  max mem: 3956
+train: [9]  [160/400]  eta: 0:01:34  lr: 0.000241  loss: 1.7262 (1.7311)  grad: 0.0502 (0.0521)  time: 0.3902  data: 0.0033  max mem: 3956
+train: [9]  [180/400]  eta: 0:01:26  lr: 0.000240  loss: 1.7110 (1.7332)  grad: 0.0514 (0.0519)  time: 0.3902  data: 0.0033  max mem: 3956
+train: [9]  [200/400]  eta: 0:01:18  lr: 0.000238  loss: 1.7479 (1.7356)  grad: 0.0508 (0.0522)  time: 0.4061  data: 0.0034  max mem: 3956
+train: [9]  [220/400]  eta: 0:01:10  lr: 0.000237  loss: 1.7541 (1.7365)  grad: 0.0516 (0.0523)  time: 0.3964  data: 0.0032  max mem: 3956
+train: [9]  [240/400]  eta: 0:01:02  lr: 0.000236  loss: 1.7123 (1.7337)  grad: 0.0495 (0.0521)  time: 0.3754  data: 0.0035  max mem: 3956
+train: [9]  [260/400]  eta: 0:00:55  lr: 0.000234  loss: 1.6978 (1.7314)  grad: 0.0511 (0.0521)  time: 0.4038  data: 0.0036  max mem: 3956
+train: [9]  [280/400]  eta: 0:00:47  lr: 0.000233  loss: 1.7276 (1.7325)  grad: 0.0511 (0.0520)  time: 0.4271  data: 0.0035  max mem: 3956
+train: [9]  [300/400]  eta: 0:00:40  lr: 0.000232  loss: 1.7187 (1.7303)  grad: 0.0473 (0.0519)  time: 0.5728  data: 0.1879  max mem: 3956
+train: [9]  [320/400]  eta: 0:00:32  lr: 0.000230  loss: 1.6939 (1.7283)  grad: 0.0530 (0.0522)  time: 0.4569  data: 0.0040  max mem: 3956
+train: [9]  [340/400]  eta: 0:00:24  lr: 0.000229  loss: 1.6876 (1.7270)  grad: 0.0530 (0.0520)  time: 0.4165  data: 0.0033  max mem: 3956
+train: [9]  [360/400]  eta: 0:00:16  lr: 0.000228  loss: 1.7151 (1.7273)  grad: 0.0523 (0.0521)  time: 0.4018  data: 0.0035  max mem: 3956
+train: [9]  [380/400]  eta: 0:00:08  lr: 0.000226  loss: 1.7158 (1.7268)  grad: 0.0539 (0.0521)  time: 0.3748  data: 0.0033  max mem: 3956
+train: [9]  [399/400]  eta: 0:00:00  lr: 0.000225  loss: 1.7203 (1.7277)  grad: 0.0515 (0.0521)  time: 0.3682  data: 0.0035  max mem: 3956
+train: [9] Total time: 0:02:42 (0.4070 s / it)
+train: [9] Summary: lr: 0.000225  loss: 1.7203 (1.7277)  grad: 0.0515 (0.0521)
+eval (validation): [9]  [ 0/63]  eta: 0:03:53    time: 3.7124  data: 3.3933  max mem: 3956
+eval (validation): [9]  [20/63]  eta: 0:00:25    time: 0.4381  data: 0.0038  max mem: 3956
+eval (validation): [9]  [40/63]  eta: 0:00:11    time: 0.3723  data: 0.0035  max mem: 3956
+eval (validation): [9]  [60/63]  eta: 0:00:01    time: 0.3560  data: 0.0034  max mem: 3956
+eval (validation): [9]  [62/63]  eta: 0:00:00    time: 0.3595  data: 0.0034  max mem: 3956
+eval (validation): [9] Total time: 0:00:28 (0.4457 s / it)
+cv: [9]  best hparam: (43, 1.0) (047) ('047_lr4.3e+01_wd1.0e+00')  loss: 0.395  acc: 0.921  f1: 0.908
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [10]  [  0/400]  eta: 0:22:25  lr: nan  time: 3.3631  data: 3.1362  max mem: 3956
+train: [10]  [ 20/400]  eta: 0:03:18  lr: 0.000224  loss: 1.7470 (1.7583)  grad: 0.0523 (0.0523)  time: 0.3803  data: 0.0039  max mem: 3956
+train: [10]  [ 40/400]  eta: 0:02:41  lr: 0.000222  loss: 1.6973 (1.7121)  grad: 0.0518 (0.0525)  time: 0.3697  data: 0.0030  max mem: 3956
+train: [10]  [ 60/400]  eta: 0:02:25  lr: 0.000221  loss: 1.6710 (1.7087)  grad: 0.0513 (0.0521)  time: 0.3841  data: 0.0033  max mem: 3956
+train: [10]  [ 80/400]  eta: 0:02:11  lr: 0.000220  loss: 1.7079 (1.7049)  grad: 0.0522 (0.0523)  time: 0.3623  data: 0.0035  max mem: 3956
+train: [10]  [100/400]  eta: 0:02:00  lr: 0.000218  loss: 1.6989 (1.7050)  grad: 0.0508 (0.0514)  time: 0.3698  data: 0.0034  max mem: 3956
+train: [10]  [120/400]  eta: 0:01:51  lr: 0.000217  loss: 1.7184 (1.7048)  grad: 0.0505 (0.0516)  time: 0.3689  data: 0.0035  max mem: 3956
+train: [10]  [140/400]  eta: 0:01:41  lr: 0.000215  loss: 1.6718 (1.7024)  grad: 0.0490 (0.0510)  time: 0.3574  data: 0.0034  max mem: 3956
+train: [10]  [160/400]  eta: 0:01:33  lr: 0.000214  loss: 1.6654 (1.6974)  grad: 0.0477 (0.0511)  time: 0.3633  data: 0.0036  max mem: 3956
+train: [10]  [180/400]  eta: 0:01:24  lr: 0.000213  loss: 1.6907 (1.7036)  grad: 0.0472 (0.0507)  time: 0.3548  data: 0.0034  max mem: 3956
+train: [10]  [200/400]  eta: 0:01:16  lr: 0.000211  loss: 1.7050 (1.7038)  grad: 0.0476 (0.0509)  time: 0.3600  data: 0.0035  max mem: 3956
+train: [10]  [220/400]  eta: 0:01:08  lr: 0.000210  loss: 1.6929 (1.7021)  grad: 0.0516 (0.0510)  time: 0.3668  data: 0.0035  max mem: 3956
+train: [10]  [240/400]  eta: 0:01:00  lr: 0.000208  loss: 1.6774 (1.6978)  grad: 0.0516 (0.0510)  time: 0.3717  data: 0.0035  max mem: 3956
+train: [10]  [260/400]  eta: 0:00:53  lr: 0.000207  loss: 1.6549 (1.6949)  grad: 0.0526 (0.0512)  time: 0.3707  data: 0.0033  max mem: 3956
+train: [10]  [280/400]  eta: 0:00:45  lr: 0.000205  loss: 1.6826 (1.6937)  grad: 0.0550 (0.0517)  time: 0.3824  data: 0.0035  max mem: 3956
+train: [10]  [300/400]  eta: 0:00:38  lr: 0.000204  loss: 1.6859 (1.6934)  grad: 0.0534 (0.0515)  time: 0.5251  data: 0.1894  max mem: 3956
+train: [10]  [320/400]  eta: 0:00:31  lr: 0.000202  loss: 1.6859 (1.6941)  grad: 0.0505 (0.0516)  time: 0.3783  data: 0.0039  max mem: 3956
+train: [10]  [340/400]  eta: 0:00:23  lr: 0.000201  loss: 1.6661 (1.6914)  grad: 0.0503 (0.0515)  time: 0.3811  data: 0.0029  max mem: 3956
+train: [10]  [360/400]  eta: 0:00:15  lr: 0.000199  loss: 1.6594 (1.6904)  grad: 0.0491 (0.0514)  time: 0.3801  data: 0.0034  max mem: 3956
+train: [10]  [380/400]  eta: 0:00:07  lr: 0.000198  loss: 1.6594 (1.6875)  grad: 0.0500 (0.0514)  time: 0.3619  data: 0.0033  max mem: 3956
+train: [10]  [399/400]  eta: 0:00:00  lr: 0.000196  loss: 1.6212 (1.6840)  grad: 0.0548 (0.0517)  time: 0.3714  data: 0.0034  max mem: 3956
+train: [10] Total time: 0:02:34 (0.3858 s / it)
+train: [10] Summary: lr: 0.000196  loss: 1.6212 (1.6840)  grad: 0.0548 (0.0517)
+eval (validation): [10]  [ 0/63]  eta: 0:03:57    time: 3.7734  data: 3.4727  max mem: 3956
+eval (validation): [10]  [20/63]  eta: 0:00:25    time: 0.4244  data: 0.0047  max mem: 3956
+eval (validation): [10]  [40/63]  eta: 0:00:10    time: 0.3547  data: 0.0030  max mem: 3956
+eval (validation): [10]  [60/63]  eta: 0:00:01    time: 0.3334  data: 0.0035  max mem: 3956
+eval (validation): [10]  [62/63]  eta: 0:00:00    time: 0.3331  data: 0.0035  max mem: 3956
+eval (validation): [10] Total time: 0:00:26 (0.4283 s / it)
+cv: [10]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.377  acc: 0.924  f1: 0.915
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [11]  [  0/400]  eta: 0:23:15  lr: nan  time: 3.4882  data: 3.2618  max mem: 3956
+train: [11]  [ 20/400]  eta: 0:03:24  lr: 0.000195  loss: 1.6671 (1.6665)  grad: 0.0505 (0.0498)  time: 0.3901  data: 0.0029  max mem: 3956
+train: [11]  [ 40/400]  eta: 0:02:46  lr: 0.000193  loss: 1.6523 (1.6458)  grad: 0.0505 (0.0516)  time: 0.3848  data: 0.0033  max mem: 3956
+train: [11]  [ 60/400]  eta: 0:02:25  lr: 0.000192  loss: 1.6446 (1.6509)  grad: 0.0488 (0.0509)  time: 0.3522  data: 0.0033  max mem: 3956
+train: [11]  [ 80/400]  eta: 0:02:11  lr: 0.000190  loss: 1.6414 (1.6487)  grad: 0.0488 (0.0508)  time: 0.3606  data: 0.0033  max mem: 3956
+train: [11]  [100/400]  eta: 0:01:59  lr: 0.000189  loss: 1.6316 (1.6518)  grad: 0.0502 (0.0513)  time: 0.3560  data: 0.0033  max mem: 3956
+train: [11]  [120/400]  eta: 0:01:50  lr: 0.000187  loss: 1.6661 (1.6529)  grad: 0.0540 (0.0512)  time: 0.3685  data: 0.0035  max mem: 3956
+train: [11]  [140/400]  eta: 0:01:41  lr: 0.000186  loss: 1.6839 (1.6635)  grad: 0.0531 (0.0513)  time: 0.3579  data: 0.0033  max mem: 3956
+train: [11]  [160/400]  eta: 0:01:32  lr: 0.000184  loss: 1.6836 (1.6615)  grad: 0.0485 (0.0508)  time: 0.3671  data: 0.0035  max mem: 3956
+train: [11]  [180/400]  eta: 0:01:24  lr: 0.000183  loss: 1.6577 (1.6608)  grad: 0.0480 (0.0509)  time: 0.3622  data: 0.0033  max mem: 3956
+train: [11]  [200/400]  eta: 0:01:16  lr: 0.000181  loss: 1.6667 (1.6614)  grad: 0.0515 (0.0511)  time: 0.3613  data: 0.0033  max mem: 3956
+train: [11]  [220/400]  eta: 0:01:08  lr: 0.000180  loss: 1.6626 (1.6640)  grad: 0.0486 (0.0508)  time: 0.3638  data: 0.0033  max mem: 3956
+train: [11]  [240/400]  eta: 0:01:00  lr: 0.000178  loss: 1.6644 (1.6660)  grad: 0.0486 (0.0507)  time: 0.3698  data: 0.0035  max mem: 3956
+train: [11]  [260/400]  eta: 0:00:52  lr: 0.000177  loss: 1.6503 (1.6653)  grad: 0.0493 (0.0506)  time: 0.3598  data: 0.0035  max mem: 3956
+train: [11]  [280/400]  eta: 0:00:45  lr: 0.000175  loss: 1.6147 (1.6626)  grad: 0.0465 (0.0504)  time: 0.3471  data: 0.0036  max mem: 3956
+train: [11]  [300/400]  eta: 0:00:38  lr: 0.000174  loss: 1.6147 (1.6613)  grad: 0.0465 (0.0501)  time: 0.5652  data: 0.1993  max mem: 3956
+train: [11]  [320/400]  eta: 0:00:31  lr: 0.000172  loss: 1.6476 (1.6620)  grad: 0.0469 (0.0501)  time: 0.3872  data: 0.0037  max mem: 3956
+train: [11]  [340/400]  eta: 0:00:23  lr: 0.000170  loss: 1.6476 (1.6613)  grad: 0.0479 (0.0501)  time: 0.3623  data: 0.0032  max mem: 3956
+train: [11]  [360/400]  eta: 0:00:15  lr: 0.000169  loss: 1.6551 (1.6615)  grad: 0.0514 (0.0503)  time: 0.3578  data: 0.0034  max mem: 3956
+train: [11]  [380/400]  eta: 0:00:07  lr: 0.000167  loss: 1.6657 (1.6601)  grad: 0.0514 (0.0503)  time: 0.3869  data: 0.0037  max mem: 3956
+train: [11]  [399/400]  eta: 0:00:00  lr: 0.000166  loss: 1.6633 (1.6596)  grad: 0.0472 (0.0502)  time: 0.3452  data: 0.0035  max mem: 3956
+train: [11] Total time: 0:02:33 (0.3832 s / it)
+train: [11] Summary: lr: 0.000166  loss: 1.6633 (1.6596)  grad: 0.0472 (0.0502)
+eval (validation): [11]  [ 0/63]  eta: 0:03:43    time: 3.5403  data: 3.2599  max mem: 3956
+eval (validation): [11]  [20/63]  eta: 0:00:23    time: 0.3938  data: 0.0037  max mem: 3956
+eval (validation): [11]  [40/63]  eta: 0:00:10    time: 0.3411  data: 0.0036  max mem: 3956
+eval (validation): [11]  [60/63]  eta: 0:00:01    time: 0.3297  data: 0.0031  max mem: 3956
+eval (validation): [11]  [62/63]  eta: 0:00:00    time: 0.3283  data: 0.0029  max mem: 3956
+eval (validation): [11] Total time: 0:00:25 (0.4091 s / it)
+cv: [11]  best hparam: (43, 1.0) (047) ('047_lr4.3e+01_wd1.0e+00')  loss: 0.385  acc: 0.920  f1: 0.911
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+train: [12]  [  0/400]  eta: 0:26:13  lr: nan  time: 3.9344  data: 3.6976  max mem: 3956
+train: [12]  [ 20/400]  eta: 0:03:22  lr: 0.000164  loss: 1.5948 (1.6058)  grad: 0.0538 (0.0540)  time: 0.3633  data: 0.0028  max mem: 3956
+train: [12]  [ 40/400]  eta: 0:02:40  lr: 0.000163  loss: 1.6326 (1.6365)  grad: 0.0532 (0.0504)  time: 0.3542  data: 0.0029  max mem: 3956
+train: [12]  [ 60/400]  eta: 0:02:21  lr: 0.000161  loss: 1.6366 (1.6319)  grad: 0.0501 (0.0516)  time: 0.3577  data: 0.0035  max mem: 3956
+train: [12]  [ 80/400]  eta: 0:02:09  lr: 0.000160  loss: 1.6371 (1.6378)  grad: 0.0491 (0.0505)  time: 0.3624  data: 0.0034  max mem: 3956
+train: [12]  [100/400]  eta: 0:01:59  lr: 0.000158  loss: 1.6310 (1.6345)  grad: 0.0488 (0.0507)  time: 0.3838  data: 0.0033  max mem: 3956
+train: [12]  [120/400]  eta: 0:01:50  lr: 0.000156  loss: 1.6037 (1.6317)  grad: 0.0498 (0.0504)  time: 0.3735  data: 0.0035  max mem: 3956
+train: [12]  [140/400]  eta: 0:01:42  lr: 0.000155  loss: 1.5995 (1.6297)  grad: 0.0505 (0.0509)  time: 0.3789  data: 0.0034  max mem: 3956
+train: [12]  [160/400]  eta: 0:01:33  lr: 0.000153  loss: 1.6118 (1.6308)  grad: 0.0505 (0.0508)  time: 0.3712  data: 0.0034  max mem: 3956
+train: [12]  [180/400]  eta: 0:01:25  lr: 0.000152  loss: 1.6337 (1.6339)  grad: 0.0469 (0.0505)  time: 0.3705  data: 0.0038  max mem: 3956
+train: [12]  [200/400]  eta: 0:01:17  lr: 0.000150  loss: 1.6549 (1.6370)  grad: 0.0481 (0.0505)  time: 0.3748  data: 0.0035  max mem: 3956
+train: [12]  [220/400]  eta: 0:01:09  lr: 0.000149  loss: 1.6661 (1.6382)  grad: 0.0485 (0.0504)  time: 0.3561  data: 0.0034  max mem: 3956
+train: [12]  [240/400]  eta: 0:01:01  lr: 0.000147  loss: 1.6558 (1.6396)  grad: 0.0489 (0.0503)  time: 0.3693  data: 0.0033  max mem: 3956
+train: [12]  [260/400]  eta: 0:00:53  lr: 0.000145  loss: 1.6288 (1.6397)  grad: 0.0522 (0.0507)  time: 0.3656  data: 0.0033  max mem: 3956
+train: [12]  [280/400]  eta: 0:00:45  lr: 0.000144  loss: 1.6236 (1.6364)  grad: 0.0535 (0.0506)  time: 0.3393  data: 0.0035  max mem: 3956
+train: [12]  [300/400]  eta: 0:00:39  lr: 0.000142  loss: 1.6007 (1.6349)  grad: 0.0489 (0.0505)  time: 0.5523  data: 0.1932  max mem: 3956
+train: [12]  [320/400]  eta: 0:00:31  lr: 0.000141  loss: 1.6056 (1.6362)  grad: 0.0503 (0.0505)  time: 0.3622  data: 0.0033  max mem: 3956
+train: [12]  [340/400]  eta: 0:00:23  lr: 0.000139  loss: 1.6059 (1.6345)  grad: 0.0479 (0.0502)  time: 0.3688  data: 0.0032  max mem: 3956
+train: [12]  [360/400]  eta: 0:00:15  lr: 0.000138  loss: 1.6069 (1.6356)  grad: 0.0479 (0.0504)  time: 0.3540  data: 0.0034  max mem: 3956
+train: [12]  [380/400]  eta: 0:00:07  lr: 0.000136  loss: 1.6131 (1.6333)  grad: 0.0515 (0.0505)  time: 0.4086  data: 0.0034  max mem: 3956
+train: [12]  [399/400]  eta: 0:00:00  lr: 0.000134  loss: 1.6133 (1.6339)  grad: 0.0503 (0.0503)  time: 0.3738  data: 0.0035  max mem: 3956
+train: [12] Total time: 0:02:34 (0.3862 s / it)
+train: [12] Summary: lr: 0.000134  loss: 1.6133 (1.6339)  grad: 0.0503 (0.0503)
+eval (validation): [12]  [ 0/63]  eta: 0:03:42    time: 3.5309  data: 3.2356  max mem: 3956
+eval (validation): [12]  [20/63]  eta: 0:00:23    time: 0.3854  data: 0.0291  max mem: 3956
+eval (validation): [12]  [40/63]  eta: 0:00:10    time: 0.3618  data: 0.0035  max mem: 3956
+eval (validation): [12]  [60/63]  eta: 0:00:01    time: 0.3200  data: 0.0033  max mem: 3956
+eval (validation): [12]  [62/63]  eta: 0:00:00    time: 0.3198  data: 0.0031  max mem: 3956
+eval (validation): [12] Total time: 0:00:25 (0.4113 s / it)
+cv: [12]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.363  acc: 0.926  f1: 0.914
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [13]  [  0/400]  eta: 0:22:47  lr: nan  time: 3.4189  data: 3.1868  max mem: 3956
+train: [13]  [ 20/400]  eta: 0:03:26  lr: 0.000133  loss: 1.5923 (1.6336)  grad: 0.0467 (0.0472)  time: 0.4007  data: 0.0026  max mem: 3956
+train: [13]  [ 40/400]  eta: 0:02:44  lr: 0.000131  loss: 1.5923 (1.6136)  grad: 0.0456 (0.0461)  time: 0.3663  data: 0.0034  max mem: 3956
+train: [13]  [ 60/400]  eta: 0:02:24  lr: 0.000130  loss: 1.5921 (1.6164)  grad: 0.0453 (0.0467)  time: 0.3613  data: 0.0034  max mem: 3956
+train: [13]  [ 80/400]  eta: 0:02:10  lr: 0.000128  loss: 1.6256 (1.6192)  grad: 0.0488 (0.0477)  time: 0.3513  data: 0.0033  max mem: 3956
+train: [13]  [100/400]  eta: 0:01:59  lr: 0.000127  loss: 1.6332 (1.6200)  grad: 0.0481 (0.0475)  time: 0.3672  data: 0.0037  max mem: 3956
+train: [13]  [120/400]  eta: 0:01:49  lr: 0.000125  loss: 1.6257 (1.6205)  grad: 0.0481 (0.0472)  time: 0.3572  data: 0.0033  max mem: 3956
+train: [13]  [140/400]  eta: 0:01:40  lr: 0.000124  loss: 1.5942 (1.6175)  grad: 0.0489 (0.0477)  time: 0.3564  data: 0.0035  max mem: 3956
+train: [13]  [160/400]  eta: 0:01:32  lr: 0.000122  loss: 1.5953 (1.6189)  grad: 0.0484 (0.0477)  time: 0.3637  data: 0.0033  max mem: 3956
+train: [13]  [180/400]  eta: 0:01:23  lr: 0.000120  loss: 1.5953 (1.6133)  grad: 0.0481 (0.0480)  time: 0.3578  data: 0.0034  max mem: 3956
+train: [13]  [200/400]  eta: 0:01:16  lr: 0.000119  loss: 1.5880 (1.6146)  grad: 0.0502 (0.0484)  time: 0.3736  data: 0.0034  max mem: 3956
+train: [13]  [220/400]  eta: 0:01:08  lr: 0.000117  loss: 1.6098 (1.6146)  grad: 0.0502 (0.0485)  time: 0.3716  data: 0.0034  max mem: 3956
+train: [13]  [240/400]  eta: 0:01:00  lr: 0.000116  loss: 1.5974 (1.6140)  grad: 0.0476 (0.0487)  time: 0.3643  data: 0.0034  max mem: 3956
+train: [13]  [260/400]  eta: 0:00:52  lr: 0.000114  loss: 1.6031 (1.6130)  grad: 0.0474 (0.0485)  time: 0.3621  data: 0.0034  max mem: 3956
+train: [13]  [280/400]  eta: 0:00:45  lr: 0.000113  loss: 1.5924 (1.6127)  grad: 0.0429 (0.0484)  time: 0.3505  data: 0.0033  max mem: 3956
+train: [13]  [300/400]  eta: 0:00:38  lr: 0.000111  loss: 1.5924 (1.6125)  grad: 0.0435 (0.0482)  time: 0.5630  data: 0.1884  max mem: 3956
+train: [13]  [320/400]  eta: 0:00:30  lr: 0.000110  loss: 1.6430 (1.6154)  grad: 0.0465 (0.0483)  time: 0.3703  data: 0.0036  max mem: 3956
+train: [13]  [340/400]  eta: 0:00:23  lr: 0.000108  loss: 1.6249 (1.6129)  grad: 0.0459 (0.0480)  time: 0.3703  data: 0.0031  max mem: 3956
+train: [13]  [360/400]  eta: 0:00:15  lr: 0.000107  loss: 1.6163 (1.6144)  grad: 0.0444 (0.0480)  time: 0.3541  data: 0.0032  max mem: 3956
+train: [13]  [380/400]  eta: 0:00:07  lr: 0.000105  loss: 1.6336 (1.6144)  grad: 0.0460 (0.0480)  time: 0.3694  data: 0.0034  max mem: 3956
+train: [13]  [399/400]  eta: 0:00:00  lr: 0.000104  loss: 1.5822 (1.6129)  grad: 0.0466 (0.0479)  time: 0.3675  data: 0.0037  max mem: 3956
+train: [13] Total time: 0:02:33 (0.3828 s / it)
+train: [13] Summary: lr: 0.000104  loss: 1.5822 (1.6129)  grad: 0.0466 (0.0479)
+eval (validation): [13]  [ 0/63]  eta: 0:03:33    time: 3.3908  data: 3.1174  max mem: 3956
+eval (validation): [13]  [20/63]  eta: 0:00:21    time: 0.3674  data: 0.0039  max mem: 3956
+eval (validation): [13]  [40/63]  eta: 0:00:09    time: 0.3541  data: 0.0033  max mem: 3956
+eval (validation): [13]  [60/63]  eta: 0:00:01    time: 0.3370  data: 0.0033  max mem: 3956
+eval (validation): [13]  [62/63]  eta: 0:00:00    time: 0.3285  data: 0.0033  max mem: 3956
+eval (validation): [13] Total time: 0:00:25 (0.4047 s / it)
+cv: [13]  best hparam: (43, 1.0) (047) ('047_lr4.3e+01_wd1.0e+00')  loss: 0.363  acc: 0.926  f1: 0.917
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+train: [14]  [  0/400]  eta: 0:22:56  lr: nan  time: 3.4408  data: 3.1544  max mem: 3956
+train: [14]  [ 20/400]  eta: 0:03:29  lr: 0.000102  loss: 1.5654 (1.5819)  grad: 0.0521 (0.0533)  time: 0.4070  data: 0.0040  max mem: 3956
+train: [14]  [ 40/400]  eta: 0:02:45  lr: 0.000101  loss: 1.5890 (1.5914)  grad: 0.0513 (0.0506)  time: 0.3637  data: 0.0031  max mem: 3956
+train: [14]  [ 60/400]  eta: 0:02:24  lr: 0.000099  loss: 1.5932 (1.5969)  grad: 0.0474 (0.0495)  time: 0.3537  data: 0.0036  max mem: 3956
+train: [14]  [ 80/400]  eta: 0:02:10  lr: 0.000098  loss: 1.5946 (1.6058)  grad: 0.0466 (0.0491)  time: 0.3596  data: 0.0035  max mem: 3956
+train: [14]  [100/400]  eta: 0:02:00  lr: 0.000096  loss: 1.5941 (1.6049)  grad: 0.0481 (0.0490)  time: 0.3794  data: 0.0035  max mem: 3956
+train: [14]  [120/400]  eta: 0:01:52  lr: 0.000095  loss: 1.5975 (1.6048)  grad: 0.0512 (0.0493)  time: 0.3903  data: 0.0037  max mem: 3956
+train: [14]  [140/400]  eta: 0:01:43  lr: 0.000093  loss: 1.6061 (1.6071)  grad: 0.0504 (0.0493)  time: 0.3743  data: 0.0034  max mem: 3956
+train: [14]  [160/400]  eta: 0:01:34  lr: 0.000092  loss: 1.5766 (1.6034)  grad: 0.0475 (0.0492)  time: 0.3593  data: 0.0037  max mem: 3956
+train: [14]  [180/400]  eta: 0:01:25  lr: 0.000090  loss: 1.5766 (1.6010)  grad: 0.0470 (0.0490)  time: 0.3699  data: 0.0036  max mem: 3956
+train: [14]  [200/400]  eta: 0:01:17  lr: 0.000089  loss: 1.6223 (1.6040)  grad: 0.0470 (0.0490)  time: 0.3710  data: 0.0035  max mem: 3956
+train: [14]  [220/400]  eta: 0:01:09  lr: 0.000088  loss: 1.6223 (1.6051)  grad: 0.0495 (0.0494)  time: 0.3909  data: 0.0034  max mem: 3956
+train: [14]  [240/400]  eta: 0:01:01  lr: 0.000086  loss: 1.5933 (1.6046)  grad: 0.0495 (0.0493)  time: 0.3728  data: 0.0035  max mem: 3956
+train: [14]  [260/400]  eta: 0:00:54  lr: 0.000085  loss: 1.5856 (1.6032)  grad: 0.0474 (0.0490)  time: 0.3723  data: 0.0034  max mem: 3956
+train: [14]  [280/400]  eta: 0:00:46  lr: 0.000083  loss: 1.5824 (1.6026)  grad: 0.0479 (0.0491)  time: 0.3511  data: 0.0035  max mem: 3956
+train: [14]  [300/400]  eta: 0:00:39  lr: 0.000082  loss: 1.5896 (1.6026)  grad: 0.0518 (0.0492)  time: 0.5736  data: 0.2072  max mem: 3956
+train: [14]  [320/400]  eta: 0:00:31  lr: 0.000081  loss: 1.5809 (1.6022)  grad: 0.0535 (0.0494)  time: 0.3986  data: 0.0044  max mem: 3956
+train: [14]  [340/400]  eta: 0:00:23  lr: 0.000079  loss: 1.5749 (1.6003)  grad: 0.0523 (0.0495)  time: 0.3570  data: 0.0023  max mem: 3956
+train: [14]  [360/400]  eta: 0:00:15  lr: 0.000078  loss: 1.5582 (1.5977)  grad: 0.0506 (0.0497)  time: 0.3583  data: 0.0032  max mem: 3956
+train: [14]  [380/400]  eta: 0:00:07  lr: 0.000076  loss: 1.5847 (1.5993)  grad: 0.0498 (0.0497)  time: 0.3731  data: 0.0035  max mem: 3956
+train: [14]  [399/400]  eta: 0:00:00  lr: 0.000075  loss: 1.6240 (1.6014)  grad: 0.0472 (0.0496)  time: 0.3848  data: 0.0037  max mem: 3956
+train: [14] Total time: 0:02:36 (0.3911 s / it)
+train: [14] Summary: lr: 0.000075  loss: 1.6240 (1.6014)  grad: 0.0472 (0.0496)
+eval (validation): [14]  [ 0/63]  eta: 0:03:34    time: 3.4109  data: 3.1326  max mem: 3956
+eval (validation): [14]  [20/63]  eta: 0:00:21    time: 0.3493  data: 0.0036  max mem: 3956
+eval (validation): [14]  [40/63]  eta: 0:00:10    time: 0.3731  data: 0.0041  max mem: 3956
+eval (validation): [14]  [60/63]  eta: 0:00:01    time: 0.3409  data: 0.0033  max mem: 3956
+eval (validation): [14]  [62/63]  eta: 0:00:00    time: 0.3338  data: 0.0032  max mem: 3956
+eval (validation): [14] Total time: 0:00:25 (0.4076 s / it)
+cv: [14]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.355  acc: 0.926  f1: 0.916
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+train: [15]  [  0/400]  eta: 0:22:45  lr: nan  time: 3.4132  data: 3.1803  max mem: 3956
+train: [15]  [ 20/400]  eta: 0:03:16  lr: 0.000074  loss: 1.6107 (1.6070)  grad: 0.0478 (0.0495)  time: 0.3716  data: 0.0050  max mem: 3956
+train: [15]  [ 40/400]  eta: 0:02:42  lr: 0.000072  loss: 1.5834 (1.5871)  grad: 0.0503 (0.0500)  time: 0.3834  data: 0.0031  max mem: 3956
+train: [15]  [ 60/400]  eta: 0:02:23  lr: 0.000071  loss: 1.5840 (1.5960)  grad: 0.0503 (0.0497)  time: 0.3618  data: 0.0033  max mem: 3956
+train: [15]  [ 80/400]  eta: 0:02:11  lr: 0.000070  loss: 1.5879 (1.5883)  grad: 0.0470 (0.0486)  time: 0.3764  data: 0.0034  max mem: 3956
+train: [15]  [100/400]  eta: 0:02:01  lr: 0.000068  loss: 1.5822 (1.5902)  grad: 0.0477 (0.0487)  time: 0.3846  data: 0.0034  max mem: 3956
+train: [15]  [120/400]  eta: 0:01:52  lr: 0.000067  loss: 1.5857 (1.5896)  grad: 0.0476 (0.0483)  time: 0.3917  data: 0.0034  max mem: 3956
+train: [15]  [140/400]  eta: 0:01:43  lr: 0.000066  loss: 1.5754 (1.5894)  grad: 0.0466 (0.0483)  time: 0.3700  data: 0.0035  max mem: 3956
+train: [15]  [160/400]  eta: 0:01:35  lr: 0.000064  loss: 1.5741 (1.5893)  grad: 0.0469 (0.0481)  time: 0.3778  data: 0.0034  max mem: 3956
+train: [15]  [180/400]  eta: 0:01:26  lr: 0.000063  loss: 1.5748 (1.5891)  grad: 0.0467 (0.0478)  time: 0.3765  data: 0.0036  max mem: 3956
+train: [15]  [200/400]  eta: 0:01:18  lr: 0.000062  loss: 1.5660 (1.5895)  grad: 0.0457 (0.0479)  time: 0.3905  data: 0.0035  max mem: 3956
+train: [15]  [220/400]  eta: 0:01:10  lr: 0.000061  loss: 1.5839 (1.5896)  grad: 0.0465 (0.0479)  time: 0.3861  data: 0.0035  max mem: 3956
+train: [15]  [240/400]  eta: 0:01:02  lr: 0.000059  loss: 1.5867 (1.5904)  grad: 0.0471 (0.0479)  time: 0.3705  data: 0.0034  max mem: 3956
+train: [15]  [260/400]  eta: 0:00:54  lr: 0.000058  loss: 1.5924 (1.5921)  grad: 0.0489 (0.0481)  time: 0.3681  data: 0.0035  max mem: 3956
+train: [15]  [280/400]  eta: 0:00:46  lr: 0.000057  loss: 1.5979 (1.5921)  grad: 0.0489 (0.0483)  time: 0.3685  data: 0.0035  max mem: 3956
+train: [15]  [300/400]  eta: 0:00:39  lr: 0.000056  loss: 1.5940 (1.5911)  grad: 0.0488 (0.0481)  time: 0.5677  data: 0.2058  max mem: 3956
+train: [15]  [320/400]  eta: 0:00:31  lr: 0.000054  loss: 1.5853 (1.5907)  grad: 0.0428 (0.0479)  time: 0.3702  data: 0.0041  max mem: 3956
+train: [15]  [340/400]  eta: 0:00:23  lr: 0.000053  loss: 1.5883 (1.5924)  grad: 0.0424 (0.0477)  time: 0.3649  data: 0.0040  max mem: 3956
+train: [15]  [360/400]  eta: 0:00:15  lr: 0.000052  loss: 1.5932 (1.5904)  grad: 0.0430 (0.0476)  time: 0.3602  data: 0.0035  max mem: 3956
+train: [15]  [380/400]  eta: 0:00:07  lr: 0.000051  loss: 1.5960 (1.5924)  grad: 0.0483 (0.0478)  time: 0.3611  data: 0.0034  max mem: 3956
+train: [15]  [399/400]  eta: 0:00:00  lr: 0.000050  loss: 1.5960 (1.5910)  grad: 0.0483 (0.0477)  time: 0.3743  data: 0.0035  max mem: 3956
+train: [15] Total time: 0:02:36 (0.3917 s / it)
+train: [15] Summary: lr: 0.000050  loss: 1.5960 (1.5910)  grad: 0.0483 (0.0477)
+eval (validation): [15]  [ 0/63]  eta: 0:03:46    time: 3.5995  data: 3.3537  max mem: 3956
+eval (validation): [15]  [20/63]  eta: 0:00:21    time: 0.3337  data: 0.0038  max mem: 3956
+eval (validation): [15]  [40/63]  eta: 0:00:09    time: 0.3504  data: 0.0034  max mem: 3956
+eval (validation): [15]  [60/63]  eta: 0:00:01    time: 0.3419  data: 0.0036  max mem: 3956
+eval (validation): [15]  [62/63]  eta: 0:00:00    time: 0.3371  data: 0.0035  max mem: 3956
+eval (validation): [15] Total time: 0:00:25 (0.3983 s / it)
+cv: [15]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.347  acc: 0.930  f1: 0.920
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [16]  [  0/400]  eta: 0:23:02  lr: nan  time: 3.4555  data: 3.2237  max mem: 3956
+train: [16]  [ 20/400]  eta: 0:03:29  lr: 0.000048  loss: 1.5776 (1.5634)  grad: 0.0444 (0.0472)  time: 0.4058  data: 0.0119  max mem: 3956
+train: [16]  [ 40/400]  eta: 0:02:47  lr: 0.000047  loss: 1.5776 (1.5745)  grad: 0.0465 (0.0467)  time: 0.3754  data: 0.0028  max mem: 3956
+train: [16]  [ 60/400]  eta: 0:02:27  lr: 0.000046  loss: 1.5747 (1.5778)  grad: 0.0484 (0.0482)  time: 0.3648  data: 0.0035  max mem: 3956
+train: [16]  [ 80/400]  eta: 0:02:12  lr: 0.000045  loss: 1.5578 (1.5697)  grad: 0.0454 (0.0467)  time: 0.3549  data: 0.0033  max mem: 3956
+train: [16]  [100/400]  eta: 0:02:00  lr: 0.000044  loss: 1.5577 (1.5750)  grad: 0.0442 (0.0473)  time: 0.3604  data: 0.0034  max mem: 3956
+train: [16]  [120/400]  eta: 0:01:51  lr: 0.000043  loss: 1.5774 (1.5790)  grad: 0.0459 (0.0472)  time: 0.3854  data: 0.0036  max mem: 3956
+train: [16]  [140/400]  eta: 0:01:42  lr: 0.000042  loss: 1.5681 (1.5755)  grad: 0.0461 (0.0472)  time: 0.3661  data: 0.0034  max mem: 3956
+train: [16]  [160/400]  eta: 0:01:34  lr: 0.000041  loss: 1.5579 (1.5752)  grad: 0.0494 (0.0475)  time: 0.3715  data: 0.0035  max mem: 3956
+train: [16]  [180/400]  eta: 0:01:25  lr: 0.000040  loss: 1.5946 (1.5804)  grad: 0.0479 (0.0474)  time: 0.3657  data: 0.0033  max mem: 3956
+train: [16]  [200/400]  eta: 0:01:17  lr: 0.000039  loss: 1.5762 (1.5786)  grad: 0.0428 (0.0471)  time: 0.3826  data: 0.0034  max mem: 3956
+train: [16]  [220/400]  eta: 0:01:09  lr: 0.000038  loss: 1.5551 (1.5772)  grad: 0.0457 (0.0476)  time: 0.3659  data: 0.0034  max mem: 3956
+train: [16]  [240/400]  eta: 0:01:02  lr: 0.000036  loss: 1.5832 (1.5794)  grad: 0.0501 (0.0477)  time: 0.4121  data: 0.0036  max mem: 3956
+train: [16]  [260/400]  eta: 0:00:54  lr: 0.000035  loss: 1.5942 (1.5798)  grad: 0.0456 (0.0475)  time: 0.3969  data: 0.0034  max mem: 3956
+train: [16]  [280/400]  eta: 0:00:46  lr: 0.000034  loss: 1.6071 (1.5822)  grad: 0.0450 (0.0475)  time: 0.3598  data: 0.0033  max mem: 3956
+train: [16]  [300/400]  eta: 0:00:39  lr: 0.000033  loss: 1.6089 (1.5846)  grad: 0.0474 (0.0476)  time: 0.5396  data: 0.1931  max mem: 3956
+train: [16]  [320/400]  eta: 0:00:31  lr: 0.000032  loss: 1.5793 (1.5842)  grad: 0.0465 (0.0474)  time: 0.3681  data: 0.0034  max mem: 3956
+train: [16]  [340/400]  eta: 0:00:23  lr: 0.000031  loss: 1.5702 (1.5830)  grad: 0.0449 (0.0475)  time: 0.3700  data: 0.0032  max mem: 3956
+train: [16]  [360/400]  eta: 0:00:15  lr: 0.000031  loss: 1.5691 (1.5830)  grad: 0.0449 (0.0473)  time: 0.3597  data: 0.0033  max mem: 3956
+train: [16]  [380/400]  eta: 0:00:07  lr: 0.000030  loss: 1.5560 (1.5816)  grad: 0.0454 (0.0473)  time: 0.3746  data: 0.0033  max mem: 3956
+train: [16]  [399/400]  eta: 0:00:00  lr: 0.000029  loss: 1.5578 (1.5807)  grad: 0.0464 (0.0473)  time: 0.3810  data: 0.0035  max mem: 3956
+train: [16] Total time: 0:02:36 (0.3910 s / it)
+train: [16] Summary: lr: 0.000029  loss: 1.5578 (1.5807)  grad: 0.0464 (0.0473)
+eval (validation): [16]  [ 0/63]  eta: 0:03:50    time: 3.6586  data: 3.4113  max mem: 3956
+eval (validation): [16]  [20/63]  eta: 0:00:21    time: 0.3487  data: 0.0127  max mem: 3956
+eval (validation): [16]  [40/63]  eta: 0:00:09    time: 0.3405  data: 0.0035  max mem: 3956
+eval (validation): [16]  [60/63]  eta: 0:00:01    time: 0.3374  data: 0.0026  max mem: 3956
+eval (validation): [16]  [62/63]  eta: 0:00:00    time: 0.3345  data: 0.0029  max mem: 3956
+eval (validation): [16] Total time: 0:00:25 (0.3995 s / it)
+cv: [16]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.343  acc: 0.930  f1: 0.921
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [17]  [  0/400]  eta: 0:22:20  lr: nan  time: 3.3517  data: 3.0691  max mem: 3956
+train: [17]  [ 20/400]  eta: 0:03:23  lr: 0.000028  loss: 1.5988 (1.5997)  grad: 0.0505 (0.0531)  time: 0.3950  data: 0.0034  max mem: 3956
+train: [17]  [ 40/400]  eta: 0:02:42  lr: 0.000027  loss: 1.5672 (1.5672)  grad: 0.0493 (0.0502)  time: 0.3608  data: 0.0031  max mem: 3956
+train: [17]  [ 60/400]  eta: 0:02:22  lr: 0.000026  loss: 1.5523 (1.5639)  grad: 0.0472 (0.0497)  time: 0.3552  data: 0.0035  max mem: 3956
+train: [17]  [ 80/400]  eta: 0:02:09  lr: 0.000025  loss: 1.5572 (1.5706)  grad: 0.0487 (0.0493)  time: 0.3563  data: 0.0032  max mem: 3956
+train: [17]  [100/400]  eta: 0:02:00  lr: 0.000024  loss: 1.5980 (1.5785)  grad: 0.0488 (0.0492)  time: 0.3873  data: 0.0032  max mem: 3956
+train: [17]  [120/400]  eta: 0:01:50  lr: 0.000023  loss: 1.5830 (1.5749)  grad: 0.0483 (0.0493)  time: 0.3628  data: 0.0035  max mem: 3956
+train: [17]  [140/400]  eta: 0:01:41  lr: 0.000023  loss: 1.5830 (1.5807)  grad: 0.0481 (0.0493)  time: 0.3556  data: 0.0032  max mem: 3956
+train: [17]  [160/400]  eta: 0:01:32  lr: 0.000022  loss: 1.6127 (1.5843)  grad: 0.0511 (0.0496)  time: 0.3699  data: 0.0032  max mem: 3956
+train: [17]  [180/400]  eta: 0:01:24  lr: 0.000021  loss: 1.6086 (1.5880)  grad: 0.0484 (0.0490)  time: 0.3727  data: 0.0033  max mem: 3956
+train: [17]  [200/400]  eta: 0:01:16  lr: 0.000020  loss: 1.5728 (1.5844)  grad: 0.0450 (0.0486)  time: 0.3604  data: 0.0034  max mem: 3956
+train: [17]  [220/400]  eta: 0:01:08  lr: 0.000019  loss: 1.5547 (1.5809)  grad: 0.0473 (0.0487)  time: 0.3591  data: 0.0035  max mem: 3956
+train: [17]  [240/400]  eta: 0:01:00  lr: 0.000019  loss: 1.5607 (1.5835)  grad: 0.0512 (0.0487)  time: 0.3751  data: 0.0035  max mem: 3956
+train: [17]  [260/400]  eta: 0:00:53  lr: 0.000018  loss: 1.6051 (1.5839)  grad: 0.0498 (0.0490)  time: 0.3821  data: 0.0035  max mem: 3956
+train: [17]  [280/400]  eta: 0:00:45  lr: 0.000017  loss: 1.5834 (1.5842)  grad: 0.0474 (0.0487)  time: 0.3479  data: 0.0032  max mem: 3956
+train: [17]  [300/400]  eta: 0:00:38  lr: 0.000016  loss: 1.5847 (1.5850)  grad: 0.0422 (0.0483)  time: 0.5534  data: 0.2017  max mem: 3956
+train: [17]  [320/400]  eta: 0:00:31  lr: 0.000016  loss: 1.5678 (1.5845)  grad: 0.0449 (0.0482)  time: 0.3645  data: 0.0039  max mem: 3956
+train: [17]  [340/400]  eta: 0:00:23  lr: 0.000015  loss: 1.5647 (1.5821)  grad: 0.0460 (0.0480)  time: 0.3586  data: 0.0027  max mem: 3956
+train: [17]  [360/400]  eta: 0:00:15  lr: 0.000014  loss: 1.5778 (1.5823)  grad: 0.0452 (0.0479)  time: 0.3502  data: 0.0035  max mem: 3956
+train: [17]  [380/400]  eta: 0:00:07  lr: 0.000014  loss: 1.5778 (1.5826)  grad: 0.0453 (0.0478)  time: 0.3659  data: 0.0035  max mem: 3956
+train: [17]  [399/400]  eta: 0:00:00  lr: 0.000013  loss: 1.5497 (1.5805)  grad: 0.0455 (0.0477)  time: 0.3672  data: 0.0033  max mem: 3956
+train: [17] Total time: 0:02:33 (0.3827 s / it)
+train: [17] Summary: lr: 0.000013  loss: 1.5497 (1.5805)  grad: 0.0455 (0.0477)
+eval (validation): [17]  [ 0/63]  eta: 0:03:41    time: 3.5150  data: 3.2879  max mem: 3956
+eval (validation): [17]  [20/63]  eta: 0:00:23    time: 0.3879  data: 0.0043  max mem: 3956
+eval (validation): [17]  [40/63]  eta: 0:00:10    time: 0.3416  data: 0.0032  max mem: 3956
+eval (validation): [17]  [60/63]  eta: 0:00:01    time: 0.3392  data: 0.0033  max mem: 3956
+eval (validation): [17]  [62/63]  eta: 0:00:00    time: 0.3408  data: 0.0033  max mem: 3956
+eval (validation): [17] Total time: 0:00:25 (0.4103 s / it)
+cv: [17]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.340  acc: 0.932  f1: 0.924
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [18]  [  0/400]  eta: 0:22:19  lr: nan  time: 3.3483  data: 3.0924  max mem: 3956
+train: [18]  [ 20/400]  eta: 0:03:20  lr: 0.000012  loss: 1.5210 (1.5343)  grad: 0.0464 (0.0455)  time: 0.3854  data: 0.0038  max mem: 3956
+train: [18]  [ 40/400]  eta: 0:02:42  lr: 0.000012  loss: 1.5677 (1.5492)  grad: 0.0468 (0.0464)  time: 0.3741  data: 0.0030  max mem: 3956
+train: [18]  [ 60/400]  eta: 0:02:24  lr: 0.000011  loss: 1.5753 (1.5559)  grad: 0.0455 (0.0460)  time: 0.3666  data: 0.0032  max mem: 3956
+train: [18]  [ 80/400]  eta: 0:02:10  lr: 0.000011  loss: 1.5995 (1.5698)  grad: 0.0463 (0.0470)  time: 0.3527  data: 0.0034  max mem: 3956
+train: [18]  [100/400]  eta: 0:01:59  lr: 0.000010  loss: 1.5821 (1.5736)  grad: 0.0469 (0.0473)  time: 0.3638  data: 0.0034  max mem: 3956
+train: [18]  [120/400]  eta: 0:01:51  lr: 0.000009  loss: 1.5657 (1.5720)  grad: 0.0470 (0.0477)  time: 0.3892  data: 0.0038  max mem: 3956
+train: [18]  [140/400]  eta: 0:01:42  lr: 0.000009  loss: 1.5610 (1.5709)  grad: 0.0477 (0.0479)  time: 0.3780  data: 0.0034  max mem: 3956
+train: [18]  [160/400]  eta: 0:01:33  lr: 0.000008  loss: 1.5803 (1.5763)  grad: 0.0470 (0.0477)  time: 0.3739  data: 0.0033  max mem: 3956
+train: [18]  [180/400]  eta: 0:01:25  lr: 0.000008  loss: 1.5964 (1.5777)  grad: 0.0459 (0.0479)  time: 0.3789  data: 0.0034  max mem: 3956
+train: [18]  [200/400]  eta: 0:01:17  lr: 0.000007  loss: 1.6002 (1.5800)  grad: 0.0456 (0.0476)  time: 0.3714  data: 0.0039  max mem: 3956
+train: [18]  [220/400]  eta: 0:01:09  lr: 0.000007  loss: 1.6049 (1.5801)  grad: 0.0438 (0.0478)  time: 0.3852  data: 0.0030  max mem: 3956
+train: [18]  [240/400]  eta: 0:01:02  lr: 0.000006  loss: 1.5985 (1.5828)  grad: 0.0457 (0.0476)  time: 0.4050  data: 0.0036  max mem: 3956
+train: [18]  [260/400]  eta: 0:00:54  lr: 0.000006  loss: 1.5985 (1.5818)  grad: 0.0450 (0.0476)  time: 0.3942  data: 0.0035  max mem: 3956
+train: [18]  [280/400]  eta: 0:00:46  lr: 0.000006  loss: 1.5707 (1.5809)  grad: 0.0448 (0.0475)  time: 0.3578  data: 0.0039  max mem: 3956
+train: [18]  [300/400]  eta: 0:00:40  lr: 0.000005  loss: 1.5707 (1.5807)  grad: 0.0478 (0.0476)  time: 0.5794  data: 0.2080  max mem: 3956
+train: [18]  [320/400]  eta: 0:00:32  lr: 0.000005  loss: 1.5474 (1.5796)  grad: 0.0478 (0.0476)  time: 0.4044  data: 0.0034  max mem: 3956
+train: [18]  [340/400]  eta: 0:00:23  lr: 0.000004  loss: 1.5668 (1.5806)  grad: 0.0443 (0.0475)  time: 0.3793  data: 0.0034  max mem: 3956
+train: [18]  [360/400]  eta: 0:00:15  lr: 0.000004  loss: 1.5819 (1.5822)  grad: 0.0489 (0.0477)  time: 0.3543  data: 0.0034  max mem: 3956
+train: [18]  [380/400]  eta: 0:00:07  lr: 0.000004  loss: 1.5970 (1.5827)  grad: 0.0489 (0.0477)  time: 0.3706  data: 0.0033  max mem: 3956
+train: [18]  [399/400]  eta: 0:00:00  lr: 0.000003  loss: 1.5852 (1.5823)  grad: 0.0457 (0.0477)  time: 0.3774  data: 0.0034  max mem: 3956
+train: [18] Total time: 0:02:37 (0.3947 s / it)
+train: [18] Summary: lr: 0.000003  loss: 1.5852 (1.5823)  grad: 0.0457 (0.0477)
+eval (validation): [18]  [ 0/63]  eta: 0:03:45    time: 3.5733  data: 3.2718  max mem: 3956
+eval (validation): [18]  [20/63]  eta: 0:00:21    time: 0.3575  data: 0.0046  max mem: 3956
+eval (validation): [18]  [40/63]  eta: 0:00:10    time: 0.3567  data: 0.0030  max mem: 3956
+eval (validation): [18]  [60/63]  eta: 0:00:01    time: 0.3292  data: 0.0032  max mem: 3956
+eval (validation): [18]  [62/63]  eta: 0:00:00    time: 0.3291  data: 0.0032  max mem: 3956
+eval (validation): [18] Total time: 0:00:25 (0.4030 s / it)
+cv: [18]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.339  acc: 0.932  f1: 0.925
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+train: [19]  [  0/400]  eta: 0:22:43  lr: nan  time: 3.4097  data: 3.1715  max mem: 3956
+train: [19]  [ 20/400]  eta: 0:03:19  lr: 0.000003  loss: 1.5982 (1.6074)  grad: 0.0457 (0.0477)  time: 0.3814  data: 0.0033  max mem: 3956
+train: [19]  [ 40/400]  eta: 0:02:40  lr: 0.000003  loss: 1.5971 (1.5930)  grad: 0.0492 (0.0494)  time: 0.3641  data: 0.0032  max mem: 3956
+train: [19]  [ 60/400]  eta: 0:02:21  lr: 0.000002  loss: 1.5581 (1.5768)  grad: 0.0458 (0.0484)  time: 0.3564  data: 0.0036  max mem: 3956
+train: [19]  [ 80/400]  eta: 0:02:08  lr: 0.000002  loss: 1.5604 (1.5776)  grad: 0.0451 (0.0481)  time: 0.3541  data: 0.0032  max mem: 3956
+train: [19]  [100/400]  eta: 0:01:58  lr: 0.000002  loss: 1.5809 (1.5747)  grad: 0.0493 (0.0484)  time: 0.3673  data: 0.0033  max mem: 3956
+train: [19]  [120/400]  eta: 0:01:48  lr: 0.000002  loss: 1.5643 (1.5763)  grad: 0.0493 (0.0484)  time: 0.3580  data: 0.0038  max mem: 3956
+train: [19]  [140/400]  eta: 0:01:39  lr: 0.000001  loss: 1.5835 (1.5771)  grad: 0.0478 (0.0483)  time: 0.3534  data: 0.0035  max mem: 3956
+train: [19]  [160/400]  eta: 0:01:31  lr: 0.000001  loss: 1.5852 (1.5754)  grad: 0.0460 (0.0479)  time: 0.3560  data: 0.0035  max mem: 3956
+train: [19]  [180/400]  eta: 0:01:23  lr: 0.000001  loss: 1.5303 (1.5702)  grad: 0.0450 (0.0477)  time: 0.3648  data: 0.0036  max mem: 3956
+train: [19]  [200/400]  eta: 0:01:15  lr: 0.000001  loss: 1.5680 (1.5723)  grad: 0.0459 (0.0479)  time: 0.3533  data: 0.0032  max mem: 3956
+train: [19]  [220/400]  eta: 0:01:07  lr: 0.000001  loss: 1.5878 (1.5730)  grad: 0.0482 (0.0479)  time: 0.3761  data: 0.0034  max mem: 3956
+train: [19]  [240/400]  eta: 0:01:00  lr: 0.000001  loss: 1.5857 (1.5744)  grad: 0.0459 (0.0477)  time: 0.3732  data: 0.0035  max mem: 3956
+train: [19]  [260/400]  eta: 0:00:52  lr: 0.000000  loss: 1.5857 (1.5756)  grad: 0.0433 (0.0473)  time: 0.3649  data: 0.0033  max mem: 3956
+train: [19]  [280/400]  eta: 0:00:44  lr: 0.000000  loss: 1.5896 (1.5773)  grad: 0.0415 (0.0472)  time: 0.3550  data: 0.0035  max mem: 3956
+train: [19]  [300/400]  eta: 0:00:38  lr: 0.000000  loss: 1.5896 (1.5789)  grad: 0.0464 (0.0472)  time: 0.5553  data: 0.2111  max mem: 3956
+train: [19]  [320/400]  eta: 0:00:30  lr: 0.000000  loss: 1.5978 (1.5797)  grad: 0.0467 (0.0473)  time: 0.4040  data: 0.0039  max mem: 3956
+train: [19]  [340/400]  eta: 0:00:23  lr: 0.000000  loss: 1.5669 (1.5774)  grad: 0.0450 (0.0471)  time: 0.3665  data: 0.0034  max mem: 3956
+train: [19]  [360/400]  eta: 0:00:15  lr: 0.000000  loss: 1.5320 (1.5768)  grad: 0.0453 (0.0472)  time: 0.3529  data: 0.0035  max mem: 3956
+train: [19]  [380/400]  eta: 0:00:07  lr: 0.000000  loss: 1.5473 (1.5772)  grad: 0.0474 (0.0471)  time: 0.3590  data: 0.0035  max mem: 3956
+train: [19]  [399/400]  eta: 0:00:00  lr: 0.000000  loss: 1.5519 (1.5741)  grad: 0.0506 (0.0474)  time: 0.3603  data: 0.0035  max mem: 3956
+train: [19] Total time: 0:02:32 (0.3817 s / it)
+train: [19] Summary: lr: 0.000000  loss: 1.5519 (1.5741)  grad: 0.0506 (0.0474)
+eval (validation): [19]  [ 0/63]  eta: 0:03:43    time: 3.5531  data: 3.2582  max mem: 3956
+eval (validation): [19]  [20/63]  eta: 0:00:21    time: 0.3532  data: 0.0040  max mem: 3956
+eval (validation): [19]  [40/63]  eta: 0:00:10    time: 0.3631  data: 0.0032  max mem: 3956
+eval (validation): [19]  [60/63]  eta: 0:00:01    time: 0.3259  data: 0.0032  max mem: 3956
+eval (validation): [19]  [62/63]  eta: 0:00:00    time: 0.3234  data: 0.0031  max mem: 3956
+eval (validation): [19] Total time: 0:00:25 (0.4027 s / it)
+cv: [19]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.339  acc: 0.932  f1: 0.924
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+evaluating last checkpoint: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-last.pth
+eval model info:
+{"score": 0.9317956349206349, "hparam": [50, 1.0], "hparam_id": 48, "epoch": 19, "is_best": false, "best_score": 0.9320436507936508}
+eval (train): [20]  [  0/297]  eta: 0:16:47    time: 3.3919  data: 3.1609  max mem: 3956
+eval (train): [20]  [ 20/297]  eta: 0:02:20    time: 0.3643  data: 0.0132  max mem: 3956
+eval (train): [20]  [ 40/297]  eta: 0:01:50    time: 0.3509  data: 0.0035  max mem: 3956
+eval (train): [20]  [ 60/297]  eta: 0:01:34    time: 0.3377  data: 0.0030  max mem: 3956
+eval (train): [20]  [ 80/297]  eta: 0:01:23    time: 0.3417  data: 0.0034  max mem: 3956
+eval (train): [20]  [100/297]  eta: 0:01:15    time: 0.3691  data: 0.0036  max mem: 3956
+eval (train): [20]  [120/297]  eta: 0:01:07    time: 0.3775  data: 0.0039  max mem: 3956
+eval (train): [20]  [140/297]  eta: 0:00:59    time: 0.3585  data: 0.0034  max mem: 3956
+eval (train): [20]  [160/297]  eta: 0:00:51    time: 0.3526  data: 0.0035  max mem: 3956
+eval (train): [20]  [180/297]  eta: 0:00:43    time: 0.3536  data: 0.0034  max mem: 3956
+eval (train): [20]  [200/297]  eta: 0:00:35    time: 0.3495  data: 0.0034  max mem: 3956
+eval (train): [20]  [220/297]  eta: 0:00:28    time: 0.3538  data: 0.0035  max mem: 3956
+eval (train): [20]  [240/297]  eta: 0:00:20    time: 0.3522  data: 0.0036  max mem: 3956
+eval (train): [20]  [260/297]  eta: 0:00:13    time: 0.3503  data: 0.0035  max mem: 3956
+eval (train): [20]  [280/297]  eta: 0:00:06    time: 0.3524  data: 0.0033  max mem: 3956
+eval (train): [20]  [296/297]  eta: 0:00:00    time: 0.3295  data: 0.0032  max mem: 3956
+eval (train): [20] Total time: 0:01:48 (0.3655 s / it)
+eval (validation): [20]  [ 0/63]  eta: 0:03:39    time: 3.4893  data: 3.2487  max mem: 3956
+eval (validation): [20]  [20/63]  eta: 0:00:22    time: 0.3805  data: 0.0279  max mem: 3956
+eval (validation): [20]  [40/63]  eta: 0:00:10    time: 0.3716  data: 0.0033  max mem: 3956
+eval (validation): [20]  [60/63]  eta: 0:00:01    time: 0.3274  data: 0.0026  max mem: 3956
+eval (validation): [20]  [62/63]  eta: 0:00:00    time: 0.3279  data: 0.0029  max mem: 3956
+eval (validation): [20] Total time: 0:00:26 (0.4129 s / it)
+eval (test): [20]  [ 0/79]  eta: 0:04:36    time: 3.4943  data: 3.1974  max mem: 3956
+eval (test): [20]  [20/79]  eta: 0:00:31    time: 0.3945  data: 0.0175  max mem: 3956
+eval (test): [20]  [40/79]  eta: 0:00:17    time: 0.3508  data: 0.0031  max mem: 3956
+eval (test): [20]  [60/79]  eta: 0:00:07    time: 0.3440  data: 0.0037  max mem: 3956
+eval (test): [20]  [78/79]  eta: 0:00:00    time: 0.3385  data: 0.0033  max mem: 3956
+eval (test): [20] Total time: 0:00:31 (0.4028 s / it)
+evaluating best checkpoint: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__patch__linear/checkpoint-best.pth
+eval model info:
+{"score": 0.9320436507936508, "hparam": [50, 1.0], "hparam_id": 48, "epoch": 18, "is_best": true, "best_score": 0.9320436507936508}
+eval (train): [20]  [  0/297]  eta: 0:17:54    time: 3.6173  data: 3.3291  max mem: 3956
+eval (train): [20]  [ 20/297]  eta: 0:02:21    time: 0.3541  data: 0.0046  max mem: 3956
+eval (train): [20]  [ 40/297]  eta: 0:01:50    time: 0.3489  data: 0.0034  max mem: 3956
+eval (train): [20]  [ 60/297]  eta: 0:01:34    time: 0.3339  data: 0.0033  max mem: 3956
+eval (train): [20]  [ 80/297]  eta: 0:01:24    time: 0.3610  data: 0.0034  max mem: 3956
+eval (train): [20]  [100/297]  eta: 0:01:14    time: 0.3259  data: 0.0035  max mem: 3956
+eval (train): [20]  [120/297]  eta: 0:01:05    time: 0.3425  data: 0.0035  max mem: 3956
+eval (train): [20]  [140/297]  eta: 0:00:57    time: 0.3381  data: 0.0035  max mem: 3956
+eval (train): [20]  [160/297]  eta: 0:00:50    time: 0.3569  data: 0.0035  max mem: 3956
+eval (train): [20]  [180/297]  eta: 0:00:42    time: 0.3574  data: 0.0038  max mem: 3956
+eval (train): [20]  [200/297]  eta: 0:00:35    time: 0.3361  data: 0.0034  max mem: 3956
+eval (train): [20]  [220/297]  eta: 0:00:27    time: 0.3627  data: 0.0037  max mem: 3956
+eval (train): [20]  [240/297]  eta: 0:00:20    time: 0.3430  data: 0.0035  max mem: 3956
+eval (train): [20]  [260/297]  eta: 0:00:13    time: 0.3573  data: 0.0036  max mem: 3956
+eval (train): [20]  [280/297]  eta: 0:00:06    time: 0.3440  data: 0.0036  max mem: 3956
+eval (train): [20]  [296/297]  eta: 0:00:00    time: 0.3297  data: 0.0033  max mem: 3956
+eval (train): [20] Total time: 0:01:46 (0.3598 s / it)
+eval (validation): [20]  [ 0/63]  eta: 0:03:53    time: 3.7013  data: 3.4093  max mem: 3956
+eval (validation): [20]  [20/63]  eta: 0:00:21    time: 0.3465  data: 0.0104  max mem: 3956
+eval (validation): [20]  [40/63]  eta: 0:00:09    time: 0.3350  data: 0.0036  max mem: 3956
+eval (validation): [20]  [60/63]  eta: 0:00:01    time: 0.3230  data: 0.0035  max mem: 3956
+eval (validation): [20]  [62/63]  eta: 0:00:00    time: 0.3129  data: 0.0028  max mem: 3956
+eval (validation): [20] Total time: 0:00:24 (0.3925 s / it)
+eval (test): [20]  [ 0/79]  eta: 0:04:51    time: 3.6948  data: 3.4539  max mem: 3956
+eval (test): [20]  [20/79]  eta: 0:00:30    time: 0.3615  data: 0.0033  max mem: 3956
+eval (test): [20]  [40/79]  eta: 0:00:16    time: 0.3327  data: 0.0030  max mem: 3956
+eval (test): [20]  [60/79]  eta: 0:00:07    time: 0.3567  data: 0.0036  max mem: 3956
+eval (test): [20]  [78/79]  eta: 0:00:00    time: 0.3240  data: 0.0032  max mem: 3956
+eval (test): [20] Total time: 0:00:30 (0.3914 s / it)
+eval results:
+
+| model    | repr   | clf    | dataset      | ckpt   |   epoch |    lr |   wd |   hparam_id | hparam    | split      |    loss |     acc |   acc_std |      f1 |    f1_std |
+|:---------|:-------|:-------|:-------------|:-------|--------:|------:|-----:|------------:|:----------|:-----------|--------:|--------:|----------:|--------:|----------:|
+| flat_mae | patch  | linear | hcpya_task21 | best   |      18 | 0.015 | 0.05 |          48 | [50, 1.0] | train      | 0.30549 | 0.94579 | 0.0015117 | 0.94245 | 0.0018296 |
+| flat_mae | patch  | linear | hcpya_task21 | best   |      18 | 0.015 | 0.05 |          48 | [50, 1.0] | validation | 0.33905 | 0.93204 | 0.0038816 | 0.92518 | 0.0047207 |
+| flat_mae | patch  | linear | hcpya_task21 | best   |      18 | 0.015 | 0.05 |          48 | [50, 1.0] | test       | 0.34095 | 0.93274 | 0.0032783 | 0.92501 | 0.0040806 |
+
+
+done! total time: 1:07:00

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/config.yaml

@@ -0,0 +1,96 @@
+output_root: experiments/decoders/output
+name_prefix: eval_probe
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (hcpya_task21 reg attn)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: false
+  norm: false
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn
+model: flat_mae
+representation: reg
+classifier: attn
+dataset: hcpya_task21
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn
+remote_dir: null

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/eval_log.json

@@ -0,0 +1 @@
+{"eval/epoch": 8, "eval/id_best": 23, "eval/lr_best": 0.00025499999999999996, "eval/wd_best": 0.05, "eval/train/loss": 0.035422105342149734, "eval/train/acc": 0.9938417811463761, "eval/train/acc_std": 0.0005515012840441916, "eval/train/f1": 0.9936347787478754, "eval/train/f1_std": 0.00062564819864731, "eval/validation/loss": 0.06714221090078354, "eval/validation/acc": 0.9818948412698413, "eval/validation/acc_std": 0.001929209224017128, "eval/validation/f1": 0.9799099655473743, "eval/validation/f1_std": 0.0024497399729675874, "eval/test/loss": 0.08606162667274475, "eval/test/acc": 0.9742063492063492, "eval/test/acc_std": 0.0022554523674217113, "eval/test/f1": 0.968385237238162, "eval/test/f1_std": 0.0030263738926539128}

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/eval_log_best.json

@@ -0,0 +1 @@
+{"eval/best/epoch": 8, "eval/best/id_best": 23, "eval/best/lr_best": 0.00025499999999999996, "eval/best/wd_best": 0.05, "eval/best/train/loss": 0.035422105342149734, "eval/best/train/acc": 0.9938417811463761, "eval/best/train/acc_std": 0.0005515012840441916, "eval/best/train/f1": 0.9936347787478754, "eval/best/train/f1_std": 0.00062564819864731, "eval/best/validation/loss": 0.06714221090078354, "eval/best/validation/acc": 0.9818948412698413, "eval/best/validation/acc_std": 0.001929209224017128, "eval/best/validation/f1": 0.9799099655473743, "eval/best/validation/f1_std": 0.0024497399729675874, "eval/best/test/loss": 0.08606162667274475, "eval/best/test/acc": 0.9742063492063492, "eval/best/test/acc_std": 0.0022554523674217113, "eval/best/test/f1": 0.968385237238162, "eval/best/test/f1_std": 0.0030263738926539128}

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/eval_log_last.json

@@ -0,0 +1 @@
+{"eval/last/epoch": 19, "eval/last/id_best": 23, "eval/last/lr_best": 0.00025499999999999996, "eval/last/wd_best": 0.05, "eval/last/train/loss": 0.01573818549513817, "eval/last/train/acc": 0.9980525290804779, "eval/last/train/acc_std": 0.0003204039388166683, "eval/last/train/f1": 0.9981245127681868, "eval/last/train/f1_std": 0.0003248137248363995, "eval/last/validation/loss": 0.06250125169754028, "eval/last/validation/acc": 0.9809027777777778, "eval/last/validation/acc_std": 0.0020652798650272228, "eval/last/validation/f1": 0.9783232753219386, "eval/last/validation/f1_std": 0.002647813729125984, "eval/last/test/loss": 0.08106201887130737, "eval/last/test/acc": 0.9759920634920635, "eval/last/test/acc_std": 0.0021589489851741514, "eval/last/test/f1": 0.970003878850155, "eval/last/test/f1_std": 0.0029274843494945402}

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/eval_table.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,reg,attn,hcpya_task21,best,8,0.00025499999999999996,0.05,23,"[0.85, 1.0]",train,0.035422105342149734,0.9938417811463761,0.0005515012840441916,0.9936347787478754,0.00062564819864731
+flat_mae,reg,attn,hcpya_task21,best,8,0.00025499999999999996,0.05,23,"[0.85, 1.0]",validation,0.06714221090078354,0.9818948412698413,0.001929209224017128,0.9799099655473743,0.0024497399729675874
+flat_mae,reg,attn,hcpya_task21,best,8,0.00025499999999999996,0.05,23,"[0.85, 1.0]",test,0.08606162667274475,0.9742063492063492,0.0022554523674217113,0.968385237238162,0.0030263738926539128

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/eval_table_best.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,reg,attn,hcpya_task21,best,8,0.00025499999999999996,0.05,23,"[0.85, 1.0]",train,0.035422105342149734,0.9938417811463761,0.0005515012840441916,0.9936347787478754,0.00062564819864731
+flat_mae,reg,attn,hcpya_task21,best,8,0.00025499999999999996,0.05,23,"[0.85, 1.0]",validation,0.06714221090078354,0.9818948412698413,0.001929209224017128,0.9799099655473743,0.0024497399729675874
+flat_mae,reg,attn,hcpya_task21,best,8,0.00025499999999999996,0.05,23,"[0.85, 1.0]",test,0.08606162667274475,0.9742063492063492,0.0022554523674217113,0.968385237238162,0.0030263738926539128

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/eval_table_last.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,reg,attn,hcpya_task21,last,19,0.00025499999999999996,0.05,23,"[0.85, 1.0]",train,0.01573818549513817,0.9980525290804779,0.0003204039388166683,0.9981245127681868,0.0003248137248363995
+flat_mae,reg,attn,hcpya_task21,last,19,0.00025499999999999996,0.05,23,"[0.85, 1.0]",validation,0.06250125169754028,0.9809027777777778,0.0020652798650272228,0.9783232753219386,0.002647813729125984
+flat_mae,reg,attn,hcpya_task21,last,19,0.00025499999999999996,0.05,23,"[0.85, 1.0]",test,0.08106201887130737,0.9759920634920635,0.0021589489851741514,0.970003878850155,0.0029274843494945402

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/log.txt

@@ -0,0 +1,886 @@
+fMRI foundation model probe eval
+version: 0.1.dev65+g4003a1397
+sha: 6c01b606db98add5848cecd23e5d599250c0bf86, status: has uncommitted changes, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-24 23:41:57
+config:
+output_root: experiments/decoders/output
+name_prefix: eval_probe
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (hcpya_task21 reg attn)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: false
+  norm: false
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn
+model: flat_mae
+representation: reg
+classifier: attn
+dataset: hcpya_task21
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=False, reg_tokens=4, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: hcpya_task21 (flat)
+train (n=18999):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 18999
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[ 832 1248 3201 1660  832  832  832  832  832 1248 1247 1243  832  416
+  416  416  416  416  416  416  416]
+)
+
+validation (n=4032):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 4032
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[176 264 688 352 176 176 176 176 176 264 264 264 176  88  88  88  88  88
+  88  88  88]
+)
+
+test (n=5040):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 5040
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[220 330 860 440 220 220 220 220 220 330 330 330 220 110 110 110 110 110
+ 110 110 110]
+)
+
+running backbone on example batch to get embedding dim
+embedding feature dim (reg): 768
+initializing sweep of classifier heads
+classifiers:
+ModuleList(
+  (0-48): 49 x AttnPoolClassifier(
+    (kv): Linear(in_features=768, out_features=1536, bias=True)
+    (linear): Linear(in_features=768, out_features=21, bias=True)
+  )
+)
+classifier params (train): 58.7M (58.7M)
+setting up optimizer
+total batch size: 128 = 64 bs per gpu x 2 accum
+lr: 3.00e-04
+full schedule: epochs = 20 (steps = 4000) (decay = True)
+warmup: epochs = 5 (steps = 1000)
+start training for 20 epochs
+train: [0]  [  0/400]  eta: 0:22:58  lr: nan  time: 3.4469  data: 3.0362  max mem: 4135
+train: [0]  [ 20/400]  eta: 0:03:22  lr: 0.000003  loss: 3.0347 (3.0349)  grad: 0.0859 (0.0857)  time: 0.3863  data: 0.0031  max mem: 4840
+train: [0]  [ 40/400]  eta: 0:02:38  lr: 0.000006  loss: 3.0255 (3.0204)  grad: 0.0861 (0.0882)  time: 0.3415  data: 0.0037  max mem: 4840
+train: [0]  [ 60/400]  eta: 0:02:18  lr: 0.000009  loss: 2.9749 (2.9961)  grad: 0.0864 (0.0882)  time: 0.3382  data: 0.0041  max mem: 4840
+train: [0]  [ 80/400]  eta: 0:02:04  lr: 0.000012  loss: 2.9128 (2.9637)  grad: 0.0868 (0.0887)  time: 0.3435  data: 0.0039  max mem: 4840
+train: [0]  [100/400]  eta: 0:01:54  lr: 0.000015  loss: 2.8187 (2.9260)  grad: 0.0868 (0.0882)  time: 0.3442  data: 0.0040  max mem: 4840
+train: [0]  [120/400]  eta: 0:01:45  lr: 0.000018  loss: 2.7005 (2.8807)  grad: 0.0872 (0.0889)  time: 0.3489  data: 0.0038  max mem: 4840
+train: [0]  [140/400]  eta: 0:01:36  lr: 0.000021  loss: 2.5915 (2.8327)  grad: 0.0837 (0.0880)  time: 0.3407  data: 0.0042  max mem: 4840
+train: [0]  [160/400]  eta: 0:01:28  lr: 0.000024  loss: 2.4918 (2.7833)  grad: 0.0803 (0.0871)  time: 0.3539  data: 0.0041  max mem: 4840
+train: [0]  [180/400]  eta: 0:01:21  lr: 0.000027  loss: 2.3732 (2.7357)  grad: 0.0811 (0.0865)  time: 0.3833  data: 0.0044  max mem: 4840
+train: [0]  [200/400]  eta: 0:01:13  lr: 0.000030  loss: 2.3194 (2.6891)  grad: 0.0824 (0.0859)  time: 0.3576  data: 0.0043  max mem: 4840
+train: [0]  [220/400]  eta: 0:01:06  lr: 0.000033  loss: 2.2266 (2.6442)  grad: 0.0831 (0.0859)  time: 0.3596  data: 0.0042  max mem: 4840
+train: [0]  [240/400]  eta: 0:00:58  lr: 0.000036  loss: 2.1643 (2.6023)  grad: 0.0808 (0.0851)  time: 0.3483  data: 0.0041  max mem: 4840
+train: [0]  [260/400]  eta: 0:00:50  lr: 0.000039  loss: 2.0924 (2.5606)  grad: 0.0750 (0.0843)  time: 0.3277  data: 0.0041  max mem: 4840
+train: [0]  [280/400]  eta: 0:00:43  lr: 0.000042  loss: 2.0384 (2.5212)  grad: 0.0750 (0.0838)  time: 0.3246  data: 0.0041  max mem: 4840
+train: [0]  [300/400]  eta: 0:00:36  lr: 0.000045  loss: 1.9733 (2.4832)  grad: 0.0764 (0.0834)  time: 0.4762  data: 0.1620  max mem: 4840
+train: [0]  [320/400]  eta: 0:00:29  lr: 0.000048  loss: 1.9234 (2.4470)  grad: 0.0773 (0.0829)  time: 0.3296  data: 0.0036  max mem: 4840
+train: [0]  [340/400]  eta: 0:00:21  lr: 0.000051  loss: 1.8917 (2.4141)  grad: 0.0741 (0.0823)  time: 0.3231  data: 0.0032  max mem: 4840
+train: [0]  [360/400]  eta: 0:00:14  lr: 0.000054  loss: 1.8509 (2.3815)  grad: 0.0714 (0.0818)  time: 0.3373  data: 0.0040  max mem: 4840
+train: [0]  [380/400]  eta: 0:00:07  lr: 0.000057  loss: 1.8141 (2.3500)  grad: 0.0716 (0.0814)  time: 0.3282  data: 0.0041  max mem: 4840
+train: [0]  [399/400]  eta: 0:00:00  lr: 0.000060  loss: 1.7458 (2.3183)  grad: 0.0723 (0.0810)  time: 0.3318  data: 0.0041  max mem: 4840
+train: [0] Total time: 0:02:23 (0.3595 s / it)
+train: [0] Summary: lr: 0.000060  loss: 1.7458 (2.3183)  grad: 0.0723 (0.0810)
+eval (validation): [0]  [ 0/63]  eta: 0:03:06    time: 2.9591  data: 2.7170  max mem: 4840
+eval (validation): [0]  [20/63]  eta: 0:00:19    time: 0.3282  data: 0.0038  max mem: 4840
+eval (validation): [0]  [40/63]  eta: 0:00:08    time: 0.2903  data: 0.0030  max mem: 4840
+eval (validation): [0]  [60/63]  eta: 0:00:01    time: 0.3028  data: 0.0031  max mem: 4840
+eval (validation): [0]  [62/63]  eta: 0:00:00    time: 0.3043  data: 0.0031  max mem: 4840
+eval (validation): [0] Total time: 0:00:22 (0.3539 s / it)
+cv: [0]  best hparam: (19, 1.0) (042) ('042_lr1.9e+01_wd1.0e+00')  loss: 0.112  acc: 0.970  f1: 0.967
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-best.pth
+train: [1]  [  0/400]  eta: 0:20:01  lr: nan  time: 3.0032  data: 2.7767  max mem: 4840
+train: [1]  [ 20/400]  eta: 0:02:54  lr: 0.000063  loss: 1.6774 (1.6939)  grad: 0.0679 (0.0707)  time: 0.3317  data: 0.0035  max mem: 4840
+train: [1]  [ 40/400]  eta: 0:02:23  lr: 0.000066  loss: 1.6774 (1.6788)  grad: 0.0683 (0.0709)  time: 0.3330  data: 0.0030  max mem: 4840
+train: [1]  [ 60/400]  eta: 0:02:07  lr: 0.000069  loss: 1.6434 (1.6597)  grad: 0.0708 (0.0713)  time: 0.3252  data: 0.0040  max mem: 4840
+train: [1]  [ 80/400]  eta: 0:01:59  lr: 0.000072  loss: 1.5837 (1.6377)  grad: 0.0717 (0.0712)  time: 0.3713  data: 0.0042  max mem: 4840
+train: [1]  [100/400]  eta: 0:01:50  lr: 0.000075  loss: 1.5705 (1.6254)  grad: 0.0691 (0.0709)  time: 0.3560  data: 0.0043  max mem: 4840
+train: [1]  [120/400]  eta: 0:01:42  lr: 0.000078  loss: 1.5591 (1.6117)  grad: 0.0673 (0.0704)  time: 0.3582  data: 0.0042  max mem: 4840
+train: [1]  [140/400]  eta: 0:01:34  lr: 0.000081  loss: 1.4840 (1.5897)  grad: 0.0691 (0.0707)  time: 0.3426  data: 0.0039  max mem: 4840
+train: [1]  [160/400]  eta: 0:01:26  lr: 0.000084  loss: 1.4654 (1.5772)  grad: 0.0692 (0.0702)  time: 0.3344  data: 0.0040  max mem: 4840
+train: [1]  [180/400]  eta: 0:01:19  lr: 0.000087  loss: 1.4674 (1.5631)  grad: 0.0667 (0.0700)  time: 0.3635  data: 0.0042  max mem: 4840
+train: [1]  [200/400]  eta: 0:01:12  lr: 0.000090  loss: 1.4233 (1.5491)  grad: 0.0667 (0.0697)  time: 0.3583  data: 0.0039  max mem: 4840
+train: [1]  [220/400]  eta: 0:01:05  lr: 0.000093  loss: 1.4027 (1.5347)  grad: 0.0665 (0.0695)  time: 0.3760  data: 0.0039  max mem: 4840
+train: [1]  [240/400]  eta: 0:00:58  lr: 0.000096  loss: 1.3869 (1.5235)  grad: 0.0656 (0.0692)  time: 0.3708  data: 0.0041  max mem: 4840
+train: [1]  [260/400]  eta: 0:00:50  lr: 0.000099  loss: 1.3462 (1.5082)  grad: 0.0652 (0.0690)  time: 0.3304  data: 0.0041  max mem: 4840
+train: [1]  [280/400]  eta: 0:00:43  lr: 0.000102  loss: 1.3164 (1.4947)  grad: 0.0677 (0.0690)  time: 0.3492  data: 0.0046  max mem: 4840
+train: [1]  [300/400]  eta: 0:00:37  lr: 0.000105  loss: 1.3244 (1.4833)  grad: 0.0659 (0.0688)  time: 0.5221  data: 0.1868  max mem: 4840
+train: [1]  [320/400]  eta: 0:00:29  lr: 0.000108  loss: 1.2948 (1.4709)  grad: 0.0638 (0.0686)  time: 0.3601  data: 0.0030  max mem: 4840
+train: [1]  [340/400]  eta: 0:00:22  lr: 0.000111  loss: 1.2721 (1.4583)  grad: 0.0650 (0.0685)  time: 0.3550  data: 0.0037  max mem: 4840
+train: [1]  [360/400]  eta: 0:00:14  lr: 0.000114  loss: 1.2497 (1.4474)  grad: 0.0637 (0.0682)  time: 0.3393  data: 0.0040  max mem: 4840
+train: [1]  [380/400]  eta: 0:00:07  lr: 0.000117  loss: 1.2332 (1.4352)  grad: 0.0646 (0.0681)  time: 0.3695  data: 0.0042  max mem: 4840
+train: [1]  [399/400]  eta: 0:00:00  lr: 0.000120  loss: 1.1928 (1.4233)  grad: 0.0677 (0.0682)  time: 0.3653  data: 0.0038  max mem: 4840
+train: [1] Total time: 0:02:27 (0.3676 s / it)
+train: [1] Summary: lr: 0.000120  loss: 1.1928 (1.4233)  grad: 0.0677 (0.0682)
+eval (validation): [1]  [ 0/63]  eta: 0:03:29    time: 3.3323  data: 3.0395  max mem: 4840
+eval (validation): [1]  [20/63]  eta: 0:00:20    time: 0.3435  data: 0.0043  max mem: 4840
+eval (validation): [1]  [40/63]  eta: 0:00:09    time: 0.3130  data: 0.0031  max mem: 4840
+eval (validation): [1]  [60/63]  eta: 0:00:01    time: 0.2925  data: 0.0033  max mem: 4840
+eval (validation): [1]  [62/63]  eta: 0:00:00    time: 0.2933  data: 0.0033  max mem: 4840
+eval (validation): [1] Total time: 0:00:23 (0.3686 s / it)
+cv: [1]  best hparam: (22, 1.0) (043) ('043_lr2.2e+01_wd1.0e+00')  loss: 0.087  acc: 0.974  f1: 0.973
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-best.pth
+train: [2]  [  0/400]  eta: 0:20:14  lr: nan  time: 3.0351  data: 2.7609  max mem: 4840
+train: [2]  [ 20/400]  eta: 0:03:04  lr: 0.000123  loss: 1.1577 (1.1685)  grad: 0.0615 (0.0643)  time: 0.3578  data: 0.0030  max mem: 4840
+train: [2]  [ 40/400]  eta: 0:02:26  lr: 0.000126  loss: 1.1581 (1.1740)  grad: 0.0650 (0.0654)  time: 0.3224  data: 0.0039  max mem: 4840
+train: [2]  [ 60/400]  eta: 0:02:09  lr: 0.000129  loss: 1.1667 (1.1712)  grad: 0.0665 (0.0656)  time: 0.3303  data: 0.0042  max mem: 4840
+train: [2]  [ 80/400]  eta: 0:01:58  lr: 0.000132  loss: 1.1459 (1.1609)  grad: 0.0667 (0.0663)  time: 0.3345  data: 0.0041  max mem: 4840
+train: [2]  [100/400]  eta: 0:01:49  lr: 0.000135  loss: 1.1409 (1.1560)  grad: 0.0673 (0.0660)  time: 0.3390  data: 0.0040  max mem: 4840
+train: [2]  [120/400]  eta: 0:01:40  lr: 0.000138  loss: 1.1099 (1.1464)  grad: 0.0638 (0.0655)  time: 0.3343  data: 0.0041  max mem: 4840
+train: [2]  [140/400]  eta: 0:01:31  lr: 0.000141  loss: 1.0895 (1.1365)  grad: 0.0607 (0.0649)  time: 0.3239  data: 0.0041  max mem: 4840
+train: [2]  [160/400]  eta: 0:01:24  lr: 0.000144  loss: 1.0812 (1.1270)  grad: 0.0625 (0.0649)  time: 0.3244  data: 0.0040  max mem: 4840
+train: [2]  [180/400]  eta: 0:01:16  lr: 0.000147  loss: 1.0643 (1.1183)  grad: 0.0629 (0.0645)  time: 0.3176  data: 0.0040  max mem: 4840
+train: [2]  [200/400]  eta: 0:01:08  lr: 0.000150  loss: 1.0504 (1.1100)  grad: 0.0605 (0.0641)  time: 0.3286  data: 0.0039  max mem: 4840
+train: [2]  [220/400]  eta: 0:01:01  lr: 0.000153  loss: 1.0341 (1.1032)  grad: 0.0605 (0.0640)  time: 0.3229  data: 0.0040  max mem: 4840
+train: [2]  [240/400]  eta: 0:00:54  lr: 0.000156  loss: 1.0191 (1.0967)  grad: 0.0621 (0.0638)  time: 0.3287  data: 0.0041  max mem: 4840
+train: [2]  [260/400]  eta: 0:00:47  lr: 0.000159  loss: 1.0065 (1.0892)  grad: 0.0633 (0.0640)  time: 0.3371  data: 0.0041  max mem: 4840
+train: [2]  [280/400]  eta: 0:00:40  lr: 0.000162  loss: 0.9867 (1.0822)  grad: 0.0627 (0.0637)  time: 0.3328  data: 0.0040  max mem: 4840
+train: [2]  [300/400]  eta: 0:00:35  lr: 0.000165  loss: 0.9747 (1.0743)  grad: 0.0625 (0.0638)  time: 0.5270  data: 0.1810  max mem: 4840
+train: [2]  [320/400]  eta: 0:00:28  lr: 0.000168  loss: 0.9490 (1.0664)  grad: 0.0643 (0.0639)  time: 0.4102  data: 0.0042  max mem: 4840
+train: [2]  [340/400]  eta: 0:00:21  lr: 0.000171  loss: 0.9416 (1.0599)  grad: 0.0652 (0.0641)  time: 0.3593  data: 0.0042  max mem: 4840
+train: [2]  [360/400]  eta: 0:00:14  lr: 0.000174  loss: 0.9225 (1.0524)  grad: 0.0668 (0.0642)  time: 0.3502  data: 0.0043  max mem: 4840
+train: [2]  [380/400]  eta: 0:00:07  lr: 0.000177  loss: 0.9036 (1.0453)  grad: 0.0655 (0.0643)  time: 0.3657  data: 0.0040  max mem: 4840
+train: [2]  [399/400]  eta: 0:00:00  lr: 0.000180  loss: 0.8949 (1.0375)  grad: 0.0661 (0.0646)  time: 0.3609  data: 0.0043  max mem: 4840
+train: [2] Total time: 0:02:23 (0.3575 s / it)
+train: [2] Summary: lr: 0.000180  loss: 0.8949 (1.0375)  grad: 0.0661 (0.0646)
+eval (validation): [2]  [ 0/63]  eta: 0:03:28    time: 3.3137  data: 3.0760  max mem: 4840
+eval (validation): [2]  [20/63]  eta: 0:00:21    time: 0.3704  data: 0.0040  max mem: 4840
+eval (validation): [2]  [40/63]  eta: 0:00:09    time: 0.3370  data: 0.0027  max mem: 4840
+eval (validation): [2]  [60/63]  eta: 0:00:01    time: 0.3254  data: 0.0033  max mem: 4840
+eval (validation): [2]  [62/63]  eta: 0:00:00    time: 0.3228  data: 0.0033  max mem: 4840
+eval (validation): [2] Total time: 0:00:24 (0.3950 s / it)
+cv: [2]  best hparam: (5.1, 1.0) (034) ('034_lr5.1e+00_wd1.0e+00')  loss: 0.081  acc: 0.977  f1: 0.976
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-best.pth
+train: [3]  [  0/400]  eta: 0:22:11  lr: nan  time: 3.3285  data: 3.0833  max mem: 4840
+train: [3]  [ 20/400]  eta: 0:03:20  lr: 0.000183  loss: 0.8712 (0.8937)  grad: 0.0702 (0.0709)  time: 0.3885  data: 0.0035  max mem: 4840
+train: [3]  [ 40/400]  eta: 0:02:34  lr: 0.000186  loss: 0.8712 (0.8929)  grad: 0.0702 (0.0715)  time: 0.3274  data: 0.0034  max mem: 4840
+train: [3]  [ 60/400]  eta: 0:02:14  lr: 0.000189  loss: 0.8898 (0.8960)  grad: 0.0745 (0.0756)  time: 0.3279  data: 0.0044  max mem: 4840
+train: [3]  [ 80/400]  eta: 0:02:02  lr: 0.000192  loss: 0.8873 (0.8935)  grad: 0.0819 (0.0789)  time: 0.3411  data: 0.0037  max mem: 4840
+train: [3]  [100/400]  eta: 0:01:53  lr: 0.000195  loss: 0.9034 (0.8983)  grad: 0.0867 (0.0813)  time: 0.3574  data: 0.0040  max mem: 4840
+train: [3]  [120/400]  eta: 0:01:46  lr: 0.000198  loss: 0.8975 (0.8979)  grad: 0.0891 (0.0842)  time: 0.3913  data: 0.0042  max mem: 4840
+train: [3]  [140/400]  eta: 0:01:38  lr: 0.000201  loss: 0.8698 (0.8973)  grad: 0.0891 (0.0850)  time: 0.3760  data: 0.0042  max mem: 4840
+train: [3]  [160/400]  eta: 0:01:30  lr: 0.000204  loss: 0.8517 (0.8910)  grad: 0.0888 (0.0861)  time: 0.3519  data: 0.0041  max mem: 4840
+train: [3]  [180/400]  eta: 0:01:22  lr: 0.000207  loss: 0.8207 (0.8859)  grad: 0.0888 (0.0865)  time: 0.3656  data: 0.0043  max mem: 4840
+train: [3]  [200/400]  eta: 0:01:14  lr: 0.000210  loss: 0.8351 (0.8835)  grad: 0.0932 (0.0877)  time: 0.3558  data: 0.0045  max mem: 4840
+train: [3]  [220/400]  eta: 0:01:07  lr: 0.000213  loss: 0.8468 (0.8809)  grad: 0.0987 (0.0895)  time: 0.3753  data: 0.0044  max mem: 4840
+train: [3]  [240/400]  eta: 0:00:59  lr: 0.000216  loss: 0.8468 (0.8798)  grad: 0.0980 (0.0904)  time: 0.3623  data: 0.0041  max mem: 4840
+train: [3]  [260/400]  eta: 0:00:51  lr: 0.000219  loss: 0.8320 (0.8753)  grad: 0.0933 (0.0907)  time: 0.3446  data: 0.0041  max mem: 4840
+train: [3]  [280/400]  eta: 0:00:44  lr: 0.000222  loss: 0.7963 (0.8698)  grad: 0.0926 (0.0911)  time: 0.3321  data: 0.0043  max mem: 4840
+train: [3]  [300/400]  eta: 0:00:37  lr: 0.000225  loss: 0.8118 (0.8685)  grad: 0.0963 (0.0917)  time: 0.5185  data: 0.1853  max mem: 4840
+train: [3]  [320/400]  eta: 0:00:29  lr: 0.000228  loss: 0.8167 (0.8653)  grad: 0.1019 (0.0926)  time: 0.3326  data: 0.0031  max mem: 4840
+train: [3]  [340/400]  eta: 0:00:22  lr: 0.000231  loss: 0.8125 (0.8614)  grad: 0.0995 (0.0928)  time: 0.3703  data: 0.0037  max mem: 4840
+train: [3]  [360/400]  eta: 0:00:14  lr: 0.000234  loss: 0.7880 (0.8580)  grad: 0.0951 (0.0931)  time: 0.3354  data: 0.0042  max mem: 4840
+train: [3]  [380/400]  eta: 0:00:07  lr: 0.000237  loss: 0.7695 (0.8538)  grad: 0.0933 (0.0931)  time: 0.3929  data: 0.0045  max mem: 4840
+train: [3]  [399/400]  eta: 0:00:00  lr: 0.000240  loss: 0.7695 (0.8496)  grad: 0.0894 (0.0930)  time: 0.3698  data: 0.0044  max mem: 4840
+train: [3] Total time: 0:02:29 (0.3737 s / it)
+train: [3] Summary: lr: 0.000240  loss: 0.7695 (0.8496)  grad: 0.0894 (0.0930)
+eval (validation): [3]  [ 0/63]  eta: 0:03:22    time: 3.2064  data: 2.9753  max mem: 4840
+eval (validation): [3]  [20/63]  eta: 0:00:21    time: 0.3675  data: 0.0046  max mem: 4840
+eval (validation): [3]  [40/63]  eta: 0:00:09    time: 0.3496  data: 0.0034  max mem: 4840
+eval (validation): [3]  [60/63]  eta: 0:00:01    time: 0.3188  data: 0.0038  max mem: 4840
+eval (validation): [3]  [62/63]  eta: 0:00:00    time: 0.3184  data: 0.0039  max mem: 4840
+eval (validation): [3] Total time: 0:00:24 (0.3947 s / it)
+cv: [3]  best hparam: (2.7, 1.0) (030) ('030_lr2.7e+00_wd1.0e+00')  loss: 0.075  acc: 0.979  f1: 0.975
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-best.pth
+train: [4]  [  0/400]  eta: 0:22:57  lr: nan  time: 3.4425  data: 3.1314  max mem: 4840
+train: [4]  [ 20/400]  eta: 0:03:08  lr: 0.000243  loss: 0.6823 (0.6854)  grad: 0.0800 (0.0840)  time: 0.3474  data: 0.0023  max mem: 4840
+train: [4]  [ 40/400]  eta: 0:02:29  lr: 0.000246  loss: 0.7244 (0.7352)  grad: 0.0896 (0.0902)  time: 0.3301  data: 0.0034  max mem: 4840
+train: [4]  [ 60/400]  eta: 0:02:11  lr: 0.000249  loss: 0.7558 (0.7298)  grad: 0.0911 (0.0903)  time: 0.3342  data: 0.0043  max mem: 4840
+train: [4]  [ 80/400]  eta: 0:02:02  lr: 0.000252  loss: 0.7404 (0.7362)  grad: 0.0911 (0.0915)  time: 0.3666  data: 0.0040  max mem: 4840
+train: [4]  [100/400]  eta: 0:01:53  lr: 0.000255  loss: 0.7391 (0.7358)  grad: 0.0934 (0.0919)  time: 0.3570  data: 0.0041  max mem: 4840
+train: [4]  [120/400]  eta: 0:01:44  lr: 0.000258  loss: 0.7402 (0.7459)  grad: 0.0947 (0.0929)  time: 0.3457  data: 0.0045  max mem: 4840
+train: [4]  [140/400]  eta: 0:01:36  lr: 0.000261  loss: 0.7329 (0.7349)  grad: 0.0883 (0.0922)  time: 0.3526  data: 0.0041  max mem: 4840
+train: [4]  [160/400]  eta: 0:01:28  lr: 0.000264  loss: 0.7065 (0.7318)  grad: 0.0881 (0.0928)  time: 0.3529  data: 0.0042  max mem: 4840
+train: [4]  [180/400]  eta: 0:01:20  lr: 0.000267  loss: 0.7295 (0.7325)  grad: 0.0910 (0.0931)  time: 0.3466  data: 0.0041  max mem: 4840
+train: [4]  [200/400]  eta: 0:01:12  lr: 0.000270  loss: 0.6747 (0.7266)  grad: 0.0912 (0.0933)  time: 0.3522  data: 0.0042  max mem: 4840
+train: [4]  [220/400]  eta: 0:01:05  lr: 0.000273  loss: 0.6634 (0.7255)  grad: 0.0912 (0.0929)  time: 0.3575  data: 0.0043  max mem: 4840
+train: [4]  [240/400]  eta: 0:00:58  lr: 0.000276  loss: 0.6831 (0.7259)  grad: 0.0921 (0.0936)  time: 0.3565  data: 0.0039  max mem: 4840
+train: [4]  [260/400]  eta: 0:00:50  lr: 0.000279  loss: 0.6817 (0.7240)  grad: 0.0972 (0.0939)  time: 0.3624  data: 0.0042  max mem: 4840
+train: [4]  [280/400]  eta: 0:00:43  lr: 0.000282  loss: 0.6496 (0.7209)  grad: 0.0999 (0.0951)  time: 0.3530  data: 0.0042  max mem: 4840
+train: [4]  [300/400]  eta: 0:00:37  lr: 0.000285  loss: 0.6579 (0.7197)  grad: 0.1016 (0.0965)  time: 0.5019  data: 0.1652  max mem: 4840
+train: [4]  [320/400]  eta: 0:00:29  lr: 0.000288  loss: 0.6720 (0.7170)  grad: 0.0986 (0.0972)  time: 0.3577  data: 0.0041  max mem: 4840
+train: [4]  [340/400]  eta: 0:00:22  lr: 0.000291  loss: 0.6245 (0.7125)  grad: 0.0969 (0.0980)  time: 0.3451  data: 0.0036  max mem: 4840
+train: [4]  [360/400]  eta: 0:00:14  lr: 0.000294  loss: 0.6245 (0.7092)  grad: 0.1016 (0.0983)  time: 0.3411  data: 0.0041  max mem: 4840
+train: [4]  [380/400]  eta: 0:00:07  lr: 0.000297  loss: 0.6157 (0.7069)  grad: 0.0939 (0.0981)  time: 0.3532  data: 0.0039  max mem: 4840
+train: [4]  [399/400]  eta: 0:00:00  lr: 0.000300  loss: 0.6157 (0.7075)  grad: 0.0976 (0.0986)  time: 0.3619  data: 0.0041  max mem: 4840
+train: [4] Total time: 0:02:26 (0.3671 s / it)
+train: [4] Summary: lr: 0.000300  loss: 0.6157 (0.7075)  grad: 0.0976 (0.0986)
+eval (validation): [4]  [ 0/63]  eta: 0:03:55    time: 3.7431  data: 3.4714  max mem: 4840
+eval (validation): [4]  [20/63]  eta: 0:00:21    time: 0.3292  data: 0.0028  max mem: 4840
+eval (validation): [4]  [40/63]  eta: 0:00:09    time: 0.3350  data: 0.0033  max mem: 4840
+eval (validation): [4]  [60/63]  eta: 0:00:01    time: 0.3070  data: 0.0034  max mem: 4840
+eval (validation): [4]  [62/63]  eta: 0:00:00    time: 0.3043  data: 0.0034  max mem: 4840
+eval (validation): [4] Total time: 0:00:24 (0.3818 s / it)
+cv: [4]  best hparam: (3.7, 1.0) (032) ('032_lr3.7e+00_wd1.0e+00')  loss: 0.069  acc: 0.979  f1: 0.977
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [5]  [  0/400]  eta: 0:27:49  lr: nan  time: 4.1737  data: 3.8651  max mem: 4840
+train: [5]  [ 20/400]  eta: 0:03:29  lr: 0.000300  loss: 0.5889 (0.5851)  grad: 0.0897 (0.0924)  time: 0.3715  data: 0.0021  max mem: 4840
+train: [5]  [ 40/400]  eta: 0:02:44  lr: 0.000300  loss: 0.5889 (0.5738)  grad: 0.0948 (0.0946)  time: 0.3542  data: 0.0039  max mem: 4840
+train: [5]  [ 60/400]  eta: 0:02:21  lr: 0.000300  loss: 0.5567 (0.5777)  grad: 0.1015 (0.0978)  time: 0.3324  data: 0.0040  max mem: 4840
+train: [5]  [ 80/400]  eta: 0:02:07  lr: 0.000300  loss: 0.5825 (0.5858)  grad: 0.1015 (0.1007)  time: 0.3506  data: 0.0041  max mem: 4840
+train: [5]  [100/400]  eta: 0:01:57  lr: 0.000300  loss: 0.5769 (0.5872)  grad: 0.0996 (0.1013)  time: 0.3657  data: 0.0040  max mem: 4840
+train: [5]  [120/400]  eta: 0:01:49  lr: 0.000300  loss: 0.5887 (0.6007)  grad: 0.1082 (0.1040)  time: 0.3777  data: 0.0043  max mem: 4840
+train: [5]  [140/400]  eta: 0:01:40  lr: 0.000300  loss: 0.6197 (0.6065)  grad: 0.1159 (0.1049)  time: 0.3668  data: 0.0041  max mem: 4840
+train: [5]  [160/400]  eta: 0:01:32  lr: 0.000299  loss: 0.5940 (0.6046)  grad: 0.1090 (0.1044)  time: 0.3594  data: 0.0044  max mem: 4840
+train: [5]  [180/400]  eta: 0:01:23  lr: 0.000299  loss: 0.5692 (0.6059)  grad: 0.0991 (0.1044)  time: 0.3636  data: 0.0041  max mem: 4840
+train: [5]  [200/400]  eta: 0:01:15  lr: 0.000299  loss: 0.6002 (0.6114)  grad: 0.1006 (0.1050)  time: 0.3660  data: 0.0041  max mem: 4840
+train: [5]  [220/400]  eta: 0:01:08  lr: 0.000299  loss: 0.5821 (0.6098)  grad: 0.1092 (0.1059)  time: 0.3621  data: 0.0043  max mem: 4840
+train: [5]  [240/400]  eta: 0:01:00  lr: 0.000299  loss: 0.5743 (0.6103)  grad: 0.1092 (0.1065)  time: 0.3588  data: 0.0043  max mem: 4840
+train: [5]  [260/400]  eta: 0:00:52  lr: 0.000299  loss: 0.6440 (0.6152)  grad: 0.1192 (0.1081)  time: 0.3568  data: 0.0040  max mem: 4840
+train: [5]  [280/400]  eta: 0:00:44  lr: 0.000298  loss: 0.6440 (0.6158)  grad: 0.1159 (0.1080)  time: 0.3642  data: 0.0042  max mem: 4840
+train: [5]  [300/400]  eta: 0:00:38  lr: 0.000298  loss: 0.5472 (0.6108)  grad: 0.0970 (0.1070)  time: 0.5047  data: 0.1743  max mem: 4840
+train: [5]  [320/400]  eta: 0:00:30  lr: 0.000298  loss: 0.4968 (0.6045)  grad: 0.0947 (0.1066)  time: 0.3234  data: 0.0035  max mem: 4840
+train: [5]  [340/400]  eta: 0:00:22  lr: 0.000298  loss: 0.4979 (0.6000)  grad: 0.0999 (0.1062)  time: 0.3452  data: 0.0036  max mem: 4840
+train: [5]  [360/400]  eta: 0:00:15  lr: 0.000297  loss: 0.5106 (0.5963)  grad: 0.0988 (0.1061)  time: 0.3591  data: 0.0040  max mem: 4840
+train: [5]  [380/400]  eta: 0:00:07  lr: 0.000297  loss: 0.5168 (0.5915)  grad: 0.0939 (0.1055)  time: 0.3424  data: 0.0040  max mem: 4840
+train: [5]  [399/400]  eta: 0:00:00  lr: 0.000297  loss: 0.5151 (0.5896)  grad: 0.0916 (0.1050)  time: 0.3555  data: 0.0040  max mem: 4840
+train: [5] Total time: 0:02:29 (0.3742 s / it)
+train: [5] Summary: lr: 0.000297  loss: 0.5151 (0.5896)  grad: 0.0916 (0.1050)
+eval (validation): [5]  [ 0/63]  eta: 0:03:24    time: 3.2503  data: 3.0189  max mem: 4840
+eval (validation): [5]  [20/63]  eta: 0:00:19    time: 0.3240  data: 0.0034  max mem: 4840
+eval (validation): [5]  [40/63]  eta: 0:00:09    time: 0.3633  data: 0.0040  max mem: 4840
+eval (validation): [5]  [60/63]  eta: 0:00:01    time: 0.3060  data: 0.0028  max mem: 4840
+eval (validation): [5]  [62/63]  eta: 0:00:00    time: 0.3028  data: 0.0031  max mem: 4840
+eval (validation): [5] Total time: 0:00:23 (0.3808 s / it)
+cv: [5]  best hparam: (2.3, 1.0) (029) ('029_lr2.3e+00_wd1.0e+00')  loss: 0.066  acc: 0.980  f1: 0.977
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-best.pth
+train: [6]  [  0/400]  eta: 0:22:13  lr: nan  time: 3.3346  data: 3.0934  max mem: 4840
+train: [6]  [ 20/400]  eta: 0:03:09  lr: 0.000296  loss: 0.4370 (0.4841)  grad: 0.0921 (0.0999)  time: 0.3563  data: 0.0037  max mem: 4840
+train: [6]  [ 40/400]  eta: 0:02:31  lr: 0.000296  loss: 0.4602 (0.4967)  grad: 0.0894 (0.0945)  time: 0.3393  data: 0.0035  max mem: 4840
+train: [6]  [ 60/400]  eta: 0:02:12  lr: 0.000296  loss: 0.5017 (0.5014)  grad: 0.0857 (0.0920)  time: 0.3230  data: 0.0030  max mem: 4840
+train: [6]  [ 80/400]  eta: 0:02:01  lr: 0.000295  loss: 0.5017 (0.5013)  grad: 0.0857 (0.0908)  time: 0.3538  data: 0.0040  max mem: 4840
+train: [6]  [100/400]  eta: 0:01:53  lr: 0.000295  loss: 0.4658 (0.5074)  grad: 0.0903 (0.0911)  time: 0.3635  data: 0.0044  max mem: 4840
+train: [6]  [120/400]  eta: 0:01:44  lr: 0.000295  loss: 0.4676 (0.5106)  grad: 0.0908 (0.0914)  time: 0.3502  data: 0.0039  max mem: 4840
+train: [6]  [140/400]  eta: 0:01:36  lr: 0.000294  loss: 0.4498 (0.5065)  grad: 0.0965 (0.0918)  time: 0.3594  data: 0.0043  max mem: 4840
+train: [6]  [160/400]  eta: 0:01:28  lr: 0.000294  loss: 0.4498 (0.5025)  grad: 0.0881 (0.0910)  time: 0.3592  data: 0.0042  max mem: 4840
+train: [6]  [180/400]  eta: 0:01:21  lr: 0.000293  loss: 0.4862 (0.5015)  grad: 0.0838 (0.0911)  time: 0.3654  data: 0.0044  max mem: 4840
+train: [6]  [200/400]  eta: 0:01:13  lr: 0.000293  loss: 0.4790 (0.5000)  grad: 0.1039 (0.0945)  time: 0.3551  data: 0.0041  max mem: 4840
+train: [6]  [220/400]  eta: 0:01:06  lr: 0.000292  loss: 0.4682 (0.4983)  grad: 0.0949 (0.0940)  time: 0.3609  data: 0.0043  max mem: 4840
+train: [6]  [240/400]  eta: 0:00:58  lr: 0.000292  loss: 0.4682 (0.4990)  grad: 0.0896 (0.0946)  time: 0.3718  data: 0.0042  max mem: 4840
+train: [6]  [260/400]  eta: 0:00:51  lr: 0.000291  loss: 0.4652 (0.4954)  grad: 0.0947 (0.0944)  time: 0.3507  data: 0.0041  max mem: 4840
+train: [6]  [280/400]  eta: 0:00:43  lr: 0.000291  loss: 0.3949 (0.4890)  grad: 0.0759 (0.0929)  time: 0.3685  data: 0.0043  max mem: 4840
+train: [6]  [300/400]  eta: 0:00:37  lr: 0.000290  loss: 0.4080 (0.4842)  grad: 0.0866 (0.0928)  time: 0.5443  data: 0.1817  max mem: 4840
+train: [6]  [320/400]  eta: 0:00:30  lr: 0.000290  loss: 0.4208 (0.4820)  grad: 0.0864 (0.0921)  time: 0.3509  data: 0.0036  max mem: 4840
+train: [6]  [340/400]  eta: 0:00:22  lr: 0.000289  loss: 0.4157 (0.4772)  grad: 0.0838 (0.0918)  time: 0.3380  data: 0.0041  max mem: 4840
+train: [6]  [360/400]  eta: 0:00:14  lr: 0.000288  loss: 0.3918 (0.4741)  grad: 0.0867 (0.0913)  time: 0.3504  data: 0.0043  max mem: 4840
+train: [6]  [380/400]  eta: 0:00:07  lr: 0.000288  loss: 0.4211 (0.4718)  grad: 0.0783 (0.0907)  time: 0.3492  data: 0.0041  max mem: 4840
+train: [6]  [399/400]  eta: 0:00:00  lr: 0.000287  loss: 0.3958 (0.4683)  grad: 0.0749 (0.0899)  time: 0.3604  data: 0.0043  max mem: 4840
+train: [6] Total time: 0:02:28 (0.3714 s / it)
+train: [6] Summary: lr: 0.000287  loss: 0.3958 (0.4683)  grad: 0.0749 (0.0899)
+eval (validation): [6]  [ 0/63]  eta: 0:03:26    time: 3.2833  data: 3.0492  max mem: 4840
+eval (validation): [6]  [20/63]  eta: 0:00:20    time: 0.3391  data: 0.0034  max mem: 4840
+eval (validation): [6]  [40/63]  eta: 0:00:09    time: 0.3586  data: 0.0031  max mem: 4840
+eval (validation): [6]  [60/63]  eta: 0:00:01    time: 0.3151  data: 0.0032  max mem: 4840
+eval (validation): [6]  [62/63]  eta: 0:00:00    time: 0.3140  data: 0.0032  max mem: 4840
+eval (validation): [6] Total time: 0:00:24 (0.3880 s / it)
+cv: [6]  best hparam: (1.6, 1.0) (027) ('027_lr1.6e+00_wd1.0e+00')  loss: 0.066  acc: 0.981  f1: 0.979
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-best.pth
+train: [7]  [  0/400]  eta: 0:21:56  lr: nan  time: 3.2914  data: 3.0386  max mem: 4840
+train: [7]  [ 20/400]  eta: 0:03:15  lr: 0.000286  loss: 0.3650 (0.3892)  grad: 0.0647 (0.0674)  time: 0.3754  data: 0.0041  max mem: 4840
+train: [7]  [ 40/400]  eta: 0:02:38  lr: 0.000286  loss: 0.3706 (0.3936)  grad: 0.0657 (0.0672)  time: 0.3645  data: 0.0038  max mem: 4840
+train: [7]  [ 60/400]  eta: 0:02:19  lr: 0.000285  loss: 0.3605 (0.3820)  grad: 0.0642 (0.0650)  time: 0.3441  data: 0.0041  max mem: 4840
+train: [7]  [ 80/400]  eta: 0:02:04  lr: 0.000284  loss: 0.3555 (0.3754)  grad: 0.0633 (0.0647)  time: 0.3322  data: 0.0042  max mem: 4840
+train: [7]  [100/400]  eta: 0:01:55  lr: 0.000284  loss: 0.3580 (0.3756)  grad: 0.0645 (0.0654)  time: 0.3707  data: 0.0043  max mem: 4840
+train: [7]  [120/400]  eta: 0:01:47  lr: 0.000283  loss: 0.4003 (0.3819)  grad: 0.0688 (0.0663)  time: 0.3612  data: 0.0043  max mem: 4840
+train: [7]  [140/400]  eta: 0:01:38  lr: 0.000282  loss: 0.4004 (0.3833)  grad: 0.0702 (0.0681)  time: 0.3610  data: 0.0042  max mem: 4840
+train: [7]  [160/400]  eta: 0:01:30  lr: 0.000282  loss: 0.3659 (0.3830)  grad: 0.0702 (0.0682)  time: 0.3483  data: 0.0042  max mem: 4840
+train: [7]  [180/400]  eta: 0:01:21  lr: 0.000281  loss: 0.3612 (0.3839)  grad: 0.0677 (0.0688)  time: 0.3414  data: 0.0043  max mem: 4840
+train: [7]  [200/400]  eta: 0:01:14  lr: 0.000280  loss: 0.3694 (0.3812)  grad: 0.0677 (0.0688)  time: 0.3630  data: 0.0041  max mem: 4840
+train: [7]  [220/400]  eta: 0:01:06  lr: 0.000279  loss: 0.3687 (0.3805)  grad: 0.0677 (0.0686)  time: 0.3604  data: 0.0037  max mem: 4840
+train: [7]  [240/400]  eta: 0:00:59  lr: 0.000278  loss: 0.3836 (0.3816)  grad: 0.0712 (0.0693)  time: 0.3612  data: 0.0042  max mem: 4840
+train: [7]  [260/400]  eta: 0:00:51  lr: 0.000278  loss: 0.3649 (0.3800)  grad: 0.0772 (0.0701)  time: 0.3605  data: 0.0043  max mem: 4840
+train: [7]  [280/400]  eta: 0:00:44  lr: 0.000277  loss: 0.3529 (0.3823)  grad: 0.0790 (0.0706)  time: 0.3661  data: 0.0043  max mem: 4840
+train: [7]  [300/400]  eta: 0:00:37  lr: 0.000276  loss: 0.3529 (0.3798)  grad: 0.0720 (0.0705)  time: 0.5297  data: 0.1804  max mem: 4840
+train: [7]  [320/400]  eta: 0:00:30  lr: 0.000275  loss: 0.3309 (0.3771)  grad: 0.0602 (0.0697)  time: 0.3737  data: 0.0041  max mem: 4840
+train: [7]  [340/400]  eta: 0:00:22  lr: 0.000274  loss: 0.3287 (0.3746)  grad: 0.0589 (0.0694)  time: 0.3459  data: 0.0035  max mem: 4840
+train: [7]  [360/400]  eta: 0:00:14  lr: 0.000273  loss: 0.3061 (0.3715)  grad: 0.0583 (0.0689)  time: 0.3438  data: 0.0041  max mem: 4840
+train: [7]  [380/400]  eta: 0:00:07  lr: 0.000272  loss: 0.3001 (0.3677)  grad: 0.0564 (0.0682)  time: 0.3461  data: 0.0040  max mem: 4840
+train: [7]  [399/400]  eta: 0:00:00  lr: 0.000271  loss: 0.3001 (0.3643)  grad: 0.0581 (0.0678)  time: 0.3517  data: 0.0043  max mem: 4840
+train: [7] Total time: 0:02:29 (0.3730 s / it)
+train: [7] Summary: lr: 0.000271  loss: 0.3001 (0.3643)  grad: 0.0581 (0.0678)
+eval (validation): [7]  [ 0/63]  eta: 0:03:37    time: 3.4584  data: 3.1499  max mem: 4840
+eval (validation): [7]  [20/63]  eta: 0:00:20    time: 0.3375  data: 0.0038  max mem: 4840
+eval (validation): [7]  [40/63]  eta: 0:00:09    time: 0.3510  data: 0.0030  max mem: 4840
+eval (validation): [7]  [60/63]  eta: 0:00:01    time: 0.3097  data: 0.0033  max mem: 4840
+eval (validation): [7]  [62/63]  eta: 0:00:00    time: 0.3097  data: 0.0030  max mem: 4840
+eval (validation): [7] Total time: 0:00:24 (0.3865 s / it)
+cv: [7]  best hparam: (1.9, 1.0) (028) ('028_lr1.9e+00_wd1.0e+00')  loss: 0.066  acc: 0.980  f1: 0.977
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [8]  [  0/400]  eta: 0:22:46  lr: nan  time: 3.4166  data: 3.0987  max mem: 4840
+train: [8]  [ 20/400]  eta: 0:03:18  lr: 0.000270  loss: 0.3193 (0.3326)  grad: 0.0631 (0.0648)  time: 0.3773  data: 0.0033  max mem: 4840
+train: [8]  [ 40/400]  eta: 0:02:40  lr: 0.000270  loss: 0.3094 (0.3243)  grad: 0.0630 (0.0610)  time: 0.3633  data: 0.0035  max mem: 4840
+train: [8]  [ 60/400]  eta: 0:02:22  lr: 0.000269  loss: 0.2893 (0.3160)  grad: 0.0523 (0.0589)  time: 0.3640  data: 0.0039  max mem: 4840
+train: [8]  [ 80/400]  eta: 0:02:08  lr: 0.000268  loss: 0.2928 (0.3229)  grad: 0.0603 (0.0597)  time: 0.3454  data: 0.0037  max mem: 4840
+train: [8]  [100/400]  eta: 0:01:56  lr: 0.000267  loss: 0.3198 (0.3239)  grad: 0.0623 (0.0599)  time: 0.3473  data: 0.0040  max mem: 4840
+train: [8]  [120/400]  eta: 0:01:48  lr: 0.000266  loss: 0.3106 (0.3195)  grad: 0.0628 (0.0603)  time: 0.3670  data: 0.0042  max mem: 4840
+train: [8]  [140/400]  eta: 0:01:39  lr: 0.000265  loss: 0.2912 (0.3172)  grad: 0.0578 (0.0596)  time: 0.3610  data: 0.0043  max mem: 4840
+train: [8]  [160/400]  eta: 0:01:31  lr: 0.000264  loss: 0.2912 (0.3173)  grad: 0.0529 (0.0602)  time: 0.3883  data: 0.0041  max mem: 4840
+train: [8]  [180/400]  eta: 0:01:23  lr: 0.000263  loss: 0.2973 (0.3160)  grad: 0.0575 (0.0607)  time: 0.3664  data: 0.0041  max mem: 4840
+train: [8]  [200/400]  eta: 0:01:15  lr: 0.000262  loss: 0.2902 (0.3133)  grad: 0.0554 (0.0600)  time: 0.3617  data: 0.0040  max mem: 4840
+train: [8]  [220/400]  eta: 0:01:08  lr: 0.000260  loss: 0.2660 (0.3118)  grad: 0.0588 (0.0602)  time: 0.3697  data: 0.0041  max mem: 4840
+train: [8]  [240/400]  eta: 0:01:00  lr: 0.000259  loss: 0.2881 (0.3106)  grad: 0.0593 (0.0601)  time: 0.3727  data: 0.0038  max mem: 4840
+train: [8]  [260/400]  eta: 0:00:52  lr: 0.000258  loss: 0.3040 (0.3127)  grad: 0.0587 (0.0604)  time: 0.3737  data: 0.0043  max mem: 4840
+train: [8]  [280/400]  eta: 0:00:45  lr: 0.000257  loss: 0.3086 (0.3112)  grad: 0.0588 (0.0601)  time: 0.3565  data: 0.0042  max mem: 4840
+train: [8]  [300/400]  eta: 0:00:38  lr: 0.000256  loss: 0.2885 (0.3098)  grad: 0.0559 (0.0601)  time: 0.4977  data: 0.1693  max mem: 4840
+train: [8]  [320/400]  eta: 0:00:30  lr: 0.000255  loss: 0.2648 (0.3071)  grad: 0.0539 (0.0601)  time: 0.3619  data: 0.0037  max mem: 4840
+train: [8]  [340/400]  eta: 0:00:22  lr: 0.000254  loss: 0.2597 (0.3051)  grad: 0.0498 (0.0595)  time: 0.3565  data: 0.0029  max mem: 4840
+train: [8]  [360/400]  eta: 0:00:15  lr: 0.000253  loss: 0.2560 (0.3030)  grad: 0.0492 (0.0591)  time: 0.3334  data: 0.0039  max mem: 4840
+train: [8]  [380/400]  eta: 0:00:07  lr: 0.000252  loss: 0.2656 (0.3022)  grad: 0.0506 (0.0588)  time: 0.3458  data: 0.0041  max mem: 4840
+train: [8]  [399/400]  eta: 0:00:00  lr: 0.000250  loss: 0.2640 (0.3005)  grad: 0.0497 (0.0585)  time: 0.3514  data: 0.0042  max mem: 4840
+train: [8] Total time: 0:02:30 (0.3763 s / it)
+train: [8] Summary: lr: 0.000250  loss: 0.2640 (0.3005)  grad: 0.0497 (0.0585)
+eval (validation): [8]  [ 0/63]  eta: 0:03:52    time: 3.6940  data: 3.3908  max mem: 4840
+eval (validation): [8]  [20/63]  eta: 0:00:20    time: 0.3275  data: 0.0028  max mem: 4840
+eval (validation): [8]  [40/63]  eta: 0:00:09    time: 0.3220  data: 0.0034  max mem: 4840
+eval (validation): [8]  [60/63]  eta: 0:00:01    time: 0.3209  data: 0.0026  max mem: 4840
+eval (validation): [8]  [62/63]  eta: 0:00:00    time: 0.3207  data: 0.0029  max mem: 4840
+eval (validation): [8] Total time: 0:00:24 (0.3817 s / it)
+cv: [8]  best hparam: (0.85, 1.0) (023) ('023_lr8.5e-01_wd1.0e+00')  loss: 0.067  acc: 0.982  f1: 0.980
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-best.pth
+train: [9]  [  0/400]  eta: 0:20:42  lr: nan  time: 3.1062  data: 2.8409  max mem: 4840
+train: [9]  [ 20/400]  eta: 0:03:07  lr: 0.000249  loss: 0.2395 (0.2504)  grad: 0.0490 (0.0529)  time: 0.3622  data: 0.0042  max mem: 4840
+train: [9]  [ 40/400]  eta: 0:02:36  lr: 0.000248  loss: 0.2456 (0.2540)  grad: 0.0490 (0.0504)  time: 0.3724  data: 0.0029  max mem: 4840
+train: [9]  [ 60/400]  eta: 0:02:20  lr: 0.000247  loss: 0.2549 (0.2600)  grad: 0.0440 (0.0527)  time: 0.3706  data: 0.0042  max mem: 4840
+train: [9]  [ 80/400]  eta: 0:02:06  lr: 0.000246  loss: 0.2616 (0.2641)  grad: 0.0478 (0.0535)  time: 0.3462  data: 0.0044  max mem: 4840
+train: [9]  [100/400]  eta: 0:01:56  lr: 0.000244  loss: 0.2658 (0.2662)  grad: 0.0520 (0.0539)  time: 0.3507  data: 0.0042  max mem: 4840
+train: [9]  [120/400]  eta: 0:01:47  lr: 0.000243  loss: 0.2477 (0.2663)  grad: 0.0480 (0.0532)  time: 0.3566  data: 0.0041  max mem: 4840
+train: [9]  [140/400]  eta: 0:01:39  lr: 0.000242  loss: 0.2507 (0.2651)  grad: 0.0461 (0.0523)  time: 0.3723  data: 0.0041  max mem: 4840
+train: [9]  [160/400]  eta: 0:01:30  lr: 0.000241  loss: 0.2391 (0.2634)  grad: 0.0461 (0.0516)  time: 0.3661  data: 0.0044  max mem: 4840
+train: [9]  [180/400]  eta: 0:01:22  lr: 0.000240  loss: 0.2313 (0.2612)  grad: 0.0474 (0.0509)  time: 0.3544  data: 0.0047  max mem: 4840
+train: [9]  [200/400]  eta: 0:01:14  lr: 0.000238  loss: 0.2499 (0.2613)  grad: 0.0470 (0.0506)  time: 0.3595  data: 0.0043  max mem: 4840
+train: [9]  [220/400]  eta: 0:01:07  lr: 0.000237  loss: 0.2532 (0.2598)  grad: 0.0483 (0.0506)  time: 0.3550  data: 0.0042  max mem: 4840
+train: [9]  [240/400]  eta: 0:00:59  lr: 0.000236  loss: 0.2495 (0.2604)  grad: 0.0517 (0.0509)  time: 0.3644  data: 0.0041  max mem: 4840
+train: [9]  [260/400]  eta: 0:00:52  lr: 0.000234  loss: 0.2594 (0.2612)  grad: 0.0517 (0.0510)  time: 0.3628  data: 0.0040  max mem: 4840
+train: [9]  [280/400]  eta: 0:00:44  lr: 0.000233  loss: 0.2793 (0.2637)  grad: 0.0546 (0.0520)  time: 0.3615  data: 0.0044  max mem: 4840
+train: [9]  [300/400]  eta: 0:00:38  lr: 0.000232  loss: 0.2763 (0.2632)  grad: 0.0578 (0.0524)  time: 0.5488  data: 0.1781  max mem: 4840
+train: [9]  [320/400]  eta: 0:00:30  lr: 0.000230  loss: 0.2437 (0.2621)  grad: 0.0490 (0.0522)  time: 0.4174  data: 0.0033  max mem: 4840
+train: [9]  [340/400]  eta: 0:00:22  lr: 0.000229  loss: 0.2458 (0.2611)  grad: 0.0478 (0.0518)  time: 0.3475  data: 0.0044  max mem: 4840
+train: [9]  [360/400]  eta: 0:00:15  lr: 0.000228  loss: 0.2481 (0.2609)  grad: 0.0443 (0.0514)  time: 0.3478  data: 0.0040  max mem: 4840
+train: [9]  [380/400]  eta: 0:00:07  lr: 0.000226  loss: 0.2154 (0.2586)  grad: 0.0429 (0.0511)  time: 0.3473  data: 0.0040  max mem: 4840
+train: [9]  [399/400]  eta: 0:00:00  lr: 0.000225  loss: 0.2230 (0.2577)  grad: 0.0447 (0.0509)  time: 0.3498  data: 0.0041  max mem: 4840
+train: [9] Total time: 0:02:31 (0.3781 s / it)
+train: [9] Summary: lr: 0.000225  loss: 0.2230 (0.2577)  grad: 0.0447 (0.0509)
+eval (validation): [9]  [ 0/63]  eta: 0:03:11    time: 3.0474  data: 2.8013  max mem: 4840
+eval (validation): [9]  [20/63]  eta: 0:00:20    time: 0.3451  data: 0.0048  max mem: 4840
+eval (validation): [9]  [40/63]  eta: 0:00:09    time: 0.3381  data: 0.0032  max mem: 4840
+eval (validation): [9]  [60/63]  eta: 0:00:01    time: 0.3073  data: 0.0030  max mem: 4840
+eval (validation): [9]  [62/63]  eta: 0:00:00    time: 0.3080  data: 0.0033  max mem: 4840
+eval (validation): [9] Total time: 0:00:23 (0.3779 s / it)
+cv: [9]  best hparam: (0.72, 1.0) (022) ('022_lr7.2e-01_wd1.0e+00')  loss: 0.069  acc: 0.981  f1: 0.978
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [10]  [  0/400]  eta: 0:22:07  lr: nan  time: 3.3182  data: 3.0618  max mem: 4840
+train: [10]  [ 20/400]  eta: 0:03:13  lr: 0.000224  loss: 0.2403 (0.2403)  grad: 0.0437 (0.0467)  time: 0.3674  data: 0.0038  max mem: 4840
+train: [10]  [ 40/400]  eta: 0:02:35  lr: 0.000222  loss: 0.2418 (0.2403)  grad: 0.0438 (0.0477)  time: 0.3543  data: 0.0034  max mem: 4840
+train: [10]  [ 60/400]  eta: 0:02:19  lr: 0.000221  loss: 0.2378 (0.2368)  grad: 0.0438 (0.0465)  time: 0.3636  data: 0.0041  max mem: 4840
+train: [10]  [ 80/400]  eta: 0:02:06  lr: 0.000220  loss: 0.2365 (0.2378)  grad: 0.0416 (0.0467)  time: 0.3472  data: 0.0039  max mem: 4840
+train: [10]  [100/400]  eta: 0:01:55  lr: 0.000218  loss: 0.2441 (0.2394)  grad: 0.0442 (0.0471)  time: 0.3396  data: 0.0043  max mem: 4840
+train: [10]  [120/400]  eta: 0:01:46  lr: 0.000217  loss: 0.2351 (0.2385)  grad: 0.0442 (0.0465)  time: 0.3540  data: 0.0039  max mem: 4840
+train: [10]  [140/400]  eta: 0:01:37  lr: 0.000215  loss: 0.2264 (0.2361)  grad: 0.0426 (0.0460)  time: 0.3560  data: 0.0039  max mem: 4840
+train: [10]  [160/400]  eta: 0:01:31  lr: 0.000214  loss: 0.2190 (0.2347)  grad: 0.0420 (0.0456)  time: 0.4047  data: 0.0043  max mem: 4840
+train: [10]  [180/400]  eta: 0:01:23  lr: 0.000213  loss: 0.2176 (0.2335)  grad: 0.0427 (0.0459)  time: 0.3817  data: 0.0040  max mem: 4840
+train: [10]  [200/400]  eta: 0:01:15  lr: 0.000211  loss: 0.2163 (0.2330)  grad: 0.0431 (0.0459)  time: 0.3718  data: 0.0044  max mem: 4840
+train: [10]  [220/400]  eta: 0:01:08  lr: 0.000210  loss: 0.2163 (0.2318)  grad: 0.0410 (0.0456)  time: 0.3716  data: 0.0043  max mem: 4840
+train: [10]  [240/400]  eta: 0:01:00  lr: 0.000208  loss: 0.2170 (0.2312)  grad: 0.0401 (0.0452)  time: 0.3689  data: 0.0041  max mem: 4840
+train: [10]  [260/400]  eta: 0:00:52  lr: 0.000207  loss: 0.2099 (0.2296)  grad: 0.0393 (0.0447)  time: 0.3666  data: 0.0041  max mem: 4840
+train: [10]  [280/400]  eta: 0:00:45  lr: 0.000205  loss: 0.2226 (0.2303)  grad: 0.0446 (0.0449)  time: 0.3619  data: 0.0043  max mem: 4840
+train: [10]  [300/400]  eta: 0:00:38  lr: 0.000204  loss: 0.2303 (0.2302)  grad: 0.0445 (0.0446)  time: 0.5522  data: 0.1873  max mem: 4840
+train: [10]  [320/400]  eta: 0:00:30  lr: 0.000202  loss: 0.2092 (0.2291)  grad: 0.0381 (0.0444)  time: 0.3688  data: 0.0031  max mem: 4840
+train: [10]  [340/400]  eta: 0:00:23  lr: 0.000201  loss: 0.2049 (0.2279)  grad: 0.0387 (0.0442)  time: 0.3687  data: 0.0035  max mem: 4840
+train: [10]  [360/400]  eta: 0:00:15  lr: 0.000199  loss: 0.2206 (0.2278)  grad: 0.0385 (0.0440)  time: 0.3523  data: 0.0040  max mem: 4840
+train: [10]  [380/400]  eta: 0:00:07  lr: 0.000198  loss: 0.2085 (0.2265)  grad: 0.0363 (0.0437)  time: 0.3303  data: 0.0042  max mem: 4840
+train: [10]  [399/400]  eta: 0:00:00  lr: 0.000196  loss: 0.2004 (0.2256)  grad: 0.0362 (0.0433)  time: 0.3690  data: 0.0042  max mem: 4840
+train: [10] Total time: 0:02:32 (0.3805 s / it)
+train: [10] Summary: lr: 0.000196  loss: 0.2004 (0.2256)  grad: 0.0362 (0.0433)
+eval (validation): [10]  [ 0/63]  eta: 0:03:24    time: 3.2469  data: 2.9799  max mem: 4840
+eval (validation): [10]  [20/63]  eta: 0:00:21    time: 0.3647  data: 0.0037  max mem: 4840
+eval (validation): [10]  [40/63]  eta: 0:00:09    time: 0.3093  data: 0.0029  max mem: 4840
+eval (validation): [10]  [60/63]  eta: 0:00:01    time: 0.3171  data: 0.0033  max mem: 4840
+eval (validation): [10]  [62/63]  eta: 0:00:00    time: 0.3179  data: 0.0033  max mem: 4840
+eval (validation): [10] Total time: 0:00:24 (0.3814 s / it)
+cv: [10]  best hparam: (1.2, 1.0) (025) ('025_lr1.2e+00_wd1.0e+00')  loss: 0.064  acc: 0.982  f1: 0.979
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [11]  [  0/400]  eta: 0:22:09  lr: nan  time: 3.3226  data: 3.0579  max mem: 4840
+train: [11]  [ 20/400]  eta: 0:03:09  lr: 0.000195  loss: 0.1903 (0.1983)  grad: 0.0365 (0.0367)  time: 0.3565  data: 0.0032  max mem: 4840
+train: [11]  [ 40/400]  eta: 0:02:33  lr: 0.000193  loss: 0.2052 (0.2107)  grad: 0.0366 (0.0380)  time: 0.3542  data: 0.0034  max mem: 4840
+train: [11]  [ 60/400]  eta: 0:02:16  lr: 0.000192  loss: 0.2061 (0.2071)  grad: 0.0366 (0.0375)  time: 0.3507  data: 0.0038  max mem: 4840
+train: [11]  [ 80/400]  eta: 0:02:04  lr: 0.000190  loss: 0.1863 (0.2036)  grad: 0.0371 (0.0377)  time: 0.3473  data: 0.0042  max mem: 4840
+train: [11]  [100/400]  eta: 0:01:53  lr: 0.000189  loss: 0.1947 (0.2051)  grad: 0.0385 (0.0381)  time: 0.3426  data: 0.0042  max mem: 4840
+train: [11]  [120/400]  eta: 0:01:45  lr: 0.000187  loss: 0.1989 (0.2038)  grad: 0.0365 (0.0379)  time: 0.3649  data: 0.0044  max mem: 4840
+train: [11]  [140/400]  eta: 0:01:37  lr: 0.000186  loss: 0.1995 (0.2038)  grad: 0.0364 (0.0378)  time: 0.3514  data: 0.0042  max mem: 4840
+train: [11]  [160/400]  eta: 0:01:28  lr: 0.000184  loss: 0.2037 (0.2025)  grad: 0.0364 (0.0375)  time: 0.3461  data: 0.0041  max mem: 4840
+train: [11]  [180/400]  eta: 0:01:21  lr: 0.000183  loss: 0.2094 (0.2031)  grad: 0.0373 (0.0376)  time: 0.3619  data: 0.0041  max mem: 4840
+train: [11]  [200/400]  eta: 0:01:13  lr: 0.000181  loss: 0.2039 (0.2034)  grad: 0.0373 (0.0375)  time: 0.3602  data: 0.0042  max mem: 4840
+train: [11]  [220/400]  eta: 0:01:06  lr: 0.000180  loss: 0.2014 (0.2042)  grad: 0.0362 (0.0376)  time: 0.3535  data: 0.0042  max mem: 4840
+train: [11]  [240/400]  eta: 0:00:58  lr: 0.000178  loss: 0.2062 (0.2044)  grad: 0.0361 (0.0375)  time: 0.3619  data: 0.0042  max mem: 4840
+train: [11]  [260/400]  eta: 0:00:51  lr: 0.000177  loss: 0.1941 (0.2043)  grad: 0.0341 (0.0372)  time: 0.3443  data: 0.0041  max mem: 4840
+train: [11]  [280/400]  eta: 0:00:43  lr: 0.000175  loss: 0.1941 (0.2038)  grad: 0.0353 (0.0372)  time: 0.3591  data: 0.0043  max mem: 4840
+train: [11]  [300/400]  eta: 0:00:37  lr: 0.000174  loss: 0.1993 (0.2036)  grad: 0.0356 (0.0372)  time: 0.5099  data: 0.1725  max mem: 4840
+train: [11]  [320/400]  eta: 0:00:29  lr: 0.000172  loss: 0.1863 (0.2031)  grad: 0.0348 (0.0371)  time: 0.3726  data: 0.0044  max mem: 4840
+train: [11]  [340/400]  eta: 0:00:22  lr: 0.000170  loss: 0.1793 (0.2018)  grad: 0.0331 (0.0369)  time: 0.3573  data: 0.0026  max mem: 4840
+train: [11]  [360/400]  eta: 0:00:14  lr: 0.000169  loss: 0.1862 (0.2015)  grad: 0.0327 (0.0367)  time: 0.3627  data: 0.0034  max mem: 4840
+train: [11]  [380/400]  eta: 0:00:07  lr: 0.000167  loss: 0.2017 (0.2013)  grad: 0.0327 (0.0365)  time: 0.3785  data: 0.0036  max mem: 4840
+train: [11]  [399/400]  eta: 0:00:00  lr: 0.000166  loss: 0.1907 (0.2005)  grad: 0.0333 (0.0364)  time: 0.3425  data: 0.0040  max mem: 4840
+train: [11] Total time: 0:02:28 (0.3719 s / it)
+train: [11] Summary: lr: 0.000166  loss: 0.1907 (0.2005)  grad: 0.0333 (0.0364)
+eval (validation): [11]  [ 0/63]  eta: 0:03:22    time: 3.2068  data: 2.9420  max mem: 4840
+eval (validation): [11]  [20/63]  eta: 0:00:22    time: 0.3861  data: 0.0057  max mem: 4840
+eval (validation): [11]  [40/63]  eta: 0:00:09    time: 0.3286  data: 0.0036  max mem: 4840
+eval (validation): [11]  [60/63]  eta: 0:00:01    time: 0.3140  data: 0.0035  max mem: 4840
+eval (validation): [11]  [62/63]  eta: 0:00:00    time: 0.3135  data: 0.0035  max mem: 4840
+eval (validation): [11] Total time: 0:00:24 (0.3924 s / it)
+cv: [11]  best hparam: (1, 1.0) (024) ('024_lr1.0e+00_wd1.0e+00')  loss: 0.062  acc: 0.982  f1: 0.979
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [12]  [  0/400]  eta: 0:22:10  lr: nan  time: 3.3274  data: 3.0633  max mem: 4840
+train: [12]  [ 20/400]  eta: 0:03:21  lr: 0.000164  loss: 0.1641 (0.1778)  grad: 0.0336 (0.0358)  time: 0.3915  data: 0.0147  max mem: 4840
+train: [12]  [ 40/400]  eta: 0:02:40  lr: 0.000163  loss: 0.1801 (0.1863)  grad: 0.0335 (0.0354)  time: 0.3582  data: 0.0037  max mem: 4840
+train: [12]  [ 60/400]  eta: 0:02:22  lr: 0.000161  loss: 0.1848 (0.1920)  grad: 0.0326 (0.0345)  time: 0.3585  data: 0.0040  max mem: 4840
+train: [12]  [ 80/400]  eta: 0:02:07  lr: 0.000160  loss: 0.1870 (0.1913)  grad: 0.0331 (0.0344)  time: 0.3442  data: 0.0037  max mem: 4840
+train: [12]  [100/400]  eta: 0:01:56  lr: 0.000158  loss: 0.1875 (0.1908)  grad: 0.0335 (0.0342)  time: 0.3403  data: 0.0041  max mem: 4840
+train: [12]  [120/400]  eta: 0:01:47  lr: 0.000156  loss: 0.1858 (0.1894)  grad: 0.0344 (0.0343)  time: 0.3621  data: 0.0042  max mem: 4840
+train: [12]  [140/400]  eta: 0:01:40  lr: 0.000155  loss: 0.1828 (0.1882)  grad: 0.0345 (0.0343)  time: 0.3909  data: 0.0041  max mem: 4840
+train: [12]  [160/400]  eta: 0:01:31  lr: 0.000153  loss: 0.1712 (0.1868)  grad: 0.0335 (0.0342)  time: 0.3673  data: 0.0044  max mem: 4840
+train: [12]  [180/400]  eta: 0:01:23  lr: 0.000152  loss: 0.1715 (0.1866)  grad: 0.0325 (0.0342)  time: 0.3542  data: 0.0040  max mem: 4840
+train: [12]  [200/400]  eta: 0:01:15  lr: 0.000150  loss: 0.1891 (0.1872)  grad: 0.0328 (0.0346)  time: 0.3816  data: 0.0042  max mem: 4840
+train: [12]  [220/400]  eta: 0:01:08  lr: 0.000149  loss: 0.1841 (0.1865)  grad: 0.0318 (0.0343)  time: 0.3731  data: 0.0042  max mem: 4840
+train: [12]  [240/400]  eta: 0:01:00  lr: 0.000147  loss: 0.1796 (0.1861)  grad: 0.0315 (0.0342)  time: 0.3593  data: 0.0043  max mem: 4840
+train: [12]  [260/400]  eta: 0:00:52  lr: 0.000145  loss: 0.1725 (0.1860)  grad: 0.0325 (0.0341)  time: 0.3635  data: 0.0042  max mem: 4840
+train: [12]  [280/400]  eta: 0:00:44  lr: 0.000144  loss: 0.1836 (0.1859)  grad: 0.0324 (0.0340)  time: 0.3492  data: 0.0041  max mem: 4840
+train: [12]  [300/400]  eta: 0:00:38  lr: 0.000142  loss: 0.1836 (0.1857)  grad: 0.0318 (0.0338)  time: 0.5108  data: 0.1831  max mem: 4840
+train: [12]  [320/400]  eta: 0:00:30  lr: 0.000141  loss: 0.1810 (0.1849)  grad: 0.0297 (0.0336)  time: 0.4149  data: 0.0042  max mem: 4840
+train: [12]  [340/400]  eta: 0:00:23  lr: 0.000139  loss: 0.1671 (0.1835)  grad: 0.0296 (0.0334)  time: 0.3865  data: 0.0037  max mem: 4840
+train: [12]  [360/400]  eta: 0:00:15  lr: 0.000138  loss: 0.1671 (0.1829)  grad: 0.0298 (0.0333)  time: 0.3790  data: 0.0040  max mem: 4840
+train: [12]  [380/400]  eta: 0:00:07  lr: 0.000136  loss: 0.1763 (0.1827)  grad: 0.0306 (0.0332)  time: 0.3494  data: 0.0040  max mem: 4840
+train: [12]  [399/400]  eta: 0:00:00  lr: 0.000134  loss: 0.1806 (0.1825)  grad: 0.0314 (0.0332)  time: 0.3546  data: 0.0040  max mem: 4840
+train: [12] Total time: 0:02:33 (0.3826 s / it)
+train: [12] Summary: lr: 0.000134  loss: 0.1806 (0.1825)  grad: 0.0314 (0.0332)
+eval (validation): [12]  [ 0/63]  eta: 0:03:15    time: 3.1098  data: 2.9053  max mem: 4840
+eval (validation): [12]  [20/63]  eta: 0:00:21    time: 0.3749  data: 0.0114  max mem: 4840
+eval (validation): [12]  [40/63]  eta: 0:00:09    time: 0.3190  data: 0.0036  max mem: 4840
+eval (validation): [12]  [60/63]  eta: 0:00:01    time: 0.3157  data: 0.0025  max mem: 4840
+eval (validation): [12]  [62/63]  eta: 0:00:00    time: 0.3140  data: 0.0023  max mem: 4840
+eval (validation): [12] Total time: 0:00:24 (0.3849 s / it)
+cv: [12]  best hparam: (1.2, 1.0) (025) ('025_lr1.2e+00_wd1.0e+00')  loss: 0.062  acc: 0.981  f1: 0.980
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [13]  [  0/400]  eta: 0:22:44  lr: nan  time: 3.4114  data: 3.1151  max mem: 4840
+train: [13]  [ 20/400]  eta: 0:03:19  lr: 0.000133  loss: 0.1726 (0.1761)  grad: 0.0306 (0.0320)  time: 0.3811  data: 0.0039  max mem: 4840
+train: [13]  [ 40/400]  eta: 0:02:43  lr: 0.000131  loss: 0.1689 (0.1734)  grad: 0.0306 (0.0319)  time: 0.3813  data: 0.0041  max mem: 4840
+train: [13]  [ 60/400]  eta: 0:02:23  lr: 0.000130  loss: 0.1693 (0.1728)  grad: 0.0313 (0.0319)  time: 0.3526  data: 0.0040  max mem: 4840
+train: [13]  [ 80/400]  eta: 0:02:07  lr: 0.000128  loss: 0.1695 (0.1720)  grad: 0.0313 (0.0320)  time: 0.3250  data: 0.0038  max mem: 4840
+train: [13]  [100/400]  eta: 0:01:55  lr: 0.000127  loss: 0.1722 (0.1740)  grad: 0.0301 (0.0318)  time: 0.3376  data: 0.0044  max mem: 4840
+train: [13]  [120/400]  eta: 0:01:46  lr: 0.000125  loss: 0.1758 (0.1732)  grad: 0.0296 (0.0315)  time: 0.3556  data: 0.0043  max mem: 4840
+train: [13]  [140/400]  eta: 0:01:36  lr: 0.000124  loss: 0.1682 (0.1735)  grad: 0.0305 (0.0317)  time: 0.3242  data: 0.0045  max mem: 4840
+train: [13]  [160/400]  eta: 0:01:28  lr: 0.000122  loss: 0.1704 (0.1734)  grad: 0.0307 (0.0316)  time: 0.3262  data: 0.0041  max mem: 4840
+train: [13]  [180/400]  eta: 0:01:19  lr: 0.000120  loss: 0.1727 (0.1735)  grad: 0.0295 (0.0315)  time: 0.3260  data: 0.0040  max mem: 4840
+train: [13]  [200/400]  eta: 0:01:11  lr: 0.000119  loss: 0.1701 (0.1727)  grad: 0.0295 (0.0313)  time: 0.3284  data: 0.0041  max mem: 4840
+train: [13]  [220/400]  eta: 0:01:04  lr: 0.000117  loss: 0.1673 (0.1731)  grad: 0.0317 (0.0314)  time: 0.3376  data: 0.0041  max mem: 4840
+train: [13]  [240/400]  eta: 0:00:56  lr: 0.000116  loss: 0.1604 (0.1718)  grad: 0.0311 (0.0313)  time: 0.3382  data: 0.0040  max mem: 4840
+train: [13]  [260/400]  eta: 0:00:49  lr: 0.000114  loss: 0.1560 (0.1706)  grad: 0.0303 (0.0312)  time: 0.3332  data: 0.0042  max mem: 4840
+train: [13]  [280/400]  eta: 0:00:42  lr: 0.000113  loss: 0.1656 (0.1709)  grad: 0.0298 (0.0312)  time: 0.3230  data: 0.0043  max mem: 4840
+train: [13]  [300/400]  eta: 0:00:36  lr: 0.000111  loss: 0.1669 (0.1703)  grad: 0.0298 (0.0312)  time: 0.4881  data: 0.1657  max mem: 4840
+train: [13]  [320/400]  eta: 0:00:28  lr: 0.000110  loss: 0.1566 (0.1696)  grad: 0.0286 (0.0311)  time: 0.3312  data: 0.0033  max mem: 4840
+train: [13]  [340/400]  eta: 0:00:21  lr: 0.000108  loss: 0.1614 (0.1693)  grad: 0.0287 (0.0310)  time: 0.3371  data: 0.0031  max mem: 4840
+train: [13]  [360/400]  eta: 0:00:14  lr: 0.000107  loss: 0.1676 (0.1688)  grad: 0.0287 (0.0308)  time: 0.3340  data: 0.0042  max mem: 4840
+train: [13]  [380/400]  eta: 0:00:07  lr: 0.000105  loss: 0.1676 (0.1692)  grad: 0.0291 (0.0308)  time: 0.3270  data: 0.0038  max mem: 4840
+train: [13]  [399/400]  eta: 0:00:00  lr: 0.000104  loss: 0.1629 (0.1688)  grad: 0.0296 (0.0307)  time: 0.3323  data: 0.0039  max mem: 4840
+train: [13] Total time: 0:02:21 (0.3541 s / it)
+train: [13] Summary: lr: 0.000104  loss: 0.1629 (0.1688)  grad: 0.0296 (0.0307)
+eval (validation): [13]  [ 0/63]  eta: 0:03:07    time: 2.9726  data: 2.7274  max mem: 4840
+eval (validation): [13]  [20/63]  eta: 0:00:19    time: 0.3236  data: 0.0181  max mem: 4840
+eval (validation): [13]  [40/63]  eta: 0:00:08    time: 0.3189  data: 0.0032  max mem: 4840
+eval (validation): [13]  [60/63]  eta: 0:00:01    time: 0.3273  data: 0.0032  max mem: 4840
+eval (validation): [13]  [62/63]  eta: 0:00:00    time: 0.3237  data: 0.0029  max mem: 4840
+eval (validation): [13] Total time: 0:00:23 (0.3699 s / it)
+cv: [13]  best hparam: (0.72, 1.0) (022) ('022_lr7.2e-01_wd1.0e+00')  loss: 0.064  acc: 0.980  f1: 0.978
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [14]  [  0/400]  eta: 0:23:40  lr: nan  time: 3.5516  data: 3.2291  max mem: 4840
+train: [14]  [ 20/400]  eta: 0:03:24  lr: 0.000102  loss: 0.1722 (0.1720)  grad: 0.0292 (0.0299)  time: 0.3867  data: 0.0030  max mem: 4840
+train: [14]  [ 40/400]  eta: 0:02:40  lr: 0.000101  loss: 0.1722 (0.1756)  grad: 0.0297 (0.0308)  time: 0.3497  data: 0.0030  max mem: 4840
+train: [14]  [ 60/400]  eta: 0:02:20  lr: 0.000099  loss: 0.1654 (0.1719)  grad: 0.0286 (0.0300)  time: 0.3503  data: 0.0041  max mem: 4840
+train: [14]  [ 80/400]  eta: 0:02:07  lr: 0.000098  loss: 0.1597 (0.1688)  grad: 0.0284 (0.0300)  time: 0.3438  data: 0.0036  max mem: 4840
+train: [14]  [100/400]  eta: 0:01:56  lr: 0.000096  loss: 0.1583 (0.1679)  grad: 0.0294 (0.0299)  time: 0.3518  data: 0.0037  max mem: 4840
+train: [14]  [120/400]  eta: 0:01:46  lr: 0.000095  loss: 0.1622 (0.1676)  grad: 0.0287 (0.0297)  time: 0.3370  data: 0.0042  max mem: 4840
+train: [14]  [140/400]  eta: 0:01:37  lr: 0.000093  loss: 0.1636 (0.1676)  grad: 0.0287 (0.0296)  time: 0.3563  data: 0.0043  max mem: 4840
+train: [14]  [160/400]  eta: 0:01:30  lr: 0.000092  loss: 0.1612 (0.1663)  grad: 0.0295 (0.0296)  time: 0.3705  data: 0.0042  max mem: 4840
+train: [14]  [180/400]  eta: 0:01:22  lr: 0.000090  loss: 0.1619 (0.1659)  grad: 0.0290 (0.0297)  time: 0.3766  data: 0.0042  max mem: 4840
+train: [14]  [200/400]  eta: 0:01:15  lr: 0.000089  loss: 0.1584 (0.1650)  grad: 0.0290 (0.0296)  time: 0.3692  data: 0.0042  max mem: 4840
+train: [14]  [220/400]  eta: 0:01:07  lr: 0.000088  loss: 0.1558 (0.1638)  grad: 0.0284 (0.0295)  time: 0.3580  data: 0.0043  max mem: 4840
+train: [14]  [240/400]  eta: 0:00:59  lr: 0.000086  loss: 0.1614 (0.1638)  grad: 0.0284 (0.0295)  time: 0.3674  data: 0.0042  max mem: 4840
+train: [14]  [260/400]  eta: 0:00:52  lr: 0.000085  loss: 0.1574 (0.1632)  grad: 0.0306 (0.0297)  time: 0.3638  data: 0.0043  max mem: 4840
+train: [14]  [280/400]  eta: 0:00:44  lr: 0.000083  loss: 0.1530 (0.1631)  grad: 0.0306 (0.0298)  time: 0.3617  data: 0.0044  max mem: 4840
+train: [14]  [300/400]  eta: 0:00:38  lr: 0.000082  loss: 0.1530 (0.1630)  grad: 0.0291 (0.0297)  time: 0.5414  data: 0.1783  max mem: 4840
+train: [14]  [320/400]  eta: 0:00:30  lr: 0.000081  loss: 0.1575 (0.1627)  grad: 0.0294 (0.0297)  time: 0.3751  data: 0.0040  max mem: 4840
+train: [14]  [340/400]  eta: 0:00:22  lr: 0.000079  loss: 0.1550 (0.1627)  grad: 0.0290 (0.0297)  time: 0.3612  data: 0.0040  max mem: 4840
+train: [14]  [360/400]  eta: 0:00:15  lr: 0.000078  loss: 0.1605 (0.1628)  grad: 0.0281 (0.0296)  time: 0.3480  data: 0.0042  max mem: 4840
+train: [14]  [380/400]  eta: 0:00:07  lr: 0.000076  loss: 0.1652 (0.1627)  grad: 0.0284 (0.0296)  time: 0.3844  data: 0.0043  max mem: 4840
+train: [14]  [399/400]  eta: 0:00:00  lr: 0.000075  loss: 0.1537 (0.1624)  grad: 0.0286 (0.0296)  time: 0.3720  data: 0.0044  max mem: 4840
+train: [14] Total time: 0:02:31 (0.3797 s / it)
+train: [14] Summary: lr: 0.000075  loss: 0.1537 (0.1624)  grad: 0.0286 (0.0296)
+eval (validation): [14]  [ 0/63]  eta: 0:03:09    time: 3.0006  data: 2.8056  max mem: 4840
+eval (validation): [14]  [20/63]  eta: 0:00:19    time: 0.3328  data: 0.0303  max mem: 4840
+eval (validation): [14]  [40/63]  eta: 0:00:09    time: 0.3375  data: 0.0032  max mem: 4840
+eval (validation): [14]  [60/63]  eta: 0:00:01    time: 0.3057  data: 0.0029  max mem: 4840
+eval (validation): [14]  [62/63]  eta: 0:00:00    time: 0.3100  data: 0.0033  max mem: 4840
+eval (validation): [14] Total time: 0:00:23 (0.3728 s / it)
+cv: [14]  best hparam: (0.61, 1.0) (021) ('021_lr6.1e-01_wd1.0e+00')  loss: 0.065  acc: 0.981  f1: 0.979
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [15]  [  0/400]  eta: 0:22:29  lr: nan  time: 3.3740  data: 3.0795  max mem: 4840
+train: [15]  [ 20/400]  eta: 0:03:17  lr: 0.000074  loss: 0.1667 (0.1695)  grad: 0.0281 (0.0290)  time: 0.3774  data: 0.0041  max mem: 4840
+train: [15]  [ 40/400]  eta: 0:02:38  lr: 0.000072  loss: 0.1667 (0.1658)  grad: 0.0289 (0.0298)  time: 0.3543  data: 0.0027  max mem: 4840
+train: [15]  [ 60/400]  eta: 0:02:17  lr: 0.000071  loss: 0.1606 (0.1640)  grad: 0.0294 (0.0296)  time: 0.3301  data: 0.0039  max mem: 4840
+train: [15]  [ 80/400]  eta: 0:02:03  lr: 0.000070  loss: 0.1458 (0.1623)  grad: 0.0274 (0.0294)  time: 0.3381  data: 0.0040  max mem: 4840
+train: [15]  [100/400]  eta: 0:01:53  lr: 0.000068  loss: 0.1458 (0.1584)  grad: 0.0281 (0.0292)  time: 0.3495  data: 0.0041  max mem: 4840
+train: [15]  [120/400]  eta: 0:01:44  lr: 0.000067  loss: 0.1588 (0.1607)  grad: 0.0281 (0.0293)  time: 0.3397  data: 0.0040  max mem: 4840
+train: [15]  [140/400]  eta: 0:01:36  lr: 0.000066  loss: 0.1609 (0.1608)  grad: 0.0294 (0.0293)  time: 0.3519  data: 0.0043  max mem: 4840
+train: [15]  [160/400]  eta: 0:01:28  lr: 0.000064  loss: 0.1483 (0.1591)  grad: 0.0294 (0.0293)  time: 0.3645  data: 0.0038  max mem: 4840
+train: [15]  [180/400]  eta: 0:01:21  lr: 0.000063  loss: 0.1511 (0.1599)  grad: 0.0284 (0.0294)  time: 0.3756  data: 0.0041  max mem: 4840
+train: [15]  [200/400]  eta: 0:01:14  lr: 0.000062  loss: 0.1554 (0.1594)  grad: 0.0279 (0.0292)  time: 0.3690  data: 0.0042  max mem: 4840
+train: [15]  [220/400]  eta: 0:01:06  lr: 0.000061  loss: 0.1496 (0.1589)  grad: 0.0276 (0.0291)  time: 0.3651  data: 0.0043  max mem: 4840
+train: [15]  [240/400]  eta: 0:00:59  lr: 0.000059  loss: 0.1416 (0.1574)  grad: 0.0296 (0.0293)  time: 0.3677  data: 0.0043  max mem: 4840
+train: [15]  [260/400]  eta: 0:00:51  lr: 0.000058  loss: 0.1419 (0.1572)  grad: 0.0291 (0.0293)  time: 0.3791  data: 0.0042  max mem: 4840
+train: [15]  [280/400]  eta: 0:00:44  lr: 0.000057  loss: 0.1499 (0.1570)  grad: 0.0274 (0.0291)  time: 0.3741  data: 0.0045  max mem: 4840
+train: [15]  [300/400]  eta: 0:00:38  lr: 0.000056  loss: 0.1517 (0.1566)  grad: 0.0272 (0.0290)  time: 0.5477  data: 0.1822  max mem: 4840
+train: [15]  [320/400]  eta: 0:00:30  lr: 0.000054  loss: 0.1454 (0.1560)  grad: 0.0270 (0.0289)  time: 0.4231  data: 0.0040  max mem: 4840
+train: [15]  [340/400]  eta: 0:00:23  lr: 0.000053  loss: 0.1539 (0.1566)  grad: 0.0273 (0.0289)  time: 0.3715  data: 0.0041  max mem: 4840
+train: [15]  [360/400]  eta: 0:00:15  lr: 0.000052  loss: 0.1633 (0.1573)  grad: 0.0273 (0.0289)  time: 0.3673  data: 0.0043  max mem: 4840
+train: [15]  [380/400]  eta: 0:00:07  lr: 0.000051  loss: 0.1643 (0.1578)  grad: 0.0282 (0.0289)  time: 0.3648  data: 0.0043  max mem: 4840
+train: [15]  [399/400]  eta: 0:00:00  lr: 0.000050  loss: 0.1512 (0.1573)  grad: 0.0282 (0.0289)  time: 0.3709  data: 0.0043  max mem: 4840
+train: [15] Total time: 0:02:32 (0.3821 s / it)
+train: [15] Summary: lr: 0.000050  loss: 0.1512 (0.1573)  grad: 0.0282 (0.0289)
+eval (validation): [15]  [ 0/63]  eta: 0:03:07    time: 2.9818  data: 2.7785  max mem: 4840
+eval (validation): [15]  [20/63]  eta: 0:00:18    time: 0.3148  data: 0.0032  max mem: 4840
+eval (validation): [15]  [40/63]  eta: 0:00:08    time: 0.3369  data: 0.0037  max mem: 4840
+eval (validation): [15]  [60/63]  eta: 0:00:01    time: 0.2993  data: 0.0026  max mem: 4840
+eval (validation): [15]  [62/63]  eta: 0:00:00    time: 0.2928  data: 0.0029  max mem: 4840
+eval (validation): [15] Total time: 0:00:22 (0.3632 s / it)
+cv: [15]  best hparam: (0.52, 1.0) (020) ('020_lr5.2e-01_wd1.0e+00')  loss: 0.066  acc: 0.981  f1: 0.979
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [16]  [  0/400]  eta: 0:24:13  lr: nan  time: 3.6339  data: 3.3106  max mem: 4840
+train: [16]  [ 20/400]  eta: 0:03:43  lr: 0.000048  loss: 0.1512 (0.1607)  grad: 0.0290 (0.0308)  time: 0.4353  data: 0.0036  max mem: 4840
+train: [16]  [ 40/400]  eta: 0:02:51  lr: 0.000047  loss: 0.1549 (0.1567)  grad: 0.0287 (0.0294)  time: 0.3588  data: 0.0041  max mem: 4840
+train: [16]  [ 60/400]  eta: 0:02:29  lr: 0.000046  loss: 0.1584 (0.1574)  grad: 0.0275 (0.0289)  time: 0.3653  data: 0.0044  max mem: 4840
+train: [16]  [ 80/400]  eta: 0:02:15  lr: 0.000045  loss: 0.1578 (0.1553)  grad: 0.0275 (0.0286)  time: 0.3687  data: 0.0039  max mem: 4840
+train: [16]  [100/400]  eta: 0:02:02  lr: 0.000044  loss: 0.1397 (0.1517)  grad: 0.0276 (0.0286)  time: 0.3534  data: 0.0040  max mem: 4840
+train: [16]  [120/400]  eta: 0:01:50  lr: 0.000043  loss: 0.1368 (0.1526)  grad: 0.0282 (0.0286)  time: 0.3311  data: 0.0040  max mem: 4840
+train: [16]  [140/400]  eta: 0:01:41  lr: 0.000042  loss: 0.1581 (0.1534)  grad: 0.0287 (0.0288)  time: 0.3501  data: 0.0043  max mem: 4840
+train: [16]  [160/400]  eta: 0:01:33  lr: 0.000041  loss: 0.1541 (0.1527)  grad: 0.0280 (0.0286)  time: 0.3782  data: 0.0041  max mem: 4840
+train: [16]  [180/400]  eta: 0:01:24  lr: 0.000040  loss: 0.1526 (0.1536)  grad: 0.0285 (0.0288)  time: 0.3626  data: 0.0040  max mem: 4840
+train: [16]  [200/400]  eta: 0:01:16  lr: 0.000039  loss: 0.1501 (0.1536)  grad: 0.0280 (0.0289)  time: 0.3454  data: 0.0037  max mem: 4840
+train: [16]  [220/400]  eta: 0:01:07  lr: 0.000038  loss: 0.1501 (0.1538)  grad: 0.0279 (0.0288)  time: 0.3397  data: 0.0041  max mem: 4840
+train: [16]  [240/400]  eta: 0:00:59  lr: 0.000036  loss: 0.1542 (0.1547)  grad: 0.0281 (0.0290)  time: 0.3398  data: 0.0043  max mem: 4840
+train: [16]  [260/400]  eta: 0:00:52  lr: 0.000035  loss: 0.1536 (0.1543)  grad: 0.0273 (0.0288)  time: 0.3534  data: 0.0043  max mem: 4840
+train: [16]  [280/400]  eta: 0:00:44  lr: 0.000034  loss: 0.1509 (0.1541)  grad: 0.0272 (0.0289)  time: 0.3738  data: 0.0043  max mem: 4840
+train: [16]  [300/400]  eta: 0:00:38  lr: 0.000033  loss: 0.1482 (0.1539)  grad: 0.0289 (0.0289)  time: 0.5404  data: 0.1848  max mem: 4840
+train: [16]  [320/400]  eta: 0:00:30  lr: 0.000032  loss: 0.1574 (0.1544)  grad: 0.0289 (0.0289)  time: 0.3787  data: 0.0040  max mem: 4840
+train: [16]  [340/400]  eta: 0:00:22  lr: 0.000031  loss: 0.1440 (0.1534)  grad: 0.0277 (0.0288)  time: 0.3637  data: 0.0040  max mem: 4840
+train: [16]  [360/400]  eta: 0:00:15  lr: 0.000031  loss: 0.1484 (0.1545)  grad: 0.0279 (0.0288)  time: 0.3666  data: 0.0042  max mem: 4840
+train: [16]  [380/400]  eta: 0:00:07  lr: 0.000030  loss: 0.1524 (0.1539)  grad: 0.0279 (0.0287)  time: 0.3567  data: 0.0043  max mem: 4840
+train: [16]  [399/400]  eta: 0:00:00  lr: 0.000029  loss: 0.1492 (0.1540)  grad: 0.0279 (0.0288)  time: 0.3763  data: 0.0041  max mem: 4840
+train: [16] Total time: 0:02:32 (0.3807 s / it)
+train: [16] Summary: lr: 0.000029  loss: 0.1492 (0.1540)  grad: 0.0279 (0.0288)
+eval (validation): [16]  [ 0/63]  eta: 0:03:33    time: 3.3902  data: 3.0855  max mem: 4840
+eval (validation): [16]  [20/63]  eta: 0:00:21    time: 0.3484  data: 0.0044  max mem: 4840
+eval (validation): [16]  [40/63]  eta: 0:00:09    time: 0.3416  data: 0.0031  max mem: 4840
+eval (validation): [16]  [60/63]  eta: 0:00:01    time: 0.3040  data: 0.0039  max mem: 4840
+eval (validation): [16]  [62/63]  eta: 0:00:00    time: 0.3032  data: 0.0040  max mem: 4840
+eval (validation): [16] Total time: 0:00:24 (0.3841 s / it)
+cv: [16]  best hparam: (0.44, 1.0) (019) ('019_lr4.4e-01_wd1.0e+00')  loss: 0.069  acc: 0.981  f1: 0.979
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [17]  [  0/400]  eta: 0:22:35  lr: nan  time: 3.3896  data: 3.1371  max mem: 4840
+train: [17]  [ 20/400]  eta: 0:03:36  lr: 0.000028  loss: 0.1444 (0.1514)  grad: 0.0262 (0.0272)  time: 0.4285  data: 0.0035  max mem: 4840
+train: [17]  [ 40/400]  eta: 0:02:49  lr: 0.000027  loss: 0.1444 (0.1489)  grad: 0.0277 (0.0277)  time: 0.3675  data: 0.0038  max mem: 4840
+train: [17]  [ 60/400]  eta: 0:02:26  lr: 0.000026  loss: 0.1322 (0.1481)  grad: 0.0275 (0.0280)  time: 0.3515  data: 0.0043  max mem: 4840
+train: [17]  [ 80/400]  eta: 0:02:13  lr: 0.000025  loss: 0.1520 (0.1502)  grad: 0.0278 (0.0281)  time: 0.3686  data: 0.0044  max mem: 4840
+train: [17]  [100/400]  eta: 0:02:02  lr: 0.000024  loss: 0.1557 (0.1514)  grad: 0.0284 (0.0283)  time: 0.3716  data: 0.0040  max mem: 4840
+train: [17]  [120/400]  eta: 0:01:51  lr: 0.000023  loss: 0.1536 (0.1519)  grad: 0.0275 (0.0281)  time: 0.3476  data: 0.0041  max mem: 4840
+train: [17]  [140/400]  eta: 0:01:42  lr: 0.000023  loss: 0.1498 (0.1520)  grad: 0.0276 (0.0281)  time: 0.3710  data: 0.0043  max mem: 4840
+train: [17]  [160/400]  eta: 0:01:33  lr: 0.000022  loss: 0.1451 (0.1511)  grad: 0.0284 (0.0281)  time: 0.3628  data: 0.0045  max mem: 4840
+train: [17]  [180/400]  eta: 0:01:25  lr: 0.000021  loss: 0.1451 (0.1508)  grad: 0.0272 (0.0280)  time: 0.3757  data: 0.0042  max mem: 4840
+train: [17]  [200/400]  eta: 0:01:17  lr: 0.000020  loss: 0.1410 (0.1494)  grad: 0.0273 (0.0281)  time: 0.3680  data: 0.0040  max mem: 4840
+train: [17]  [220/400]  eta: 0:01:08  lr: 0.000019  loss: 0.1450 (0.1503)  grad: 0.0287 (0.0282)  time: 0.3499  data: 0.0039  max mem: 4840
+train: [17]  [240/400]  eta: 0:01:00  lr: 0.000019  loss: 0.1517 (0.1498)  grad: 0.0293 (0.0281)  time: 0.3589  data: 0.0041  max mem: 4840
+train: [17]  [260/400]  eta: 0:00:53  lr: 0.000018  loss: 0.1482 (0.1505)  grad: 0.0279 (0.0282)  time: 0.3598  data: 0.0040  max mem: 4840
+train: [17]  [280/400]  eta: 0:00:45  lr: 0.000017  loss: 0.1482 (0.1505)  grad: 0.0280 (0.0283)  time: 0.3651  data: 0.0043  max mem: 4840
+train: [17]  [300/400]  eta: 0:00:39  lr: 0.000016  loss: 0.1502 (0.1507)  grad: 0.0278 (0.0282)  time: 0.5578  data: 0.1904  max mem: 4840
+train: [17]  [320/400]  eta: 0:00:31  lr: 0.000016  loss: 0.1502 (0.1506)  grad: 0.0266 (0.0281)  time: 0.3578  data: 0.0032  max mem: 4840
+train: [17]  [340/400]  eta: 0:00:23  lr: 0.000015  loss: 0.1420 (0.1502)  grad: 0.0276 (0.0282)  time: 0.3554  data: 0.0034  max mem: 4840
+train: [17]  [360/400]  eta: 0:00:15  lr: 0.000014  loss: 0.1413 (0.1501)  grad: 0.0280 (0.0282)  time: 0.3456  data: 0.0044  max mem: 4840
+train: [17]  [380/400]  eta: 0:00:07  lr: 0.000014  loss: 0.1515 (0.1508)  grad: 0.0280 (0.0282)  time: 0.3595  data: 0.0043  max mem: 4840
+train: [17]  [399/400]  eta: 0:00:00  lr: 0.000013  loss: 0.1575 (0.1510)  grad: 0.0278 (0.0282)  time: 0.3497  data: 0.0040  max mem: 4840
+train: [17] Total time: 0:02:32 (0.3817 s / it)
+train: [17] Summary: lr: 0.000013  loss: 0.1575 (0.1510)  grad: 0.0278 (0.0282)
+eval (validation): [17]  [ 0/63]  eta: 0:03:32    time: 3.3742  data: 3.0932  max mem: 4840
+eval (validation): [17]  [20/63]  eta: 0:00:21    time: 0.3610  data: 0.0044  max mem: 4840
+eval (validation): [17]  [40/63]  eta: 0:00:09    time: 0.3248  data: 0.0033  max mem: 4840
+eval (validation): [17]  [60/63]  eta: 0:00:01    time: 0.3213  data: 0.0036  max mem: 4840
+eval (validation): [17]  [62/63]  eta: 0:00:00    time: 0.3125  data: 0.0036  max mem: 4840
+eval (validation): [17] Total time: 0:00:24 (0.3868 s / it)
+cv: [17]  best hparam: (1, 1.0) (024) ('024_lr1.0e+00_wd1.0e+00')  loss: 0.063  acc: 0.981  f1: 0.979
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [18]  [  0/400]  eta: 0:22:45  lr: nan  time: 3.4142  data: 3.1027  max mem: 4840
+train: [18]  [ 20/400]  eta: 0:03:21  lr: 0.000012  loss: 0.1561 (0.1573)  grad: 0.0280 (0.0288)  time: 0.3863  data: 0.0042  max mem: 4840
+train: [18]  [ 40/400]  eta: 0:02:41  lr: 0.000012  loss: 0.1540 (0.1565)  grad: 0.0282 (0.0289)  time: 0.3624  data: 0.0037  max mem: 4840
+train: [18]  [ 60/400]  eta: 0:02:22  lr: 0.000011  loss: 0.1527 (0.1536)  grad: 0.0280 (0.0286)  time: 0.3632  data: 0.0042  max mem: 4840
+train: [18]  [ 80/400]  eta: 0:02:10  lr: 0.000011  loss: 0.1482 (0.1519)  grad: 0.0271 (0.0284)  time: 0.3635  data: 0.0044  max mem: 4840
+train: [18]  [100/400]  eta: 0:01:59  lr: 0.000010  loss: 0.1434 (0.1503)  grad: 0.0277 (0.0282)  time: 0.3614  data: 0.0044  max mem: 4840
+train: [18]  [120/400]  eta: 0:01:49  lr: 0.000009  loss: 0.1418 (0.1498)  grad: 0.0272 (0.0282)  time: 0.3574  data: 0.0043  max mem: 4840
+train: [18]  [140/400]  eta: 0:01:39  lr: 0.000009  loss: 0.1469 (0.1506)  grad: 0.0283 (0.0285)  time: 0.3395  data: 0.0043  max mem: 4840
+train: [18]  [160/400]  eta: 0:01:31  lr: 0.000008  loss: 0.1491 (0.1510)  grad: 0.0274 (0.0283)  time: 0.3573  data: 0.0045  max mem: 4840
+train: [18]  [180/400]  eta: 0:01:23  lr: 0.000008  loss: 0.1428 (0.1498)  grad: 0.0271 (0.0283)  time: 0.3813  data: 0.0040  max mem: 4840
+train: [18]  [200/400]  eta: 0:01:15  lr: 0.000007  loss: 0.1359 (0.1486)  grad: 0.0275 (0.0282)  time: 0.3633  data: 0.0043  max mem: 4840
+train: [18]  [220/400]  eta: 0:01:08  lr: 0.000007  loss: 0.1359 (0.1489)  grad: 0.0280 (0.0283)  time: 0.3704  data: 0.0041  max mem: 4840
+train: [18]  [240/400]  eta: 0:01:00  lr: 0.000006  loss: 0.1496 (0.1493)  grad: 0.0278 (0.0283)  time: 0.3658  data: 0.0044  max mem: 4840
+train: [18]  [260/400]  eta: 0:00:52  lr: 0.000006  loss: 0.1496 (0.1493)  grad: 0.0274 (0.0283)  time: 0.3522  data: 0.0042  max mem: 4840
+train: [18]  [280/400]  eta: 0:00:44  lr: 0.000006  loss: 0.1427 (0.1484)  grad: 0.0274 (0.0282)  time: 0.3594  data: 0.0044  max mem: 4840
+train: [18]  [300/400]  eta: 0:00:38  lr: 0.000005  loss: 0.1436 (0.1482)  grad: 0.0293 (0.0284)  time: 0.5491  data: 0.1856  max mem: 4840
+train: [18]  [320/400]  eta: 0:00:30  lr: 0.000005  loss: 0.1464 (0.1482)  grad: 0.0291 (0.0283)  time: 0.3977  data: 0.0039  max mem: 4840
+train: [18]  [340/400]  eta: 0:00:23  lr: 0.000004  loss: 0.1554 (0.1489)  grad: 0.0279 (0.0284)  time: 0.3613  data: 0.0043  max mem: 4840
+train: [18]  [360/400]  eta: 0:00:15  lr: 0.000004  loss: 0.1562 (0.1491)  grad: 0.0287 (0.0284)  time: 0.3614  data: 0.0044  max mem: 4840
+train: [18]  [380/400]  eta: 0:00:07  lr: 0.000004  loss: 0.1455 (0.1488)  grad: 0.0284 (0.0284)  time: 0.3605  data: 0.0044  max mem: 4840
+train: [18]  [399/400]  eta: 0:00:00  lr: 0.000003  loss: 0.1454 (0.1490)  grad: 0.0274 (0.0283)  time: 0.3534  data: 0.0039  max mem: 4840
+train: [18] Total time: 0:02:32 (0.3815 s / it)
+train: [18] Summary: lr: 0.000003  loss: 0.1454 (0.1490)  grad: 0.0274 (0.0283)
+eval (validation): [18]  [ 0/63]  eta: 0:03:26    time: 3.2789  data: 3.0582  max mem: 4840
+eval (validation): [18]  [20/63]  eta: 0:00:20    time: 0.3336  data: 0.0183  max mem: 4840
+eval (validation): [18]  [40/63]  eta: 0:00:09    time: 0.3304  data: 0.0068  max mem: 4840
+eval (validation): [18]  [60/63]  eta: 0:00:01    time: 0.3074  data: 0.0029  max mem: 4840
+eval (validation): [18]  [62/63]  eta: 0:00:00    time: 0.3048  data: 0.0029  max mem: 4840
+eval (validation): [18] Total time: 0:00:23 (0.3749 s / it)
+cv: [18]  best hparam: (0.85, 1.0) (023) ('023_lr8.5e-01_wd1.0e+00')  loss: 0.063  acc: 0.981  f1: 0.978
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+train: [19]  [  0/400]  eta: 0:24:04  lr: nan  time: 3.6123  data: 3.3470  max mem: 4840
+train: [19]  [ 20/400]  eta: 0:03:09  lr: 0.000003  loss: 0.1507 (0.1509)  grad: 0.0276 (0.0277)  time: 0.3423  data: 0.0036  max mem: 4840
+train: [19]  [ 40/400]  eta: 0:02:35  lr: 0.000003  loss: 0.1428 (0.1439)  grad: 0.0269 (0.0277)  time: 0.3638  data: 0.0035  max mem: 4840
+train: [19]  [ 60/400]  eta: 0:02:18  lr: 0.000002  loss: 0.1407 (0.1431)  grad: 0.0269 (0.0281)  time: 0.3537  data: 0.0031  max mem: 4840
+train: [19]  [ 80/400]  eta: 0:02:05  lr: 0.000002  loss: 0.1460 (0.1456)  grad: 0.0275 (0.0280)  time: 0.3478  data: 0.0041  max mem: 4840
+train: [19]  [100/400]  eta: 0:01:56  lr: 0.000002  loss: 0.1535 (0.1461)  grad: 0.0275 (0.0281)  time: 0.3771  data: 0.0042  max mem: 4840
+train: [19]  [120/400]  eta: 0:01:48  lr: 0.000002  loss: 0.1554 (0.1477)  grad: 0.0279 (0.0281)  time: 0.3791  data: 0.0042  max mem: 4840
+train: [19]  [140/400]  eta: 0:01:39  lr: 0.000001  loss: 0.1560 (0.1496)  grad: 0.0283 (0.0281)  time: 0.3469  data: 0.0039  max mem: 4840
+train: [19]  [160/400]  eta: 0:01:30  lr: 0.000001  loss: 0.1481 (0.1499)  grad: 0.0283 (0.0283)  time: 0.3493  data: 0.0043  max mem: 4840
+train: [19]  [180/400]  eta: 0:01:22  lr: 0.000001  loss: 0.1473 (0.1503)  grad: 0.0275 (0.0281)  time: 0.3663  data: 0.0044  max mem: 4840
+train: [19]  [200/400]  eta: 0:01:15  lr: 0.000001  loss: 0.1525 (0.1511)  grad: 0.0270 (0.0281)  time: 0.3809  data: 0.0043  max mem: 4840
+train: [19]  [220/400]  eta: 0:01:07  lr: 0.000001  loss: 0.1530 (0.1524)  grad: 0.0271 (0.0281)  time: 0.3563  data: 0.0039  max mem: 4840
+train: [19]  [240/400]  eta: 0:00:59  lr: 0.000001  loss: 0.1582 (0.1533)  grad: 0.0274 (0.0282)  time: 0.3571  data: 0.0040  max mem: 4840
+train: [19]  [260/400]  eta: 0:00:52  lr: 0.000000  loss: 0.1488 (0.1528)  grad: 0.0283 (0.0282)  time: 0.3881  data: 0.0042  max mem: 4840
+train: [19]  [280/400]  eta: 0:00:44  lr: 0.000000  loss: 0.1419 (0.1526)  grad: 0.0279 (0.0282)  time: 0.3752  data: 0.0041  max mem: 4840
+train: [19]  [300/400]  eta: 0:00:38  lr: 0.000000  loss: 0.1481 (0.1533)  grad: 0.0269 (0.0282)  time: 0.5551  data: 0.1854  max mem: 4840
+train: [19]  [320/400]  eta: 0:00:30  lr: 0.000000  loss: 0.1572 (0.1537)  grad: 0.0271 (0.0282)  time: 0.3716  data: 0.0034  max mem: 4840
+train: [19]  [340/400]  eta: 0:00:23  lr: 0.000000  loss: 0.1565 (0.1536)  grad: 0.0274 (0.0282)  time: 0.3707  data: 0.0041  max mem: 4840
+train: [19]  [360/400]  eta: 0:00:15  lr: 0.000000  loss: 0.1600 (0.1536)  grad: 0.0271 (0.0282)  time: 0.3681  data: 0.0042  max mem: 4840
+train: [19]  [380/400]  eta: 0:00:07  lr: 0.000000  loss: 0.1594 (0.1541)  grad: 0.0278 (0.0283)  time: 0.3515  data: 0.0040  max mem: 4840
+train: [19]  [399/400]  eta: 0:00:00  lr: 0.000000  loss: 0.1579 (0.1544)  grad: 0.0278 (0.0283)  time: 0.3727  data: 0.0044  max mem: 4840
+train: [19] Total time: 0:02:32 (0.3823 s / it)
+train: [19] Summary: lr: 0.000000  loss: 0.1579 (0.1544)  grad: 0.0278 (0.0283)
+eval (validation): [19]  [ 0/63]  eta: 0:03:29    time: 3.3249  data: 3.1005  max mem: 4840
+eval (validation): [19]  [20/63]  eta: 0:00:19    time: 0.3208  data: 0.0049  max mem: 4840
+eval (validation): [19]  [40/63]  eta: 0:00:09    time: 0.3273  data: 0.0023  max mem: 4840
+eval (validation): [19]  [60/63]  eta: 0:00:01    time: 0.3213  data: 0.0033  max mem: 4840
+eval (validation): [19]  [62/63]  eta: 0:00:00    time: 0.3191  data: 0.0032  max mem: 4840
+eval (validation): [19] Total time: 0:00:23 (0.3757 s / it)
+cv: [19]  best hparam: (0.85, 1.0) (023) ('023_lr8.5e-01_wd1.0e+00')  loss: 0.063  acc: 0.981  f1: 0.978
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+evaluating last checkpoint: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-last.pth
+eval model info:
+{"score": 0.9809027777777778, "hparam": [0.85, 1.0], "hparam_id": 23, "epoch": 19, "is_best": false, "best_score": 0.9818948412698413}
+eval (train): [20]  [  0/297]  eta: 0:13:45    time: 2.7781  data: 2.5479  max mem: 4840
+eval (train): [20]  [ 20/297]  eta: 0:01:57    time: 0.3056  data: 0.0026  max mem: 4840
+eval (train): [20]  [ 40/297]  eta: 0:01:40    time: 0.3540  data: 0.0040  max mem: 4840
+eval (train): [20]  [ 60/297]  eta: 0:01:31    time: 0.3802  data: 0.0041  max mem: 4840
+eval (train): [20]  [ 80/297]  eta: 0:01:21    time: 0.3434  data: 0.0036  max mem: 4840
+eval (train): [20]  [100/297]  eta: 0:01:12    time: 0.3366  data: 0.0034  max mem: 4840
+eval (train): [20]  [120/297]  eta: 0:01:04    time: 0.3381  data: 0.0034  max mem: 4840
+eval (train): [20]  [140/297]  eta: 0:00:56    time: 0.3362  data: 0.0034  max mem: 4840
+eval (train): [20]  [160/297]  eta: 0:00:48    time: 0.3420  data: 0.0035  max mem: 4840
+eval (train): [20]  [180/297]  eta: 0:00:41    time: 0.3591  data: 0.0037  max mem: 4840
+eval (train): [20]  [200/297]  eta: 0:00:34    time: 0.3441  data: 0.0033  max mem: 4840
+eval (train): [20]  [220/297]  eta: 0:00:27    time: 0.3522  data: 0.0037  max mem: 4840
+eval (train): [20]  [240/297]  eta: 0:00:20    time: 0.3545  data: 0.0034  max mem: 4840
+eval (train): [20]  [260/297]  eta: 0:00:13    time: 0.3387  data: 0.0037  max mem: 4840
+eval (train): [20]  [280/297]  eta: 0:00:06    time: 0.3649  data: 0.0038  max mem: 4840
+eval (train): [20]  [296/297]  eta: 0:00:00    time: 0.3369  data: 0.0037  max mem: 4840
+eval (train): [20] Total time: 0:01:45 (0.3545 s / it)
+eval (validation): [20]  [ 0/63]  eta: 0:03:17    time: 3.1380  data: 2.8735  max mem: 4840
+eval (validation): [20]  [20/63]  eta: 0:00:20    time: 0.3492  data: 0.0044  max mem: 4840
+eval (validation): [20]  [40/63]  eta: 0:00:09    time: 0.3500  data: 0.0034  max mem: 4840
+eval (validation): [20]  [60/63]  eta: 0:00:01    time: 0.3266  data: 0.0031  max mem: 4840
+eval (validation): [20]  [62/63]  eta: 0:00:00    time: 0.3150  data: 0.0032  max mem: 4840
+eval (validation): [20] Total time: 0:00:24 (0.3891 s / it)
+eval (test): [20]  [ 0/79]  eta: 0:04:08    time: 3.1514  data: 2.8668  max mem: 4840
+eval (test): [20]  [20/79]  eta: 0:00:30    time: 0.3868  data: 0.0043  max mem: 4840
+eval (test): [20]  [40/79]  eta: 0:00:17    time: 0.3575  data: 0.0034  max mem: 4840
+eval (test): [20]  [60/79]  eta: 0:00:07    time: 0.3336  data: 0.0037  max mem: 4840
+eval (test): [20]  [78/79]  eta: 0:00:00    time: 0.2950  data: 0.0030  max mem: 4840
+eval (test): [20] Total time: 0:00:30 (0.3832 s / it)
+evaluating best checkpoint: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__attn/checkpoint-best.pth
+eval model info:
+{"score": 0.9818948412698413, "hparam": [0.85, 1.0], "hparam_id": 23, "epoch": 8, "is_best": true, "best_score": 0.9818948412698413}
+eval (train): [20]  [  0/297]  eta: 0:15:37    time: 3.1577  data: 2.8724  max mem: 4840
+eval (train): [20]  [ 20/297]  eta: 0:02:21    time: 0.3779  data: 0.0046  max mem: 4840
+eval (train): [20]  [ 40/297]  eta: 0:01:48    time: 0.3306  data: 0.0030  max mem: 4840
+eval (train): [20]  [ 60/297]  eta: 0:01:31    time: 0.3149  data: 0.0032  max mem: 4840
+eval (train): [20]  [ 80/297]  eta: 0:01:20    time: 0.3213  data: 0.0035  max mem: 4840
+eval (train): [20]  [100/297]  eta: 0:01:13    time: 0.3824  data: 0.0039  max mem: 4840
+eval (train): [20]  [120/297]  eta: 0:01:05    time: 0.3626  data: 0.0037  max mem: 4840
+eval (train): [20]  [140/297]  eta: 0:00:58    time: 0.3713  data: 0.0037  max mem: 4840
+eval (train): [20]  [160/297]  eta: 0:00:50    time: 0.3432  data: 0.0038  max mem: 4840
+eval (train): [20]  [180/297]  eta: 0:00:42    time: 0.3187  data: 0.0033  max mem: 4840
+eval (train): [20]  [200/297]  eta: 0:00:35    time: 0.3490  data: 0.0039  max mem: 4840
+eval (train): [20]  [220/297]  eta: 0:00:28    time: 0.3907  data: 0.0040  max mem: 4840
+eval (train): [20]  [240/297]  eta: 0:00:20    time: 0.3616  data: 0.0034  max mem: 4840
+eval (train): [20]  [260/297]  eta: 0:00:13    time: 0.3412  data: 0.0037  max mem: 4840
+eval (train): [20]  [280/297]  eta: 0:00:06    time: 0.3214  data: 0.0031  max mem: 4840
+eval (train): [20]  [296/297]  eta: 0:00:00    time: 0.2961  data: 0.0033  max mem: 4840
+eval (train): [20] Total time: 0:01:46 (0.3579 s / it)
+eval (validation): [20]  [ 0/63]  eta: 0:03:13    time: 3.0700  data: 2.8401  max mem: 4840
+eval (validation): [20]  [20/63]  eta: 0:00:21    time: 0.3734  data: 0.0037  max mem: 4840
+eval (validation): [20]  [40/63]  eta: 0:00:10    time: 0.3928  data: 0.0033  max mem: 4840
+eval (validation): [20]  [60/63]  eta: 0:00:01    time: 0.3164  data: 0.0030  max mem: 4840
+eval (validation): [20]  [62/63]  eta: 0:00:00    time: 0.3156  data: 0.0032  max mem: 4840
+eval (validation): [20] Total time: 0:00:25 (0.4062 s / it)
+eval (test): [20]  [ 0/79]  eta: 0:04:06    time: 3.1213  data: 2.8250  max mem: 4840
+eval (test): [20]  [20/79]  eta: 0:00:28    time: 0.3435  data: 0.0037  max mem: 4840
+eval (test): [20]  [40/79]  eta: 0:00:16    time: 0.3537  data: 0.0032  max mem: 4840
+eval (test): [20]  [60/79]  eta: 0:00:07    time: 0.3345  data: 0.0030  max mem: 4840
+eval (test): [20]  [78/79]  eta: 0:00:00    time: 0.3188  data: 0.0034  max mem: 4840
+eval (test): [20] Total time: 0:00:29 (0.3757 s / it)
+eval results:
+
+| model    | repr   | clf   | dataset      | ckpt   |   epoch |       lr |   wd |   hparam_id | hparam      | split      |     loss |     acc |   acc_std |      f1 |     f1_std |
+|:---------|:-------|:------|:-------------|:-------|--------:|---------:|-----:|------------:|:------------|:-----------|---------:|--------:|----------:|--------:|-----------:|
+| flat_mae | reg    | attn  | hcpya_task21 | best   |       8 | 0.000255 | 0.05 |          23 | [0.85, 1.0] | train      | 0.035422 | 0.99384 | 0.0005515 | 0.99363 | 0.00062565 |
+| flat_mae | reg    | attn  | hcpya_task21 | best   |       8 | 0.000255 | 0.05 |          23 | [0.85, 1.0] | validation | 0.067142 | 0.98189 | 0.0019292 | 0.97991 | 0.0024497  |
+| flat_mae | reg    | attn  | hcpya_task21 | best   |       8 | 0.000255 | 0.05 |          23 | [0.85, 1.0] | test       | 0.086062 | 0.97421 | 0.0022555 | 0.96839 | 0.0030264  |
+
+
+done! total time: 1:06:05

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/config.yaml

@@ -0,0 +1,96 @@
+output_root: experiments/decoders/output
+name_prefix: eval_probe
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (hcpya_task21 reg linear)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: false
+  norm: false
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear
+model: flat_mae
+representation: reg
+classifier: linear
+dataset: hcpya_task21
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear
+remote_dir: null

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/eval_log.json

@@ -0,0 +1 @@
+{"eval/epoch": 16, "eval/id_best": 47, "eval/lr_best": 0.012899999999999998, "eval/wd_best": 0.05, "eval/train/loss": 0.045580506324768066, "eval/train/acc": 0.9932628033054371, "eval/train/acc_std": 0.0005759690612889856, "eval/train/f1": 0.9945752581741296, "eval/train/f1_std": 0.0004984929288620381, "eval/validation/loss": 0.11011836677789688, "eval/validation/acc": 0.9692460317460317, "eval/validation/acc_std": 0.002672485185662656, "eval/validation/f1": 0.9634397574434674, "eval/validation/f1_std": 0.0036045798428122238, "eval/test/loss": 0.13361142575740814, "eval/test/acc": 0.9626984126984127, "eval/test/acc_std": 0.002597767655873538, "eval/test/f1": 0.9567782414613959, "eval/test/f1_std": 0.0034142196925447286}

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/eval_log_best.json

@@ -0,0 +1 @@
+{"eval/best/epoch": 16, "eval/best/id_best": 47, "eval/best/lr_best": 0.012899999999999998, "eval/best/wd_best": 0.05, "eval/best/train/loss": 0.045580506324768066, "eval/best/train/acc": 0.9932628033054371, "eval/best/train/acc_std": 0.0005759690612889856, "eval/best/train/f1": 0.9945752581741296, "eval/best/train/f1_std": 0.0004984929288620381, "eval/best/validation/loss": 0.11011836677789688, "eval/best/validation/acc": 0.9692460317460317, "eval/best/validation/acc_std": 0.002672485185662656, "eval/best/validation/f1": 0.9634397574434674, "eval/best/validation/f1_std": 0.0036045798428122238, "eval/best/test/loss": 0.13361142575740814, "eval/best/test/acc": 0.9626984126984127, "eval/best/test/acc_std": 0.002597767655873538, "eval/best/test/f1": 0.9567782414613959, "eval/best/test/f1_std": 0.0034142196925447286}

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/eval_log_last.json

@@ -0,0 +1 @@
+{"eval/last/epoch": 19, "eval/last/id_best": 43, "eval/last/lr_best": 0.006599999999999999, "eval/last/wd_best": 0.05, "eval/last/train/loss": 0.05881621316075325, "eval/last/train/acc": 0.990157376704037, "eval/last/train/acc_std": 0.0007111911502121867, "eval/last/train/f1": 0.9912159372513739, "eval/last/train/f1_std": 0.0007002959171880783, "eval/last/validation/loss": 0.11311419308185577, "eval/last/validation/acc": 0.9682539682539683, "eval/last/validation/acc_std": 0.0027225725311416115, "eval/last/validation/f1": 0.9626198768348221, "eval/last/validation/f1_std": 0.003632717506997901, "eval/last/test/loss": 0.13546596467494965, "eval/last/test/acc": 0.9625, "eval/last/test/acc_std": 0.002650503619719388, "eval/last/test/f1": 0.9569064705527082, "eval/last/test/f1_std": 0.0034210302434791212}

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/eval_table.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,reg,linear,hcpya_task21,best,16,0.012899999999999998,0.05,47,"[43, 1.0]",train,0.045580506324768066,0.9932628033054371,0.0005759690612889856,0.9945752581741296,0.0004984929288620381
+flat_mae,reg,linear,hcpya_task21,best,16,0.012899999999999998,0.05,47,"[43, 1.0]",validation,0.11011836677789688,0.9692460317460317,0.002672485185662656,0.9634397574434674,0.0036045798428122238
+flat_mae,reg,linear,hcpya_task21,best,16,0.012899999999999998,0.05,47,"[43, 1.0]",test,0.13361142575740814,0.9626984126984127,0.002597767655873538,0.9567782414613959,0.0034142196925447286

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/eval_table_best.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,reg,linear,hcpya_task21,best,16,0.012899999999999998,0.05,47,"[43, 1.0]",train,0.045580506324768066,0.9932628033054371,0.0005759690612889856,0.9945752581741296,0.0004984929288620381
+flat_mae,reg,linear,hcpya_task21,best,16,0.012899999999999998,0.05,47,"[43, 1.0]",validation,0.11011836677789688,0.9692460317460317,0.002672485185662656,0.9634397574434674,0.0036045798428122238
+flat_mae,reg,linear,hcpya_task21,best,16,0.012899999999999998,0.05,47,"[43, 1.0]",test,0.13361142575740814,0.9626984126984127,0.002597767655873538,0.9567782414613959,0.0034142196925447286

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/eval_table_last.csv

@@ -0,0 +1,4 @@
+model,repr,clf,dataset,ckpt,epoch,lr,wd,hparam_id,hparam,split,loss,acc,acc_std,f1,f1_std
+flat_mae,reg,linear,hcpya_task21,last,19,0.006599999999999999,0.05,43,"[22, 1.0]",train,0.05881621316075325,0.990157376704037,0.0007111911502121867,0.9912159372513739,0.0007002959171880783
+flat_mae,reg,linear,hcpya_task21,last,19,0.006599999999999999,0.05,43,"[22, 1.0]",validation,0.11311419308185577,0.9682539682539683,0.0027225725311416115,0.9626198768348221,0.003632717506997901
+flat_mae,reg,linear,hcpya_task21,last,19,0.006599999999999999,0.05,43,"[22, 1.0]",test,0.13546596467494965,0.9625,0.002650503619719388,0.9569064705527082,0.0034210302434791212

decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/log.txt

@@ -0,0 +1,887 @@
+fMRI foundation model probe eval
+version: 0.1.dev65+g4003a1397
+sha: 6c01b606db98add5848cecd23e5d599250c0bf86, status: clean, branch: dev/clane9
+cwd: /data/connor/fmri-fm
+start: 2026-02-24 21:17:48
+config:
+output_root: experiments/decoders/output
+name_prefix: eval_probe
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (hcpya_task21 reg linear)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: false
+  norm: false
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear
+model: flat_mae
+representation: reg
+classifier: linear
+dataset: hcpya_task21
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear
+remote_dir: null
+
+creating frozen backbone model: flat_mae
+backbone:
+MaskedEncoderWrapper(
+  (model): MaskedEncoder(
+    class_token=False, reg_tokens=4, no_embed_class=True, mask_drop_scale=False
+    (patchify): Patchify3D((16, 224, 560), (4, 16, 16), in_chans=1)
+    (patch_embed): Linear(in_features=1024, out_features=768, bias=True)
+    (pos_embed): SeparablePosEmbed(768, (4, 14, 35))
+    (blocks): ModuleList(
+      (0-11): 12 x Block(
+        (norm1): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (attn): Attention(
+          num_heads=12
+          (q): Linear(in_features=768, out_features=768, bias=True)
+          (k): Linear(in_features=768, out_features=768, bias=True)
+          (v): Linear(in_features=768, out_features=768, bias=True)
+          (proj): Linear(in_features=768, out_features=768, bias=True)
+        )
+        (drop_path1): Identity()
+        (norm2): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+        (mlp): Mlp(
+          (fc1): Linear(in_features=768, out_features=3072, bias=True)
+          (act): GELU(approximate='none')
+          (fc2): Linear(in_features=3072, out_features=768, bias=True)
+        )
+        (drop_path2): Identity()
+      )
+    )
+    (norm): LayerNorm((768,), eps=1e-06, elementwise_affine=True)
+  )
+)
+creating dataset: hcpya_task21 (flat)
+train (n=18999):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 18999
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[ 832 1248 3201 1660  832  832  832  832  832 1248 1247 1243  832  416
+  416  416  416  416  416  416  416]
+)
+
+validation (n=4032):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 4032
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[176 264 688 352 176 176 176 176 176 264 264 264 176  88  88  88  88  88
+  88  88  88]
+)
+
+test (n=5040):
+HFDataset(
+    dataset=Dataset({
+    features: ['sub', 'task', 'cond', 'cond_id', 'path', 'start', 'end', 'n_frames', 'tr', 'bold', 'mean', 'std'],
+    num_rows: 5040
+}),
+    labels=[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20],
+    counts=[220 330 860 440 220 220 220 220 220 330 330 330 220 110 110 110 110 110
+ 110 110 110]
+)
+
+running backbone on example batch to get embedding dim
+embedding feature dim (reg): 768
+initializing sweep of classifier heads
+classifiers:
+ModuleList(
+  (0-48): 49 x LinearClassifier(
+    (linear): Linear(in_features=768, out_features=21, bias=True)
+  )
+)
+classifier params (train): 0.8M (0.8M)
+setting up optimizer
+total batch size: 128 = 64 bs per gpu x 2 accum
+lr: 3.00e-04
+full schedule: epochs = 20 (steps = 4000) (decay = True)
+warmup: epochs = 5 (steps = 1000)
+start training for 20 epochs
+train: [0]  [  0/400]  eta: 0:21:38  lr: nan  time: 3.2469  data: 2.8724  max mem: 3914
+train: [0]  [ 20/400]  eta: 0:03:07  lr: 0.000003  loss: 3.0185 (3.0137)  grad: 0.1066 (0.1051)  time: 0.3554  data: 0.0041  max mem: 3956
+train: [0]  [ 40/400]  eta: 0:02:30  lr: 0.000006  loss: 3.0185 (3.0139)  grad: 0.1030 (0.1025)  time: 0.3416  data: 0.0041  max mem: 3956
+train: [0]  [ 60/400]  eta: 0:02:16  lr: 0.000009  loss: 3.0020 (3.0067)  grad: 0.1026 (0.1039)  time: 0.3649  data: 0.0038  max mem: 3956
+train: [0]  [ 80/400]  eta: 0:02:02  lr: 0.000012  loss: 2.9815 (2.9998)  grad: 0.1052 (0.1047)  time: 0.3298  data: 0.0037  max mem: 3956
+train: [0]  [100/400]  eta: 0:01:52  lr: 0.000015  loss: 2.9654 (2.9902)  grad: 0.1078 (0.1063)  time: 0.3374  data: 0.0041  max mem: 3956
+train: [0]  [120/400]  eta: 0:01:42  lr: 0.000018  loss: 2.9373 (2.9780)  grad: 0.1092 (0.1059)  time: 0.3320  data: 0.0041  max mem: 3956
+train: [0]  [140/400]  eta: 0:01:34  lr: 0.000021  loss: 2.9045 (2.9665)  grad: 0.1041 (0.1063)  time: 0.3331  data: 0.0041  max mem: 3956
+train: [0]  [160/400]  eta: 0:01:26  lr: 0.000024  loss: 2.8766 (2.9544)  grad: 0.1005 (0.1054)  time: 0.3315  data: 0.0041  max mem: 3956
+train: [0]  [180/400]  eta: 0:01:18  lr: 0.000027  loss: 2.8463 (2.9378)  grad: 0.0989 (0.1051)  time: 0.3422  data: 0.0042  max mem: 3956
+train: [0]  [200/400]  eta: 0:01:11  lr: 0.000030  loss: 2.7871 (2.9214)  grad: 0.0989 (0.1046)  time: 0.3639  data: 0.0041  max mem: 3956
+train: [0]  [220/400]  eta: 0:01:04  lr: 0.000033  loss: 2.7667 (2.9060)  grad: 0.0990 (0.1040)  time: 0.3361  data: 0.0042  max mem: 3956
+train: [0]  [240/400]  eta: 0:00:56  lr: 0.000036  loss: 2.7244 (2.8881)  grad: 0.0991 (0.1039)  time: 0.3230  data: 0.0039  max mem: 3956
+train: [0]  [260/400]  eta: 0:00:49  lr: 0.000039  loss: 2.6737 (2.8697)  grad: 0.1040 (0.1038)  time: 0.3627  data: 0.0045  max mem: 3956
+train: [0]  [280/400]  eta: 0:00:42  lr: 0.000042  loss: 2.6433 (2.8526)  grad: 0.0973 (0.1031)  time: 0.3384  data: 0.0044  max mem: 3956
+train: [0]  [300/400]  eta: 0:00:36  lr: 0.000045  loss: 2.6005 (2.8346)  grad: 0.0905 (0.1026)  time: 0.4930  data: 0.1794  max mem: 3956
+train: [0]  [320/400]  eta: 0:00:28  lr: 0.000048  loss: 2.5648 (2.8161)  grad: 0.0949 (0.1024)  time: 0.3390  data: 0.0032  max mem: 3956
+train: [0]  [340/400]  eta: 0:00:21  lr: 0.000051  loss: 2.5061 (2.7977)  grad: 0.1002 (0.1022)  time: 0.3271  data: 0.0033  max mem: 3956
+train: [0]  [360/400]  eta: 0:00:14  lr: 0.000054  loss: 2.4902 (2.7805)  grad: 0.0900 (0.1015)  time: 0.3226  data: 0.0036  max mem: 3956
+train: [0]  [380/400]  eta: 0:00:07  lr: 0.000057  loss: 2.4540 (2.7620)  grad: 0.0894 (0.1010)  time: 0.3411  data: 0.0039  max mem: 3956
+train: [0]  [399/400]  eta: 0:00:00  lr: 0.000060  loss: 2.4033 (2.7436)  grad: 0.0900 (0.1006)  time: 0.3335  data: 0.0038  max mem: 3956
+train: [0] Total time: 0:02:22 (0.3552 s / it)
+train: [0] Summary: lr: 0.000060  loss: 2.4033 (2.7436)  grad: 0.0900 (0.1006)
+eval (validation): [0]  [ 0/63]  eta: 0:03:21    time: 3.1955  data: 2.9930  max mem: 3956
+eval (validation): [0]  [20/63]  eta: 0:00:19    time: 0.3064  data: 0.0033  max mem: 3956
+eval (validation): [0]  [40/63]  eta: 0:00:09    time: 0.3578  data: 0.0038  max mem: 3956
+eval (validation): [0]  [60/63]  eta: 0:00:01    time: 0.3121  data: 0.0026  max mem: 3956
+eval (validation): [0]  [62/63]  eta: 0:00:00    time: 0.3054  data: 0.0028  max mem: 3956
+eval (validation): [0] Total time: 0:00:23 (0.3754 s / it)
+cv: [0]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.522  acc: 0.932  f1: 0.928
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-best.pth
+train: [1]  [  0/400]  eta: 0:20:53  lr: nan  time: 3.1338  data: 2.9243  max mem: 3956
+train: [1]  [ 20/400]  eta: 0:02:50  lr: 0.000063  loss: 2.3715 (2.3656)  grad: 0.0824 (0.0866)  time: 0.3151  data: 0.0033  max mem: 3956
+train: [1]  [ 40/400]  eta: 0:02:23  lr: 0.000066  loss: 2.3599 (2.3607)  grad: 0.0825 (0.0863)  time: 0.3429  data: 0.0034  max mem: 3956
+train: [1]  [ 60/400]  eta: 0:02:07  lr: 0.000069  loss: 2.3333 (2.3436)  grad: 0.0868 (0.0880)  time: 0.3292  data: 0.0040  max mem: 3956
+train: [1]  [ 80/400]  eta: 0:01:56  lr: 0.000072  loss: 2.3071 (2.3376)  grad: 0.0864 (0.0867)  time: 0.3363  data: 0.0037  max mem: 3956
+train: [1]  [100/400]  eta: 0:01:47  lr: 0.000075  loss: 2.2919 (2.3241)  grad: 0.0872 (0.0874)  time: 0.3291  data: 0.0041  max mem: 3956
+train: [1]  [120/400]  eta: 0:01:38  lr: 0.000078  loss: 2.2210 (2.3061)  grad: 0.0872 (0.0877)  time: 0.3270  data: 0.0036  max mem: 3956
+train: [1]  [140/400]  eta: 0:01:30  lr: 0.000081  loss: 2.2164 (2.2922)  grad: 0.0861 (0.0873)  time: 0.3311  data: 0.0039  max mem: 3956
+train: [1]  [160/400]  eta: 0:01:23  lr: 0.000084  loss: 2.1965 (2.2797)  grad: 0.0849 (0.0872)  time: 0.3208  data: 0.0040  max mem: 3956
+train: [1]  [180/400]  eta: 0:01:15  lr: 0.000087  loss: 2.1697 (2.2650)  grad: 0.0876 (0.0873)  time: 0.3343  data: 0.0040  max mem: 3956
+train: [1]  [200/400]  eta: 0:01:08  lr: 0.000090  loss: 2.1403 (2.2501)  grad: 0.0876 (0.0877)  time: 0.3296  data: 0.0034  max mem: 3956
+train: [1]  [220/400]  eta: 0:01:02  lr: 0.000093  loss: 2.1134 (2.2360)  grad: 0.0870 (0.0878)  time: 0.3682  data: 0.0042  max mem: 3956
+train: [1]  [240/400]  eta: 0:00:55  lr: 0.000096  loss: 2.0867 (2.2230)  grad: 0.0862 (0.0876)  time: 0.3336  data: 0.0029  max mem: 3956
+train: [1]  [260/400]  eta: 0:00:48  lr: 0.000099  loss: 2.0867 (2.2127)  grad: 0.0822 (0.0871)  time: 0.3414  data: 0.0041  max mem: 3956
+train: [1]  [280/400]  eta: 0:00:41  lr: 0.000102  loss: 2.0539 (2.2014)  grad: 0.0822 (0.0871)  time: 0.3469  data: 0.0043  max mem: 3956
+train: [1]  [300/400]  eta: 0:00:35  lr: 0.000105  loss: 2.0466 (2.1899)  grad: 0.0807 (0.0869)  time: 0.4923  data: 0.1813  max mem: 3956
+train: [1]  [320/400]  eta: 0:00:28  lr: 0.000108  loss: 2.0009 (2.1767)  grad: 0.0807 (0.0867)  time: 0.3310  data: 0.0036  max mem: 3956
+train: [1]  [340/400]  eta: 0:00:21  lr: 0.000111  loss: 1.9939 (2.1666)  grad: 0.0805 (0.0861)  time: 0.3470  data: 0.0031  max mem: 3956
+train: [1]  [360/400]  eta: 0:00:14  lr: 0.000114  loss: 1.9939 (2.1561)  grad: 0.0796 (0.0858)  time: 0.3489  data: 0.0040  max mem: 3956
+train: [1]  [380/400]  eta: 0:00:07  lr: 0.000117  loss: 1.9697 (2.1453)  grad: 0.0796 (0.0855)  time: 0.3703  data: 0.0216  max mem: 3956
+train: [1]  [399/400]  eta: 0:00:00  lr: 0.000120  loss: 1.9443 (2.1350)  grad: 0.0812 (0.0854)  time: 0.3497  data: 0.0092  max mem: 3956
+train: [1] Total time: 0:02:21 (0.3533 s / it)
+train: [1] Summary: lr: 0.000120  loss: 1.9443 (2.1350)  grad: 0.0812 (0.0854)
+eval (validation): [1]  [ 0/63]  eta: 0:03:25    time: 3.2601  data: 2.9842  max mem: 3956
+eval (validation): [1]  [20/63]  eta: 0:00:21    time: 0.3683  data: 0.0199  max mem: 3956
+eval (validation): [1]  [40/63]  eta: 0:00:09    time: 0.3377  data: 0.0034  max mem: 3956
+eval (validation): [1]  [60/63]  eta: 0:00:01    time: 0.3032  data: 0.0033  max mem: 3956
+eval (validation): [1]  [62/63]  eta: 0:00:00    time: 0.3017  data: 0.0033  max mem: 3956
+eval (validation): [1] Total time: 0:00:24 (0.3863 s / it)
+cv: [1]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.230  acc: 0.949  f1: 0.946
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-best.pth
+train: [2]  [  0/400]  eta: 0:20:46  lr: nan  time: 3.1165  data: 2.8584  max mem: 3956
+train: [2]  [ 20/400]  eta: 0:02:52  lr: 0.000123  loss: 1.8838 (1.8871)  grad: 0.0791 (0.0824)  time: 0.3214  data: 0.0037  max mem: 3956
+train: [2]  [ 40/400]  eta: 0:02:32  lr: 0.000126  loss: 1.8939 (1.9002)  grad: 0.0795 (0.0817)  time: 0.3913  data: 0.0033  max mem: 3956
+train: [2]  [ 60/400]  eta: 0:02:16  lr: 0.000129  loss: 1.8939 (1.8940)  grad: 0.0770 (0.0798)  time: 0.3564  data: 0.0043  max mem: 3956
+train: [2]  [ 80/400]  eta: 0:02:02  lr: 0.000132  loss: 1.8776 (1.8906)  grad: 0.0734 (0.0784)  time: 0.3281  data: 0.0043  max mem: 3956
+train: [2]  [100/400]  eta: 0:01:52  lr: 0.000135  loss: 1.8664 (1.8810)  grad: 0.0752 (0.0788)  time: 0.3402  data: 0.0044  max mem: 3956
+train: [2]  [120/400]  eta: 0:01:43  lr: 0.000138  loss: 1.8262 (1.8724)  grad: 0.0777 (0.0786)  time: 0.3484  data: 0.0043  max mem: 3956
+train: [2]  [140/400]  eta: 0:01:35  lr: 0.000141  loss: 1.8282 (1.8665)  grad: 0.0765 (0.0784)  time: 0.3434  data: 0.0041  max mem: 3956
+train: [2]  [160/400]  eta: 0:01:26  lr: 0.000144  loss: 1.8113 (1.8591)  grad: 0.0762 (0.0783)  time: 0.3235  data: 0.0038  max mem: 3956
+train: [2]  [180/400]  eta: 0:01:19  lr: 0.000147  loss: 1.7929 (1.8497)  grad: 0.0794 (0.0786)  time: 0.3489  data: 0.0041  max mem: 3956
+train: [2]  [200/400]  eta: 0:01:11  lr: 0.000150  loss: 1.7854 (1.8434)  grad: 0.0775 (0.0782)  time: 0.3322  data: 0.0036  max mem: 3956
+train: [2]  [220/400]  eta: 0:01:04  lr: 0.000153  loss: 1.7854 (1.8353)  grad: 0.0752 (0.0783)  time: 0.3556  data: 0.0034  max mem: 3956
+train: [2]  [240/400]  eta: 0:00:56  lr: 0.000156  loss: 1.7289 (1.8273)  grad: 0.0752 (0.0780)  time: 0.3342  data: 0.0041  max mem: 3956
+train: [2]  [260/400]  eta: 0:00:49  lr: 0.000159  loss: 1.7306 (1.8218)  grad: 0.0764 (0.0777)  time: 0.3608  data: 0.0042  max mem: 3956
+train: [2]  [280/400]  eta: 0:00:42  lr: 0.000162  loss: 1.7243 (1.8138)  grad: 0.0735 (0.0775)  time: 0.3282  data: 0.0042  max mem: 3956
+train: [2]  [300/400]  eta: 0:00:36  lr: 0.000165  loss: 1.7089 (1.8077)  grad: 0.0727 (0.0774)  time: 0.4930  data: 0.1811  max mem: 3956
+train: [2]  [320/400]  eta: 0:00:28  lr: 0.000168  loss: 1.7017 (1.8004)  grad: 0.0727 (0.0772)  time: 0.3555  data: 0.0110  max mem: 3956
+train: [2]  [340/400]  eta: 0:00:21  lr: 0.000171  loss: 1.6937 (1.7944)  grad: 0.0700 (0.0768)  time: 0.3406  data: 0.0025  max mem: 3956
+train: [2]  [360/400]  eta: 0:00:14  lr: 0.000174  loss: 1.6850 (1.7874)  grad: 0.0695 (0.0766)  time: 0.3966  data: 0.0441  max mem: 3956
+train: [2]  [380/400]  eta: 0:00:07  lr: 0.000177  loss: 1.6776 (1.7812)  grad: 0.0695 (0.0763)  time: 0.3569  data: 0.0040  max mem: 3956
+train: [2]  [399/400]  eta: 0:00:00  lr: 0.000180  loss: 1.6556 (1.7744)  grad: 0.0713 (0.0762)  time: 0.3406  data: 0.0042  max mem: 3956
+train: [2] Total time: 0:02:24 (0.3624 s / it)
+train: [2] Summary: lr: 0.000180  loss: 1.6556 (1.7744)  grad: 0.0713 (0.0762)
+eval (validation): [2]  [ 0/63]  eta: 0:03:28    time: 3.3038  data: 3.0373  max mem: 3956
+eval (validation): [2]  [20/63]  eta: 0:00:21    time: 0.3544  data: 0.0035  max mem: 3956
+eval (validation): [2]  [40/63]  eta: 0:00:09    time: 0.3069  data: 0.0033  max mem: 3956
+eval (validation): [2]  [60/63]  eta: 0:00:01    time: 0.3199  data: 0.0036  max mem: 3956
+eval (validation): [2]  [62/63]  eta: 0:00:00    time: 0.3156  data: 0.0037  max mem: 3956
+eval (validation): [2] Total time: 0:00:23 (0.3787 s / it)
+cv: [2]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.169  acc: 0.956  f1: 0.952
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-best.pth
+train: [3]  [  0/400]  eta: 0:21:25  lr: nan  time: 3.2146  data: 2.9676  max mem: 3956
+train: [3]  [ 20/400]  eta: 0:03:10  lr: 0.000183  loss: 1.6312 (1.6241)  grad: 0.0691 (0.0720)  time: 0.3659  data: 0.0032  max mem: 3956
+train: [3]  [ 40/400]  eta: 0:02:31  lr: 0.000186  loss: 1.6104 (1.6131)  grad: 0.0691 (0.0722)  time: 0.3363  data: 0.0040  max mem: 3956
+train: [3]  [ 60/400]  eta: 0:02:12  lr: 0.000189  loss: 1.5797 (1.6007)  grad: 0.0697 (0.0718)  time: 0.3232  data: 0.0042  max mem: 3956
+train: [3]  [ 80/400]  eta: 0:01:59  lr: 0.000192  loss: 1.5850 (1.6024)  grad: 0.0702 (0.0720)  time: 0.3297  data: 0.0041  max mem: 3956
+train: [3]  [100/400]  eta: 0:01:51  lr: 0.000195  loss: 1.5912 (1.5966)  grad: 0.0700 (0.0716)  time: 0.3569  data: 0.0042  max mem: 3956
+train: [3]  [120/400]  eta: 0:01:41  lr: 0.000198  loss: 1.5505 (1.5882)  grad: 0.0689 (0.0715)  time: 0.3282  data: 0.0041  max mem: 3956
+train: [3]  [140/400]  eta: 0:01:33  lr: 0.000201  loss: 1.5513 (1.5872)  grad: 0.0691 (0.0715)  time: 0.3275  data: 0.0039  max mem: 3956
+train: [3]  [160/400]  eta: 0:01:24  lr: 0.000204  loss: 1.5492 (1.5832)  grad: 0.0691 (0.0713)  time: 0.3210  data: 0.0043  max mem: 3956
+train: [3]  [180/400]  eta: 0:01:17  lr: 0.000207  loss: 1.5468 (1.5786)  grad: 0.0689 (0.0712)  time: 0.3340  data: 0.0032  max mem: 3956
+train: [3]  [200/400]  eta: 0:01:09  lr: 0.000210  loss: 1.5505 (1.5748)  grad: 0.0685 (0.0712)  time: 0.3337  data: 0.0039  max mem: 3956
+train: [3]  [220/400]  eta: 0:01:02  lr: 0.000213  loss: 1.5363 (1.5712)  grad: 0.0660 (0.0708)  time: 0.3308  data: 0.0041  max mem: 3956
+train: [3]  [240/400]  eta: 0:00:55  lr: 0.000216  loss: 1.5239 (1.5670)  grad: 0.0656 (0.0708)  time: 0.3315  data: 0.0039  max mem: 3956
+train: [3]  [260/400]  eta: 0:00:48  lr: 0.000219  loss: 1.5132 (1.5634)  grad: 0.0720 (0.0709)  time: 0.3486  data: 0.0038  max mem: 3956
+train: [3]  [280/400]  eta: 0:00:41  lr: 0.000222  loss: 1.4907 (1.5578)  grad: 0.0677 (0.0706)  time: 0.3415  data: 0.0043  max mem: 3956
+train: [3]  [300/400]  eta: 0:00:35  lr: 0.000225  loss: 1.4844 (1.5536)  grad: 0.0650 (0.0705)  time: 0.5292  data: 0.2139  max mem: 3956
+train: [3]  [320/400]  eta: 0:00:28  lr: 0.000228  loss: 1.4775 (1.5499)  grad: 0.0650 (0.0702)  time: 0.3445  data: 0.0065  max mem: 3956
+train: [3]  [340/400]  eta: 0:00:21  lr: 0.000231  loss: 1.4665 (1.5445)  grad: 0.0670 (0.0700)  time: 0.3397  data: 0.0020  max mem: 3956
+train: [3]  [360/400]  eta: 0:00:14  lr: 0.000234  loss: 1.4556 (1.5401)  grad: 0.0659 (0.0696)  time: 0.3353  data: 0.0089  max mem: 3956
+train: [3]  [380/400]  eta: 0:00:07  lr: 0.000237  loss: 1.4514 (1.5359)  grad: 0.0608 (0.0694)  time: 0.3519  data: 0.0031  max mem: 3956
+train: [3]  [399/400]  eta: 0:00:00  lr: 0.000240  loss: 1.4472 (1.5300)  grad: 0.0662 (0.0693)  time: 0.3365  data: 0.0032  max mem: 3956
+train: [3] Total time: 0:02:22 (0.3551 s / it)
+train: [3] Summary: lr: 0.000240  loss: 1.4472 (1.5300)  grad: 0.0662 (0.0693)
+eval (validation): [3]  [ 0/63]  eta: 0:03:27    time: 3.2987  data: 3.0479  max mem: 3956
+eval (validation): [3]  [20/63]  eta: 0:00:21    time: 0.3653  data: 0.0044  max mem: 3956
+eval (validation): [3]  [40/63]  eta: 0:00:09    time: 0.3443  data: 0.0032  max mem: 3956
+eval (validation): [3]  [60/63]  eta: 0:00:01    time: 0.3069  data: 0.0036  max mem: 3956
+eval (validation): [3]  [62/63]  eta: 0:00:00    time: 0.3052  data: 0.0035  max mem: 3956
+eval (validation): [3] Total time: 0:00:24 (0.3890 s / it)
+cv: [3]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.143  acc: 0.962  f1: 0.957
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-best.pth
+train: [4]  [  0/400]  eta: 0:20:12  lr: nan  time: 3.0313  data: 2.7540  max mem: 3956
+train: [4]  [ 20/400]  eta: 0:03:06  lr: 0.000243  loss: 1.4146 (1.4406)  grad: 0.0633 (0.0667)  time: 0.3643  data: 0.0037  max mem: 3956
+train: [4]  [ 40/400]  eta: 0:02:31  lr: 0.000246  loss: 1.4096 (1.4261)  grad: 0.0658 (0.0665)  time: 0.3484  data: 0.0035  max mem: 3956
+train: [4]  [ 60/400]  eta: 0:02:13  lr: 0.000249  loss: 1.3985 (1.4163)  grad: 0.0660 (0.0667)  time: 0.3323  data: 0.0043  max mem: 3956
+train: [4]  [ 80/400]  eta: 0:02:01  lr: 0.000252  loss: 1.3977 (1.4114)  grad: 0.0660 (0.0668)  time: 0.3400  data: 0.0062  max mem: 3956
+train: [4]  [100/400]  eta: 0:01:50  lr: 0.000255  loss: 1.4034 (1.4088)  grad: 0.0641 (0.0661)  time: 0.3265  data: 0.0039  max mem: 3956
+train: [4]  [120/400]  eta: 0:01:41  lr: 0.000258  loss: 1.4145 (1.4083)  grad: 0.0624 (0.0657)  time: 0.3393  data: 0.0041  max mem: 3956
+train: [4]  [140/400]  eta: 0:01:34  lr: 0.000261  loss: 1.3536 (1.4009)  grad: 0.0665 (0.0658)  time: 0.3484  data: 0.0044  max mem: 3956
+train: [4]  [160/400]  eta: 0:01:26  lr: 0.000264  loss: 1.3536 (1.3983)  grad: 0.0647 (0.0655)  time: 0.3434  data: 0.0041  max mem: 3956
+train: [4]  [180/400]  eta: 0:01:18  lr: 0.000267  loss: 1.3609 (1.3937)  grad: 0.0651 (0.0655)  time: 0.3511  data: 0.0042  max mem: 3956
+train: [4]  [200/400]  eta: 0:01:11  lr: 0.000270  loss: 1.3371 (1.3866)  grad: 0.0652 (0.0654)  time: 0.3466  data: 0.0043  max mem: 3956
+train: [4]  [220/400]  eta: 0:01:03  lr: 0.000273  loss: 1.3360 (1.3834)  grad: 0.0610 (0.0648)  time: 0.3205  data: 0.0038  max mem: 3956
+train: [4]  [240/400]  eta: 0:00:56  lr: 0.000276  loss: 1.3355 (1.3777)  grad: 0.0613 (0.0648)  time: 0.3196  data: 0.0037  max mem: 3956
+train: [4]  [260/400]  eta: 0:00:49  lr: 0.000279  loss: 1.3181 (1.3736)  grad: 0.0623 (0.0647)  time: 0.3424  data: 0.0040  max mem: 3956
+train: [4]  [280/400]  eta: 0:00:41  lr: 0.000282  loss: 1.3146 (1.3702)  grad: 0.0604 (0.0644)  time: 0.3331  data: 0.0037  max mem: 3956
+train: [4]  [300/400]  eta: 0:00:35  lr: 0.000285  loss: 1.3036 (1.3653)  grad: 0.0602 (0.0643)  time: 0.4982  data: 0.1862  max mem: 3956
+train: [4]  [320/400]  eta: 0:00:28  lr: 0.000288  loss: 1.3090 (1.3621)  grad: 0.0635 (0.0644)  time: 0.3377  data: 0.0035  max mem: 3956
+train: [4]  [340/400]  eta: 0:00:21  lr: 0.000291  loss: 1.3227 (1.3586)  grad: 0.0605 (0.0641)  time: 0.3148  data: 0.0038  max mem: 3956
+train: [4]  [360/400]  eta: 0:00:14  lr: 0.000294  loss: 1.2832 (1.3544)  grad: 0.0601 (0.0639)  time: 0.3350  data: 0.0031  max mem: 3956
+train: [4]  [380/400]  eta: 0:00:07  lr: 0.000297  loss: 1.2865 (1.3521)  grad: 0.0596 (0.0637)  time: 0.3386  data: 0.0068  max mem: 3956
+train: [4]  [399/400]  eta: 0:00:00  lr: 0.000300  loss: 1.2865 (1.3480)  grad: 0.0592 (0.0636)  time: 0.3804  data: 0.0526  max mem: 3956
+train: [4] Total time: 0:02:22 (0.3553 s / it)
+train: [4] Summary: lr: 0.000300  loss: 1.2865 (1.3480)  grad: 0.0592 (0.0636)
+eval (validation): [4]  [ 0/63]  eta: 0:03:29    time: 3.3282  data: 3.0629  max mem: 3956
+eval (validation): [4]  [20/63]  eta: 0:00:20    time: 0.3302  data: 0.0038  max mem: 3956
+eval (validation): [4]  [40/63]  eta: 0:00:09    time: 0.3178  data: 0.0034  max mem: 3956
+eval (validation): [4]  [60/63]  eta: 0:00:01    time: 0.2928  data: 0.0035  max mem: 3956
+eval (validation): [4]  [62/63]  eta: 0:00:00    time: 0.2920  data: 0.0035  max mem: 3956
+eval (validation): [4] Total time: 0:00:22 (0.3649 s / it)
+cv: [4]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.129  acc: 0.963  f1: 0.957
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-best.pth
+train: [5]  [  0/400]  eta: 0:21:42  lr: nan  time: 3.2571  data: 2.9840  max mem: 3956
+train: [5]  [ 20/400]  eta: 0:02:58  lr: 0.000300  loss: 1.2425 (1.2610)  grad: 0.0550 (0.0585)  time: 0.3302  data: 0.0270  max mem: 3956
+train: [5]  [ 40/400]  eta: 0:02:32  lr: 0.000300  loss: 1.2573 (1.2730)  grad: 0.0594 (0.0590)  time: 0.3757  data: 0.0499  max mem: 3956
+train: [5]  [ 60/400]  eta: 0:02:12  lr: 0.000300  loss: 1.2626 (1.2633)  grad: 0.0612 (0.0605)  time: 0.3190  data: 0.0035  max mem: 3956
+train: [5]  [ 80/400]  eta: 0:01:59  lr: 0.000300  loss: 1.2417 (1.2587)  grad: 0.0612 (0.0603)  time: 0.3237  data: 0.0041  max mem: 3956
+train: [5]  [100/400]  eta: 0:01:48  lr: 0.000300  loss: 1.2284 (1.2523)  grad: 0.0611 (0.0610)  time: 0.3192  data: 0.0038  max mem: 3956
+train: [5]  [120/400]  eta: 0:01:39  lr: 0.000300  loss: 1.2284 (1.2502)  grad: 0.0625 (0.0613)  time: 0.3209  data: 0.0039  max mem: 3956
+train: [5]  [140/400]  eta: 0:01:31  lr: 0.000300  loss: 1.2295 (1.2478)  grad: 0.0607 (0.0611)  time: 0.3301  data: 0.0043  max mem: 3956
+train: [5]  [160/400]  eta: 0:01:23  lr: 0.000299  loss: 1.2108 (1.2442)  grad: 0.0594 (0.0610)  time: 0.3171  data: 0.0040  max mem: 3956
+train: [5]  [180/400]  eta: 0:01:16  lr: 0.000299  loss: 1.2128 (1.2426)  grad: 0.0575 (0.0607)  time: 0.3510  data: 0.0041  max mem: 3956
+train: [5]  [200/400]  eta: 0:01:09  lr: 0.000299  loss: 1.2078 (1.2400)  grad: 0.0572 (0.0605)  time: 0.3363  data: 0.0041  max mem: 3956
+train: [5]  [220/400]  eta: 0:01:02  lr: 0.000299  loss: 1.2059 (1.2363)  grad: 0.0589 (0.0605)  time: 0.3348  data: 0.0042  max mem: 3956
+train: [5]  [240/400]  eta: 0:00:55  lr: 0.000299  loss: 1.1929 (1.2328)  grad: 0.0582 (0.0605)  time: 0.3434  data: 0.0043  max mem: 3956
+train: [5]  [260/400]  eta: 0:00:48  lr: 0.000299  loss: 1.1929 (1.2308)  grad: 0.0570 (0.0603)  time: 0.3410  data: 0.0043  max mem: 3956
+train: [5]  [280/400]  eta: 0:00:41  lr: 0.000298  loss: 1.1916 (1.2276)  grad: 0.0568 (0.0601)  time: 0.3284  data: 0.0043  max mem: 3956
+train: [5]  [300/400]  eta: 0:00:35  lr: 0.000298  loss: 1.1752 (1.2248)  grad: 0.0550 (0.0598)  time: 0.5280  data: 0.1803  max mem: 3956
+train: [5]  [320/400]  eta: 0:00:28  lr: 0.000298  loss: 1.1752 (1.2222)  grad: 0.0562 (0.0596)  time: 0.4032  data: 0.0322  max mem: 3956
+train: [5]  [340/400]  eta: 0:00:21  lr: 0.000298  loss: 1.1721 (1.2192)  grad: 0.0568 (0.0594)  time: 0.3499  data: 0.0328  max mem: 3956
+train: [5]  [360/400]  eta: 0:00:14  lr: 0.000297  loss: 1.1709 (1.2165)  grad: 0.0577 (0.0593)  time: 0.3366  data: 0.0039  max mem: 3956
+train: [5]  [380/400]  eta: 0:00:07  lr: 0.000297  loss: 1.1765 (1.2155)  grad: 0.0570 (0.0592)  time: 0.3235  data: 0.0023  max mem: 3956
+train: [5]  [399/400]  eta: 0:00:00  lr: 0.000297  loss: 1.1646 (1.2117)  grad: 0.0587 (0.0592)  time: 0.3252  data: 0.0022  max mem: 3956
+train: [5] Total time: 0:02:21 (0.3548 s / it)
+train: [5] Summary: lr: 0.000297  loss: 1.1646 (1.2117)  grad: 0.0587 (0.0592)
+eval (validation): [5]  [ 0/63]  eta: 0:03:28    time: 3.3082  data: 3.1031  max mem: 3956
+eval (validation): [5]  [20/63]  eta: 0:00:25    time: 0.4595  data: 0.1448  max mem: 3956
+eval (validation): [5]  [40/63]  eta: 0:00:10    time: 0.3421  data: 0.0036  max mem: 3956
+eval (validation): [5]  [60/63]  eta: 0:00:01    time: 0.2923  data: 0.0033  max mem: 3956
+eval (validation): [5]  [62/63]  eta: 0:00:00    time: 0.2955  data: 0.0032  max mem: 3956
+eval (validation): [5] Total time: 0:00:26 (0.4153 s / it)
+cv: [5]  best hparam: (43, 1.0) (047) ('047_lr4.3e+01_wd1.0e+00')  loss: 0.129  acc: 0.964  f1: 0.959
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-best.pth
+train: [6]  [  0/400]  eta: 0:21:08  lr: nan  time: 3.1724  data: 2.9456  max mem: 3956
+train: [6]  [ 20/400]  eta: 0:03:40  lr: 0.000296  loss: 1.1560 (1.1665)  grad: 0.0547 (0.0564)  time: 0.4515  data: 0.1134  max mem: 3956
+train: [6]  [ 40/400]  eta: 0:02:43  lr: 0.000296  loss: 1.1592 (1.1660)  grad: 0.0546 (0.0562)  time: 0.3236  data: 0.0035  max mem: 3956
+train: [6]  [ 60/400]  eta: 0:02:20  lr: 0.000296  loss: 1.1581 (1.1617)  grad: 0.0555 (0.0568)  time: 0.3230  data: 0.0040  max mem: 3956
+train: [6]  [ 80/400]  eta: 0:02:05  lr: 0.000295  loss: 1.1203 (1.1536)  grad: 0.0565 (0.0567)  time: 0.3262  data: 0.0035  max mem: 3956
+train: [6]  [100/400]  eta: 0:01:55  lr: 0.000295  loss: 1.1096 (1.1471)  grad: 0.0562 (0.0564)  time: 0.3530  data: 0.0035  max mem: 3956
+train: [6]  [120/400]  eta: 0:01:44  lr: 0.000295  loss: 1.1112 (1.1438)  grad: 0.0545 (0.0564)  time: 0.3328  data: 0.0041  max mem: 3956
+train: [6]  [140/400]  eta: 0:01:35  lr: 0.000294  loss: 1.1061 (1.1381)  grad: 0.0555 (0.0569)  time: 0.3267  data: 0.0040  max mem: 3956
+train: [6]  [160/400]  eta: 0:01:27  lr: 0.000294  loss: 1.1061 (1.1346)  grad: 0.0571 (0.0571)  time: 0.3452  data: 0.0042  max mem: 3956
+train: [6]  [180/400]  eta: 0:01:19  lr: 0.000293  loss: 1.1287 (1.1338)  grad: 0.0566 (0.0572)  time: 0.3365  data: 0.0042  max mem: 3956
+train: [6]  [200/400]  eta: 0:01:11  lr: 0.000293  loss: 1.1267 (1.1318)  grad: 0.0554 (0.0570)  time: 0.3258  data: 0.0037  max mem: 3956
+train: [6]  [220/400]  eta: 0:01:04  lr: 0.000292  loss: 1.0909 (1.1275)  grad: 0.0553 (0.0569)  time: 0.3374  data: 0.0039  max mem: 3956
+train: [6]  [240/400]  eta: 0:00:56  lr: 0.000292  loss: 1.0869 (1.1250)  grad: 0.0549 (0.0567)  time: 0.3372  data: 0.0044  max mem: 3956
+train: [6]  [260/400]  eta: 0:00:49  lr: 0.000291  loss: 1.1026 (1.1242)  grad: 0.0554 (0.0567)  time: 0.3365  data: 0.0041  max mem: 3956
+train: [6]  [280/400]  eta: 0:00:42  lr: 0.000291  loss: 1.0823 (1.1207)  grad: 0.0549 (0.0566)  time: 0.3339  data: 0.0040  max mem: 3956
+train: [6]  [300/400]  eta: 0:00:36  lr: 0.000290  loss: 1.0508 (1.1175)  grad: 0.0545 (0.0564)  time: 0.4875  data: 0.1818  max mem: 3956
+train: [6]  [320/400]  eta: 0:00:29  lr: 0.000290  loss: 1.0508 (1.1138)  grad: 0.0549 (0.0563)  time: 0.4131  data: 0.0606  max mem: 3956
+train: [6]  [340/400]  eta: 0:00:21  lr: 0.000289  loss: 1.0619 (1.1122)  grad: 0.0556 (0.0562)  time: 0.3521  data: 0.0041  max mem: 3956
+train: [6]  [360/400]  eta: 0:00:14  lr: 0.000288  loss: 1.0701 (1.1091)  grad: 0.0562 (0.0563)  time: 0.3373  data: 0.0042  max mem: 3956
+train: [6]  [380/400]  eta: 0:00:07  lr: 0.000288  loss: 1.0679 (1.1078)  grad: 0.0529 (0.0561)  time: 0.3553  data: 0.0044  max mem: 3956
+train: [6]  [399/400]  eta: 0:00:00  lr: 0.000287  loss: 1.0846 (1.1062)  grad: 0.0530 (0.0562)  time: 0.3460  data: 0.0044  max mem: 3956
+train: [6] Total time: 0:02:24 (0.3617 s / it)
+train: [6] Summary: lr: 0.000287  loss: 1.0846 (1.1062)  grad: 0.0530 (0.0562)
+eval (validation): [6]  [ 0/63]  eta: 0:03:25    time: 3.2566  data: 3.0383  max mem: 3956
+eval (validation): [6]  [20/63]  eta: 0:00:22    time: 0.3948  data: 0.0648  max mem: 3956
+eval (validation): [6]  [40/63]  eta: 0:00:10    time: 0.3899  data: 0.0029  max mem: 3956
+eval (validation): [6]  [60/63]  eta: 0:00:01    time: 0.3329  data: 0.0037  max mem: 3956
+eval (validation): [6]  [62/63]  eta: 0:00:00    time: 0.3257  data: 0.0037  max mem: 3956
+eval (validation): [6] Total time: 0:00:26 (0.4222 s / it)
+cv: [6]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.119  acc: 0.968  f1: 0.963
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-best.pth
+train: [7]  [  0/400]  eta: 0:36:44  lr: nan  time: 5.5107  data: 5.2551  max mem: 3956
+train: [7]  [ 20/400]  eta: 0:03:43  lr: 0.000286  loss: 1.0392 (1.0661)  grad: 0.0533 (0.0556)  time: 0.3416  data: 0.0208  max mem: 3956
+train: [7]  [ 40/400]  eta: 0:02:43  lr: 0.000286  loss: 1.0579 (1.0654)  grad: 0.0534 (0.0555)  time: 0.3125  data: 0.0027  max mem: 3956
+train: [7]  [ 60/400]  eta: 0:02:19  lr: 0.000285  loss: 1.0564 (1.0565)  grad: 0.0539 (0.0562)  time: 0.3254  data: 0.0039  max mem: 3956
+train: [7]  [ 80/400]  eta: 0:02:06  lr: 0.000284  loss: 1.0570 (1.0620)  grad: 0.0532 (0.0552)  time: 0.3405  data: 0.0044  max mem: 3956
+train: [7]  [100/400]  eta: 0:01:54  lr: 0.000284  loss: 1.0540 (1.0566)  grad: 0.0520 (0.0546)  time: 0.3291  data: 0.0037  max mem: 3956
+train: [7]  [120/400]  eta: 0:01:43  lr: 0.000283  loss: 1.0347 (1.0552)  grad: 0.0529 (0.0545)  time: 0.3140  data: 0.0042  max mem: 3956
+train: [7]  [140/400]  eta: 0:01:34  lr: 0.000282  loss: 1.0416 (1.0535)  grad: 0.0530 (0.0542)  time: 0.3366  data: 0.0042  max mem: 3956
+train: [7]  [160/400]  eta: 0:01:26  lr: 0.000282  loss: 1.0316 (1.0502)  grad: 0.0530 (0.0540)  time: 0.3301  data: 0.0036  max mem: 3956
+train: [7]  [180/400]  eta: 0:01:18  lr: 0.000281  loss: 1.0260 (1.0472)  grad: 0.0518 (0.0539)  time: 0.3302  data: 0.0037  max mem: 3956
+train: [7]  [200/400]  eta: 0:01:11  lr: 0.000280  loss: 1.0260 (1.0473)  grad: 0.0533 (0.0538)  time: 0.3327  data: 0.0043  max mem: 3956
+train: [7]  [220/400]  eta: 0:01:03  lr: 0.000279  loss: 1.0440 (1.0476)  grad: 0.0537 (0.0538)  time: 0.3274  data: 0.0038  max mem: 3956
+train: [7]  [240/400]  eta: 0:00:56  lr: 0.000278  loss: 1.0281 (1.0468)  grad: 0.0523 (0.0537)  time: 0.3470  data: 0.0039  max mem: 3956
+train: [7]  [260/400]  eta: 0:00:49  lr: 0.000278  loss: 1.0041 (1.0439)  grad: 0.0519 (0.0536)  time: 0.3458  data: 0.0038  max mem: 3956
+train: [7]  [280/400]  eta: 0:00:42  lr: 0.000277  loss: 1.0190 (1.0424)  grad: 0.0498 (0.0534)  time: 0.3381  data: 0.0038  max mem: 3956
+train: [7]  [300/400]  eta: 0:00:36  lr: 0.000276  loss: 1.0193 (1.0410)  grad: 0.0493 (0.0532)  time: 0.5547  data: 0.2312  max mem: 3956
+train: [7]  [320/400]  eta: 0:00:29  lr: 0.000275  loss: 1.0086 (1.0388)  grad: 0.0515 (0.0532)  time: 0.3487  data: 0.0035  max mem: 3956
+train: [7]  [340/400]  eta: 0:00:21  lr: 0.000274  loss: 0.9791 (1.0350)  grad: 0.0538 (0.0533)  time: 0.3363  data: 0.0036  max mem: 3956
+train: [7]  [360/400]  eta: 0:00:14  lr: 0.000273  loss: 1.0008 (1.0352)  grad: 0.0536 (0.0532)  time: 0.3380  data: 0.0036  max mem: 3956
+train: [7]  [380/400]  eta: 0:00:07  lr: 0.000272  loss: 1.0389 (1.0340)  grad: 0.0531 (0.0533)  time: 0.3422  data: 0.0037  max mem: 3956
+train: [7]  [399/400]  eta: 0:00:00  lr: 0.000271  loss: 0.9997 (1.0321)  grad: 0.0520 (0.0533)  time: 0.3293  data: 0.0042  max mem: 3956
+train: [7] Total time: 0:02:23 (0.3585 s / it)
+train: [7] Summary: lr: 0.000271  loss: 0.9997 (1.0321)  grad: 0.0520 (0.0533)
+eval (validation): [7]  [ 0/63]  eta: 0:06:48    time: 6.4913  data: 6.2382  max mem: 3956
+eval (validation): [7]  [20/63]  eta: 0:00:25    time: 0.3014  data: 0.0029  max mem: 3956
+eval (validation): [7]  [40/63]  eta: 0:00:10    time: 0.3126  data: 0.0030  max mem: 3956
+eval (validation): [7]  [60/63]  eta: 0:00:01    time: 0.2997  data: 0.0033  max mem: 3956
+eval (validation): [7]  [62/63]  eta: 0:00:00    time: 0.2989  data: 0.0031  max mem: 3956
+eval (validation): [7] Total time: 0:00:25 (0.4088 s / it)
+cv: [7]  best hparam: (43, 1.0) (047) ('047_lr4.3e+01_wd1.0e+00')  loss: 0.118  acc: 0.968  f1: 0.963
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+train: [8]  [  0/400]  eta: 0:21:16  lr: nan  time: 3.1914  data: 2.9813  max mem: 3956
+train: [8]  [ 20/400]  eta: 0:03:14  lr: 0.000270  loss: 0.9779 (0.9844)  grad: 0.0488 (0.0514)  time: 0.3765  data: 0.0039  max mem: 3956
+train: [8]  [ 40/400]  eta: 0:02:32  lr: 0.000270  loss: 0.9861 (0.9936)  grad: 0.0506 (0.0509)  time: 0.3303  data: 0.0035  max mem: 3956
+train: [8]  [ 60/400]  eta: 0:02:13  lr: 0.000269  loss: 0.9942 (0.9971)  grad: 0.0512 (0.0514)  time: 0.3302  data: 0.0042  max mem: 3956
+train: [8]  [ 80/400]  eta: 0:02:00  lr: 0.000268  loss: 1.0144 (1.0022)  grad: 0.0500 (0.0511)  time: 0.3309  data: 0.0043  max mem: 3956
+train: [8]  [100/400]  eta: 0:01:50  lr: 0.000267  loss: 1.0038 (0.9991)  grad: 0.0491 (0.0511)  time: 0.3276  data: 0.0043  max mem: 3956
+train: [8]  [120/400]  eta: 0:01:41  lr: 0.000266  loss: 0.9876 (0.9975)  grad: 0.0523 (0.0513)  time: 0.3290  data: 0.0040  max mem: 3956
+train: [8]  [140/400]  eta: 0:01:33  lr: 0.000265  loss: 0.9806 (0.9956)  grad: 0.0504 (0.0512)  time: 0.3491  data: 0.0040  max mem: 3956
+train: [8]  [160/400]  eta: 0:01:25  lr: 0.000264  loss: 0.9605 (0.9905)  grad: 0.0492 (0.0512)  time: 0.3441  data: 0.0037  max mem: 3956
+train: [8]  [180/400]  eta: 0:01:18  lr: 0.000263  loss: 0.9505 (0.9873)  grad: 0.0514 (0.0515)  time: 0.3324  data: 0.0039  max mem: 3956
+train: [8]  [200/400]  eta: 0:01:10  lr: 0.000262  loss: 0.9725 (0.9862)  grad: 0.0504 (0.0515)  time: 0.3172  data: 0.0038  max mem: 3956
+train: [8]  [220/400]  eta: 0:01:02  lr: 0.000260  loss: 0.9792 (0.9863)  grad: 0.0503 (0.0515)  time: 0.3142  data: 0.0040  max mem: 3956
+train: [8]  [240/400]  eta: 0:00:55  lr: 0.000259  loss: 0.9638 (0.9831)  grad: 0.0533 (0.0517)  time: 0.3431  data: 0.0036  max mem: 3956
+train: [8]  [260/400]  eta: 0:00:48  lr: 0.000258  loss: 0.9638 (0.9823)  grad: 0.0516 (0.0516)  time: 0.3515  data: 0.0041  max mem: 3956
+train: [8]  [280/400]  eta: 0:00:41  lr: 0.000257  loss: 0.9532 (0.9801)  grad: 0.0499 (0.0516)  time: 0.3396  data: 0.0041  max mem: 3956
+train: [8]  [300/400]  eta: 0:00:35  lr: 0.000256  loss: 0.9529 (0.9793)  grad: 0.0499 (0.0515)  time: 0.4939  data: 0.1688  max mem: 3956
+train: [8]  [320/400]  eta: 0:00:28  lr: 0.000255  loss: 0.9577 (0.9774)  grad: 0.0478 (0.0513)  time: 0.3580  data: 0.0040  max mem: 3956
+train: [8]  [340/400]  eta: 0:00:21  lr: 0.000254  loss: 0.9577 (0.9766)  grad: 0.0485 (0.0512)  time: 0.3915  data: 0.0041  max mem: 3956
+train: [8]  [360/400]  eta: 0:00:14  lr: 0.000253  loss: 0.9417 (0.9746)  grad: 0.0498 (0.0512)  time: 0.3287  data: 0.0039  max mem: 3956
+train: [8]  [380/400]  eta: 0:00:07  lr: 0.000252  loss: 0.9417 (0.9742)  grad: 0.0496 (0.0511)  time: 0.3428  data: 0.0041  max mem: 3956
+train: [8]  [399/400]  eta: 0:00:00  lr: 0.000250  loss: 0.9564 (0.9735)  grad: 0.0484 (0.0511)  time: 0.3346  data: 0.0035  max mem: 3956
+train: [8] Total time: 0:02:22 (0.3558 s / it)
+train: [8] Summary: lr: 0.000250  loss: 0.9564 (0.9735)  grad: 0.0484 (0.0511)
+eval (validation): [8]  [ 0/63]  eta: 0:03:17    time: 3.1306  data: 2.9324  max mem: 3956
+eval (validation): [8]  [20/63]  eta: 0:00:19    time: 0.3133  data: 0.0136  max mem: 3956
+eval (validation): [8]  [40/63]  eta: 0:00:09    time: 0.3462  data: 0.0032  max mem: 3956
+eval (validation): [8]  [60/63]  eta: 0:00:01    time: 0.3156  data: 0.0031  max mem: 3956
+eval (validation): [8]  [62/63]  eta: 0:00:00    time: 0.3104  data: 0.0024  max mem: 3956
+eval (validation): [8] Total time: 0:00:23 (0.3730 s / it)
+cv: [8]  best hparam: (36, 1.0) (046) ('046_lr3.6e+01_wd1.0e+00')  loss: 0.117  acc: 0.968  f1: 0.963
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+train: [9]  [  0/400]  eta: 0:21:43  lr: nan  time: 3.2597  data: 3.0418  max mem: 3956
+train: [9]  [ 20/400]  eta: 0:03:05  lr: 0.000249  loss: 0.9247 (0.9298)  grad: 0.0521 (0.0513)  time: 0.3493  data: 0.0041  max mem: 3956
+train: [9]  [ 40/400]  eta: 0:02:30  lr: 0.000248  loss: 0.9247 (0.9353)  grad: 0.0514 (0.0510)  time: 0.3426  data: 0.0035  max mem: 3956
+train: [9]  [ 60/400]  eta: 0:02:13  lr: 0.000247  loss: 0.9209 (0.9333)  grad: 0.0488 (0.0504)  time: 0.3413  data: 0.0041  max mem: 3956
+train: [9]  [ 80/400]  eta: 0:02:01  lr: 0.000246  loss: 0.9273 (0.9322)  grad: 0.0487 (0.0507)  time: 0.3392  data: 0.0039  max mem: 3956
+train: [9]  [100/400]  eta: 0:01:51  lr: 0.000244  loss: 0.9224 (0.9266)  grad: 0.0513 (0.0508)  time: 0.3376  data: 0.0034  max mem: 3956
+train: [9]  [120/400]  eta: 0:01:41  lr: 0.000243  loss: 0.9224 (0.9281)  grad: 0.0486 (0.0504)  time: 0.3294  data: 0.0037  max mem: 3956
+train: [9]  [140/400]  eta: 0:01:34  lr: 0.000242  loss: 0.9391 (0.9298)  grad: 0.0473 (0.0502)  time: 0.3535  data: 0.0036  max mem: 3956
+train: [9]  [160/400]  eta: 0:01:26  lr: 0.000241  loss: 0.9245 (0.9274)  grad: 0.0481 (0.0500)  time: 0.3545  data: 0.0041  max mem: 3956
+train: [9]  [180/400]  eta: 0:01:19  lr: 0.000240  loss: 0.9095 (0.9287)  grad: 0.0476 (0.0498)  time: 0.3440  data: 0.0042  max mem: 3956
+train: [9]  [200/400]  eta: 0:01:11  lr: 0.000238  loss: 0.9266 (0.9288)  grad: 0.0470 (0.0497)  time: 0.3525  data: 0.0038  max mem: 3956
+train: [9]  [220/400]  eta: 0:01:04  lr: 0.000237  loss: 0.9189 (0.9288)  grad: 0.0474 (0.0495)  time: 0.3486  data: 0.0036  max mem: 3956
+train: [9]  [240/400]  eta: 0:00:57  lr: 0.000236  loss: 0.9189 (0.9281)  grad: 0.0485 (0.0496)  time: 0.3529  data: 0.0036  max mem: 3956
+train: [9]  [260/400]  eta: 0:00:49  lr: 0.000234  loss: 0.9143 (0.9259)  grad: 0.0493 (0.0496)  time: 0.3385  data: 0.0039  max mem: 3956
+train: [9]  [280/400]  eta: 0:00:42  lr: 0.000233  loss: 0.9142 (0.9262)  grad: 0.0490 (0.0495)  time: 0.3435  data: 0.0040  max mem: 3956
+train: [9]  [300/400]  eta: 0:00:36  lr: 0.000232  loss: 0.9142 (0.9250)  grad: 0.0481 (0.0493)  time: 0.4710  data: 0.1606  max mem: 3956
+train: [9]  [320/400]  eta: 0:00:28  lr: 0.000230  loss: 0.8975 (0.9233)  grad: 0.0471 (0.0494)  time: 0.3546  data: 0.0041  max mem: 3956
+train: [9]  [340/400]  eta: 0:00:21  lr: 0.000229  loss: 0.9041 (0.9221)  grad: 0.0472 (0.0493)  time: 0.3506  data: 0.0034  max mem: 3956
+train: [9]  [360/400]  eta: 0:00:14  lr: 0.000228  loss: 0.9072 (0.9227)  grad: 0.0473 (0.0493)  time: 0.3758  data: 0.0041  max mem: 3956
+train: [9]  [380/400]  eta: 0:00:07  lr: 0.000226  loss: 0.9180 (0.9218)  grad: 0.0477 (0.0493)  time: 0.3616  data: 0.0043  max mem: 3956
+train: [9]  [399/400]  eta: 0:00:00  lr: 0.000225  loss: 0.9161 (0.9216)  grad: 0.0476 (0.0492)  time: 0.3306  data: 0.0040  max mem: 3956
+train: [9] Total time: 0:02:24 (0.3613 s / it)
+train: [9] Summary: lr: 0.000225  loss: 0.9161 (0.9216)  grad: 0.0476 (0.0492)
+eval (validation): [9]  [ 0/63]  eta: 0:03:31    time: 3.3587  data: 3.0815  max mem: 3956
+eval (validation): [9]  [20/63]  eta: 0:00:20    time: 0.3430  data: 0.0043  max mem: 3956
+eval (validation): [9]  [40/63]  eta: 0:00:09    time: 0.3379  data: 0.0031  max mem: 3956
+eval (validation): [9]  [60/63]  eta: 0:00:01    time: 0.3207  data: 0.0038  max mem: 3956
+eval (validation): [9]  [62/63]  eta: 0:00:00    time: 0.3182  data: 0.0039  max mem: 3956
+eval (validation): [9] Total time: 0:00:24 (0.3859 s / it)
+cv: [9]  best hparam: (50, 1.0) (048) ('048_lr5.0e+01_wd1.0e+00')  loss: 0.118  acc: 0.965  f1: 0.960
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+train: [10]  [  0/400]  eta: 0:20:24  lr: nan  time: 3.0616  data: 2.8085  max mem: 3956
+train: [10]  [ 20/400]  eta: 0:03:06  lr: 0.000224  loss: 0.9126 (0.9278)  grad: 0.0468 (0.0470)  time: 0.3610  data: 0.0038  max mem: 3956
+train: [10]  [ 40/400]  eta: 0:02:32  lr: 0.000222  loss: 0.8986 (0.9017)  grad: 0.0479 (0.0492)  time: 0.3530  data: 0.0038  max mem: 3956
+train: [10]  [ 60/400]  eta: 0:02:16  lr: 0.000221  loss: 0.8661 (0.8986)  grad: 0.0502 (0.0492)  time: 0.3545  data: 0.0041  max mem: 3956
+train: [10]  [ 80/400]  eta: 0:02:04  lr: 0.000220  loss: 0.8924 (0.8966)  grad: 0.0467 (0.0487)  time: 0.3504  data: 0.0042  max mem: 3956
+train: [10]  [100/400]  eta: 0:01:53  lr: 0.000218  loss: 0.8992 (0.8993)  grad: 0.0463 (0.0483)  time: 0.3342  data: 0.0038  max mem: 3956
+train: [10]  [120/400]  eta: 0:01:43  lr: 0.000217  loss: 0.9080 (0.8963)  grad: 0.0475 (0.0482)  time: 0.3385  data: 0.0039  max mem: 3956
+train: [10]  [140/400]  eta: 0:01:36  lr: 0.000215  loss: 0.8786 (0.8950)  grad: 0.0459 (0.0479)  time: 0.3671  data: 0.0041  max mem: 3956
+train: [10]  [160/400]  eta: 0:01:28  lr: 0.000214  loss: 0.8650 (0.8920)  grad: 0.0472 (0.0480)  time: 0.3546  data: 0.0040  max mem: 3956
+train: [10]  [180/400]  eta: 0:01:20  lr: 0.000213  loss: 0.8895 (0.8967)  grad: 0.0472 (0.0477)  time: 0.3489  data: 0.0042  max mem: 3956
+train: [10]  [200/400]  eta: 0:01:12  lr: 0.000211  loss: 0.8982 (0.8966)  grad: 0.0461 (0.0477)  time: 0.3510  data: 0.0041  max mem: 3956
+train: [10]  [220/400]  eta: 0:01:05  lr: 0.000210  loss: 0.8982 (0.8962)  grad: 0.0476 (0.0478)  time: 0.3390  data: 0.0041  max mem: 3956
+train: [10]  [240/400]  eta: 0:00:57  lr: 0.000208  loss: 0.8864 (0.8930)  grad: 0.0486 (0.0480)  time: 0.3586  data: 0.0043  max mem: 3956
+train: [10]  [260/400]  eta: 0:00:50  lr: 0.000207  loss: 0.8538 (0.8911)  grad: 0.0481 (0.0480)  time: 0.3455  data: 0.0043  max mem: 3956
+train: [10]  [280/400]  eta: 0:00:43  lr: 0.000205  loss: 0.8635 (0.8907)  grad: 0.0494 (0.0482)  time: 0.3612  data: 0.0045  max mem: 3956
+train: [10]  [300/400]  eta: 0:00:37  lr: 0.000204  loss: 0.8817 (0.8903)  grad: 0.0494 (0.0481)  time: 0.5029  data: 0.1813  max mem: 3956
+train: [10]  [320/400]  eta: 0:00:29  lr: 0.000202  loss: 0.8776 (0.8899)  grad: 0.0477 (0.0480)  time: 0.3623  data: 0.0046  max mem: 3956
+train: [10]  [340/400]  eta: 0:00:22  lr: 0.000201  loss: 0.8621 (0.8875)  grad: 0.0470 (0.0479)  time: 0.3933  data: 0.0032  max mem: 3956
+train: [10]  [360/400]  eta: 0:00:14  lr: 0.000199  loss: 0.8537 (0.8861)  grad: 0.0470 (0.0479)  time: 0.3487  data: 0.0040  max mem: 3956
+train: [10]  [380/400]  eta: 0:00:07  lr: 0.000198  loss: 0.8520 (0.8842)  grad: 0.0480 (0.0479)  time: 0.3532  data: 0.0042  max mem: 3956
+train: [10]  [399/400]  eta: 0:00:00  lr: 0.000196  loss: 0.8510 (0.8819)  grad: 0.0499 (0.0481)  time: 0.3520  data: 0.0043  max mem: 3956
+train: [10] Total time: 0:02:27 (0.3688 s / it)
+train: [10] Summary: lr: 0.000196  loss: 0.8510 (0.8819)  grad: 0.0499 (0.0481)
+eval (validation): [10]  [ 0/63]  eta: 0:03:31    time: 3.3511  data: 3.0817  max mem: 3956
+eval (validation): [10]  [20/63]  eta: 0:00:21    time: 0.3607  data: 0.0036  max mem: 3956
+eval (validation): [10]  [40/63]  eta: 0:00:09    time: 0.3355  data: 0.0033  max mem: 3956
+eval (validation): [10]  [60/63]  eta: 0:00:01    time: 0.3362  data: 0.0037  max mem: 3956
+eval (validation): [10]  [62/63]  eta: 0:00:00    time: 0.3365  data: 0.0036  max mem: 3956
+eval (validation): [10] Total time: 0:00:24 (0.3961 s / it)
+cv: [10]  best hparam: (26, 1.0) (044) ('044_lr2.6e+01_wd1.0e+00')  loss: 0.117  acc: 0.968  f1: 0.961
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+train: [11]  [  0/400]  eta: 0:21:38  lr: nan  time: 3.2454  data: 2.9815  max mem: 3956
+train: [11]  [ 20/400]  eta: 0:03:11  lr: 0.000195  loss: 0.8617 (0.8563)  grad: 0.0455 (0.0459)  time: 0.3661  data: 0.0186  max mem: 3956
+train: [11]  [ 40/400]  eta: 0:02:34  lr: 0.000193  loss: 0.8610 (0.8483)  grad: 0.0463 (0.0477)  time: 0.3505  data: 0.0033  max mem: 3956
+train: [11]  [ 60/400]  eta: 0:02:16  lr: 0.000192  loss: 0.8570 (0.8543)  grad: 0.0453 (0.0472)  time: 0.3419  data: 0.0040  max mem: 3956
+train: [11]  [ 80/400]  eta: 0:02:03  lr: 0.000190  loss: 0.8563 (0.8522)  grad: 0.0463 (0.0474)  time: 0.3471  data: 0.0039  max mem: 3956
+train: [11]  [100/400]  eta: 0:01:53  lr: 0.000189  loss: 0.8399 (0.8528)  grad: 0.0491 (0.0476)  time: 0.3473  data: 0.0043  max mem: 3956
+train: [11]  [120/400]  eta: 0:01:44  lr: 0.000187  loss: 0.8503 (0.8546)  grad: 0.0491 (0.0478)  time: 0.3366  data: 0.0038  max mem: 3956
+train: [11]  [140/400]  eta: 0:01:35  lr: 0.000186  loss: 0.8552 (0.8576)  grad: 0.0465 (0.0474)  time: 0.3477  data: 0.0038  max mem: 3956
+train: [11]  [160/400]  eta: 0:01:27  lr: 0.000184  loss: 0.8543 (0.8568)  grad: 0.0447 (0.0473)  time: 0.3325  data: 0.0041  max mem: 3956
+train: [11]  [180/400]  eta: 0:01:19  lr: 0.000183  loss: 0.8412 (0.8560)  grad: 0.0464 (0.0474)  time: 0.3439  data: 0.0043  max mem: 3956
+train: [11]  [200/400]  eta: 0:01:12  lr: 0.000181  loss: 0.8526 (0.8572)  grad: 0.0463 (0.0473)  time: 0.3423  data: 0.0041  max mem: 3956
+train: [11]  [220/400]  eta: 0:01:04  lr: 0.000180  loss: 0.8608 (0.8580)  grad: 0.0449 (0.0472)  time: 0.3687  data: 0.0041  max mem: 3956
+train: [11]  [240/400]  eta: 0:00:57  lr: 0.000178  loss: 0.8658 (0.8599)  grad: 0.0460 (0.0472)  time: 0.3550  data: 0.0042  max mem: 3956
+train: [11]  [260/400]  eta: 0:00:50  lr: 0.000177  loss: 0.8678 (0.8601)  grad: 0.0456 (0.0471)  time: 0.3361  data: 0.0041  max mem: 3956
+train: [11]  [280/400]  eta: 0:00:43  lr: 0.000175  loss: 0.8522 (0.8582)  grad: 0.0441 (0.0469)  time: 0.3625  data: 0.0041  max mem: 3956
+train: [11]  [300/400]  eta: 0:00:36  lr: 0.000174  loss: 0.8328 (0.8574)  grad: 0.0433 (0.0468)  time: 0.4705  data: 0.1678  max mem: 3956
+train: [11]  [320/400]  eta: 0:00:29  lr: 0.000172  loss: 0.8330 (0.8566)  grad: 0.0450 (0.0467)  time: 0.3430  data: 0.0060  max mem: 3956
+train: [11]  [340/400]  eta: 0:00:21  lr: 0.000170  loss: 0.8616 (0.8573)  grad: 0.0460 (0.0467)  time: 0.3422  data: 0.0031  max mem: 3956
+train: [11]  [360/400]  eta: 0:00:14  lr: 0.000169  loss: 0.8713 (0.8573)  grad: 0.0473 (0.0468)  time: 0.3297  data: 0.0042  max mem: 3956
+train: [11]  [380/400]  eta: 0:00:07  lr: 0.000167  loss: 0.8467 (0.8562)  grad: 0.0460 (0.0468)  time: 0.3307  data: 0.0040  max mem: 3956
+train: [11]  [399/400]  eta: 0:00:00  lr: 0.000166  loss: 0.8467 (0.8554)  grad: 0.0461 (0.0467)  time: 0.3542  data: 0.0042  max mem: 3956
+train: [11] Total time: 0:02:24 (0.3602 s / it)
+train: [11] Summary: lr: 0.000166  loss: 0.8467 (0.8554)  grad: 0.0461 (0.0467)
+eval (validation): [11]  [ 0/63]  eta: 0:03:26    time: 3.2801  data: 3.0205  max mem: 3956
+eval (validation): [11]  [20/63]  eta: 0:00:20    time: 0.3258  data: 0.0194  max mem: 3956
+eval (validation): [11]  [40/63]  eta: 0:00:09    time: 0.3324  data: 0.0039  max mem: 3956
+eval (validation): [11]  [60/63]  eta: 0:00:01    time: 0.2884  data: 0.0026  max mem: 3956
+eval (validation): [11]  [62/63]  eta: 0:00:00    time: 0.2855  data: 0.0027  max mem: 3956
+eval (validation): [11] Total time: 0:00:23 (0.3659 s / it)
+cv: [11]  best hparam: (26, 1.0) (044) ('044_lr2.6e+01_wd1.0e+00')  loss: 0.117  acc: 0.968  f1: 0.962
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+train: [12]  [  0/400]  eta: 0:19:53  lr: nan  time: 2.9842  data: 2.7634  max mem: 3956
+train: [12]  [ 20/400]  eta: 0:02:53  lr: 0.000164  loss: 0.8304 (0.8296)  grad: 0.0491 (0.0483)  time: 0.3294  data: 0.0031  max mem: 3956
+train: [12]  [ 40/400]  eta: 0:02:21  lr: 0.000163  loss: 0.8372 (0.8381)  grad: 0.0474 (0.0467)  time: 0.3262  data: 0.0038  max mem: 3956
+train: [12]  [ 60/400]  eta: 0:02:06  lr: 0.000161  loss: 0.8234 (0.8337)  grad: 0.0459 (0.0471)  time: 0.3260  data: 0.0042  max mem: 3956
+train: [12]  [ 80/400]  eta: 0:01:54  lr: 0.000160  loss: 0.8212 (0.8337)  grad: 0.0457 (0.0466)  time: 0.3215  data: 0.0041  max mem: 3956
+train: [12]  [100/400]  eta: 0:01:45  lr: 0.000158  loss: 0.8140 (0.8294)  grad: 0.0457 (0.0468)  time: 0.3302  data: 0.0042  max mem: 3956
+train: [12]  [120/400]  eta: 0:01:37  lr: 0.000156  loss: 0.8102 (0.8289)  grad: 0.0459 (0.0468)  time: 0.3232  data: 0.0039  max mem: 3956
+train: [12]  [140/400]  eta: 0:01:29  lr: 0.000155  loss: 0.8148 (0.8277)  grad: 0.0454 (0.0470)  time: 0.3278  data: 0.0039  max mem: 3956
+train: [12]  [160/400]  eta: 0:01:22  lr: 0.000153  loss: 0.8120 (0.8258)  grad: 0.0453 (0.0468)  time: 0.3265  data: 0.0043  max mem: 3956
+train: [12]  [180/400]  eta: 0:01:14  lr: 0.000152  loss: 0.8117 (0.8262)  grad: 0.0446 (0.0466)  time: 0.3196  data: 0.0036  max mem: 3956
+train: [12]  [200/400]  eta: 0:01:07  lr: 0.000150  loss: 0.8429 (0.8288)  grad: 0.0446 (0.0465)  time: 0.3149  data: 0.0038  max mem: 3956
+train: [12]  [220/400]  eta: 0:01:00  lr: 0.000149  loss: 0.8544 (0.8308)  grad: 0.0453 (0.0464)  time: 0.3189  data: 0.0042  max mem: 3956
+train: [12]  [240/400]  eta: 0:00:53  lr: 0.000147  loss: 0.8405 (0.8316)  grad: 0.0443 (0.0462)  time: 0.3219  data: 0.0039  max mem: 3956
+train: [12]  [260/400]  eta: 0:00:46  lr: 0.000145  loss: 0.8229 (0.8314)  grad: 0.0453 (0.0463)  time: 0.3241  data: 0.0036  max mem: 3956
+train: [12]  [280/400]  eta: 0:00:40  lr: 0.000144  loss: 0.8138 (0.8296)  grad: 0.0470 (0.0462)  time: 0.3566  data: 0.0043  max mem: 3956
+train: [12]  [300/400]  eta: 0:00:34  lr: 0.000142  loss: 0.8118 (0.8286)  grad: 0.0464 (0.0462)  time: 0.4690  data: 0.1631  max mem: 3956
+train: [12]  [320/400]  eta: 0:00:27  lr: 0.000141  loss: 0.8166 (0.8288)  grad: 0.0454 (0.0462)  time: 0.3235  data: 0.0029  max mem: 3956
+train: [12]  [340/400]  eta: 0:00:20  lr: 0.000139  loss: 0.8186 (0.8280)  grad: 0.0437 (0.0461)  time: 0.3620  data: 0.0031  max mem: 3956
+train: [12]  [360/400]  eta: 0:00:13  lr: 0.000138  loss: 0.8214 (0.8288)  grad: 0.0435 (0.0461)  time: 0.3653  data: 0.0044  max mem: 3956
+train: [12]  [380/400]  eta: 0:00:06  lr: 0.000136  loss: 0.8179 (0.8269)  grad: 0.0471 (0.0462)  time: 0.3497  data: 0.0041  max mem: 3956
+train: [12]  [399/400]  eta: 0:00:00  lr: 0.000134  loss: 0.8166 (0.8270)  grad: 0.0464 (0.0461)  time: 0.3506  data: 0.0045  max mem: 3956
+train: [12] Total time: 0:02:18 (0.3463 s / it)
+train: [12] Summary: lr: 0.000134  loss: 0.8166 (0.8270)  grad: 0.0464 (0.0461)
+eval (validation): [12]  [ 0/63]  eta: 0:03:25    time: 3.2651  data: 3.0508  max mem: 3956
+eval (validation): [12]  [20/63]  eta: 0:00:21    time: 0.3596  data: 0.0041  max mem: 3956
+eval (validation): [12]  [40/63]  eta: 0:00:09    time: 0.3452  data: 0.0032  max mem: 3956
+eval (validation): [12]  [60/63]  eta: 0:00:01    time: 0.3093  data: 0.0033  max mem: 3956
+eval (validation): [12]  [62/63]  eta: 0:00:00    time: 0.3066  data: 0.0032  max mem: 3956
+eval (validation): [12] Total time: 0:00:24 (0.3888 s / it)
+cv: [12]  best hparam: (26, 1.0) (044) ('044_lr2.6e+01_wd1.0e+00')  loss: 0.115  acc: 0.969  f1: 0.963
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-best.pth
+train: [13]  [  0/400]  eta: 0:22:12  lr: nan  time: 3.3317  data: 3.0861  max mem: 3956
+train: [13]  [ 20/400]  eta: 0:03:06  lr: 0.000133  loss: 0.7946 (0.8238)  grad: 0.0429 (0.0440)  time: 0.3500  data: 0.0039  max mem: 3956
+train: [13]  [ 40/400]  eta: 0:02:32  lr: 0.000131  loss: 0.7974 (0.8095)  grad: 0.0439 (0.0439)  time: 0.3535  data: 0.0032  max mem: 3956
+train: [13]  [ 60/400]  eta: 0:02:14  lr: 0.000130  loss: 0.7974 (0.8123)  grad: 0.0440 (0.0438)  time: 0.3350  data: 0.0044  max mem: 3956
+train: [13]  [ 80/400]  eta: 0:02:02  lr: 0.000128  loss: 0.8081 (0.8107)  grad: 0.0440 (0.0443)  time: 0.3452  data: 0.0044  max mem: 3956
+train: [13]  [100/400]  eta: 0:01:52  lr: 0.000127  loss: 0.8108 (0.8132)  grad: 0.0443 (0.0442)  time: 0.3496  data: 0.0044  max mem: 3956
+train: [13]  [120/400]  eta: 0:01:42  lr: 0.000125  loss: 0.8108 (0.8131)  grad: 0.0440 (0.0441)  time: 0.3191  data: 0.0039  max mem: 3956
+train: [13]  [140/400]  eta: 0:01:34  lr: 0.000124  loss: 0.7970 (0.8109)  grad: 0.0441 (0.0444)  time: 0.3350  data: 0.0041  max mem: 3956
+train: [13]  [160/400]  eta: 0:01:26  lr: 0.000122  loss: 0.7961 (0.8117)  grad: 0.0458 (0.0444)  time: 0.3338  data: 0.0042  max mem: 3956
+train: [13]  [180/400]  eta: 0:01:18  lr: 0.000120  loss: 0.7922 (0.8082)  grad: 0.0458 (0.0447)  time: 0.3486  data: 0.0045  max mem: 3956
+train: [13]  [200/400]  eta: 0:01:11  lr: 0.000119  loss: 0.7919 (0.8094)  grad: 0.0456 (0.0448)  time: 0.3359  data: 0.0042  max mem: 3956
+train: [13]  [220/400]  eta: 0:01:03  lr: 0.000117  loss: 0.7959 (0.8091)  grad: 0.0453 (0.0448)  time: 0.3332  data: 0.0043  max mem: 3956
+train: [13]  [240/400]  eta: 0:00:56  lr: 0.000116  loss: 0.8024 (0.8094)  grad: 0.0453 (0.0449)  time: 0.3404  data: 0.0040  max mem: 3956
+train: [13]  [260/400]  eta: 0:00:49  lr: 0.000114  loss: 0.8079 (0.8087)  grad: 0.0439 (0.0449)  time: 0.3419  data: 0.0040  max mem: 3956
+train: [13]  [280/400]  eta: 0:00:42  lr: 0.000113  loss: 0.7805 (0.8076)  grad: 0.0445 (0.0449)  time: 0.3671  data: 0.0043  max mem: 3956
+train: [13]  [300/400]  eta: 0:00:36  lr: 0.000111  loss: 0.7830 (0.8072)  grad: 0.0445 (0.0448)  time: 0.5141  data: 0.1702  max mem: 3956
+train: [13]  [320/400]  eta: 0:00:28  lr: 0.000110  loss: 0.8279 (0.8091)  grad: 0.0441 (0.0448)  time: 0.3379  data: 0.0036  max mem: 3956
+train: [13]  [340/400]  eta: 0:00:21  lr: 0.000108  loss: 0.8138 (0.8076)  grad: 0.0438 (0.0447)  time: 0.3339  data: 0.0037  max mem: 3956
+train: [13]  [360/400]  eta: 0:00:14  lr: 0.000107  loss: 0.7944 (0.8078)  grad: 0.0434 (0.0447)  time: 0.3420  data: 0.0040  max mem: 3956
+train: [13]  [380/400]  eta: 0:00:07  lr: 0.000105  loss: 0.8071 (0.8081)  grad: 0.0434 (0.0447)  time: 0.3541  data: 0.0042  max mem: 3956
+train: [13]  [399/400]  eta: 0:00:00  lr: 0.000104  loss: 0.7961 (0.8069)  grad: 0.0434 (0.0447)  time: 0.3383  data: 0.0044  max mem: 3956
+train: [13] Total time: 0:02:23 (0.3585 s / it)
+train: [13] Summary: lr: 0.000104  loss: 0.7961 (0.8069)  grad: 0.0434 (0.0447)
+eval (validation): [13]  [ 0/63]  eta: 0:03:20    time: 3.1828  data: 2.9779  max mem: 3956
+eval (validation): [13]  [20/63]  eta: 0:00:21    time: 0.3676  data: 0.0237  max mem: 3956
+eval (validation): [13]  [40/63]  eta: 0:00:09    time: 0.3400  data: 0.0034  max mem: 3956
+eval (validation): [13]  [60/63]  eta: 0:00:01    time: 0.3140  data: 0.0037  max mem: 3956
+eval (validation): [13]  [62/63]  eta: 0:00:00    time: 0.3120  data: 0.0037  max mem: 3956
+eval (validation): [13] Total time: 0:00:24 (0.3897 s / it)
+cv: [13]  best hparam: (36, 1.0) (046) ('046_lr3.6e+01_wd1.0e+00')  loss: 0.113  acc: 0.967  f1: 0.962
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+train: [14]  [  0/400]  eta: 0:21:56  lr: nan  time: 3.2921  data: 3.0138  max mem: 3956
+train: [14]  [ 20/400]  eta: 0:03:09  lr: 0.000102  loss: 0.7662 (0.7854)  grad: 0.0469 (0.0479)  time: 0.3578  data: 0.0050  max mem: 3956
+train: [14]  [ 40/400]  eta: 0:02:30  lr: 0.000101  loss: 0.7674 (0.7843)  grad: 0.0448 (0.0452)  time: 0.3339  data: 0.0042  max mem: 3956
+train: [14]  [ 60/400]  eta: 0:02:12  lr: 0.000099  loss: 0.7938 (0.7894)  grad: 0.0439 (0.0450)  time: 0.3363  data: 0.0026  max mem: 3956
+train: [14]  [ 80/400]  eta: 0:02:00  lr: 0.000098  loss: 0.8004 (0.7949)  grad: 0.0449 (0.0450)  time: 0.3348  data: 0.0035  max mem: 3956
+train: [14]  [100/400]  eta: 0:01:50  lr: 0.000096  loss: 0.8121 (0.7957)  grad: 0.0440 (0.0448)  time: 0.3372  data: 0.0026  max mem: 3956
+train: [14]  [120/400]  eta: 0:01:41  lr: 0.000095  loss: 0.7980 (0.7966)  grad: 0.0440 (0.0449)  time: 0.3313  data: 0.0036  max mem: 3956
+train: [14]  [140/400]  eta: 0:01:33  lr: 0.000093  loss: 0.8036 (0.7982)  grad: 0.0449 (0.0449)  time: 0.3452  data: 0.0025  max mem: 3956
+train: [14]  [160/400]  eta: 0:01:25  lr: 0.000092  loss: 0.7934 (0.7970)  grad: 0.0453 (0.0450)  time: 0.3243  data: 0.0038  max mem: 3956
+train: [14]  [180/400]  eta: 0:01:18  lr: 0.000090  loss: 0.7901 (0.7978)  grad: 0.0446 (0.0449)  time: 0.3623  data: 0.0038  max mem: 3956
+train: [14]  [200/400]  eta: 0:01:12  lr: 0.000089  loss: 0.8125 (0.7999)  grad: 0.0421 (0.0447)  time: 0.4137  data: 0.0043  max mem: 3956
+train: [14]  [220/400]  eta: 0:01:05  lr: 0.000088  loss: 0.7953 (0.7985)  grad: 0.0423 (0.0447)  time: 0.3572  data: 0.0041  max mem: 3956
+train: [14]  [240/400]  eta: 0:00:57  lr: 0.000086  loss: 0.7813 (0.7976)  grad: 0.0449 (0.0448)  time: 0.3396  data: 0.0042  max mem: 3956
+train: [14]  [260/400]  eta: 0:00:50  lr: 0.000085  loss: 0.7813 (0.7957)  grad: 0.0446 (0.0447)  time: 0.3509  data: 0.0043  max mem: 3956
+train: [14]  [280/400]  eta: 0:00:43  lr: 0.000083  loss: 0.7824 (0.7955)  grad: 0.0446 (0.0447)  time: 0.3472  data: 0.0042  max mem: 3956
+train: [14]  [300/400]  eta: 0:00:36  lr: 0.000082  loss: 0.7845 (0.7953)  grad: 0.0456 (0.0448)  time: 0.4841  data: 0.1761  max mem: 3956
+train: [14]  [320/400]  eta: 0:00:29  lr: 0.000081  loss: 0.7845 (0.7953)  grad: 0.0447 (0.0448)  time: 0.3527  data: 0.0287  max mem: 3956
+train: [14]  [340/400]  eta: 0:00:21  lr: 0.000079  loss: 0.7782 (0.7938)  grad: 0.0459 (0.0449)  time: 0.3314  data: 0.0027  max mem: 3956
+train: [14]  [360/400]  eta: 0:00:14  lr: 0.000078  loss: 0.7635 (0.7926)  grad: 0.0462 (0.0450)  time: 0.3565  data: 0.0043  max mem: 3956
+train: [14]  [380/400]  eta: 0:00:07  lr: 0.000076  loss: 0.7800 (0.7929)  grad: 0.0455 (0.0450)  time: 0.3580  data: 0.0043  max mem: 3956
+train: [14]  [399/400]  eta: 0:00:00  lr: 0.000075  loss: 0.7965 (0.7934)  grad: 0.0441 (0.0450)  time: 0.3472  data: 0.0044  max mem: 3956
+train: [14] Total time: 0:02:25 (0.3629 s / it)
+train: [14] Summary: lr: 0.000075  loss: 0.7965 (0.7934)  grad: 0.0441 (0.0450)
+eval (validation): [14]  [ 0/63]  eta: 0:03:26    time: 3.2773  data: 3.0644  max mem: 3956
+eval (validation): [14]  [20/63]  eta: 0:00:21    time: 0.3676  data: 0.0041  max mem: 3956
+eval (validation): [14]  [40/63]  eta: 0:00:09    time: 0.3462  data: 0.0030  max mem: 3956
+eval (validation): [14]  [60/63]  eta: 0:00:01    time: 0.3309  data: 0.0036  max mem: 3956
+eval (validation): [14]  [62/63]  eta: 0:00:00    time: 0.3276  data: 0.0036  max mem: 3956
+eval (validation): [14] Total time: 0:00:25 (0.3987 s / it)
+cv: [14]  best hparam: (26, 1.0) (044) ('044_lr2.6e+01_wd1.0e+00')  loss: 0.113  acc: 0.969  f1: 0.963
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+train: [15]  [  0/400]  eta: 0:22:06  lr: nan  time: 3.3175  data: 3.0274  max mem: 3956
+train: [15]  [ 20/400]  eta: 0:03:15  lr: 0.000074  loss: 0.7923 (0.7923)  grad: 0.0435 (0.0444)  time: 0.3752  data: 0.0046  max mem: 3956
+train: [15]  [ 40/400]  eta: 0:02:34  lr: 0.000072  loss: 0.7747 (0.7812)  grad: 0.0449 (0.0451)  time: 0.3387  data: 0.0038  max mem: 3956
+train: [15]  [ 60/400]  eta: 0:02:18  lr: 0.000071  loss: 0.7713 (0.7835)  grad: 0.0458 (0.0449)  time: 0.3615  data: 0.0043  max mem: 3956
+train: [15]  [ 80/400]  eta: 0:02:06  lr: 0.000070  loss: 0.7679 (0.7773)  grad: 0.0433 (0.0445)  time: 0.3623  data: 0.0042  max mem: 3956
+train: [15]  [100/400]  eta: 0:01:55  lr: 0.000068  loss: 0.7679 (0.7788)  grad: 0.0434 (0.0447)  time: 0.3421  data: 0.0042  max mem: 3956
+train: [15]  [120/400]  eta: 0:01:45  lr: 0.000067  loss: 0.7938 (0.7807)  grad: 0.0452 (0.0447)  time: 0.3331  data: 0.0037  max mem: 3956
+train: [15]  [140/400]  eta: 0:01:37  lr: 0.000066  loss: 0.7823 (0.7826)  grad: 0.0452 (0.0445)  time: 0.3581  data: 0.0042  max mem: 3956
+train: [15]  [160/400]  eta: 0:01:30  lr: 0.000064  loss: 0.7794 (0.7835)  grad: 0.0432 (0.0443)  time: 0.3859  data: 0.0043  max mem: 3956
+train: [15]  [180/400]  eta: 0:01:21  lr: 0.000063  loss: 0.7787 (0.7837)  grad: 0.0426 (0.0441)  time: 0.3448  data: 0.0042  max mem: 3956
+train: [15]  [200/400]  eta: 0:01:14  lr: 0.000062  loss: 0.7873 (0.7849)  grad: 0.0425 (0.0441)  time: 0.3511  data: 0.0038  max mem: 3956
+train: [15]  [220/400]  eta: 0:01:06  lr: 0.000061  loss: 0.7920 (0.7843)  grad: 0.0440 (0.0441)  time: 0.3468  data: 0.0037  max mem: 3956
+train: [15]  [240/400]  eta: 0:00:58  lr: 0.000059  loss: 0.7746 (0.7840)  grad: 0.0440 (0.0442)  time: 0.3292  data: 0.0036  max mem: 3956
+train: [15]  [260/400]  eta: 0:00:50  lr: 0.000058  loss: 0.7772 (0.7847)  grad: 0.0434 (0.0441)  time: 0.3369  data: 0.0037  max mem: 3956
+train: [15]  [280/400]  eta: 0:00:43  lr: 0.000057  loss: 0.7772 (0.7842)  grad: 0.0438 (0.0442)  time: 0.3647  data: 0.0040  max mem: 3956
+train: [15]  [300/400]  eta: 0:00:37  lr: 0.000056  loss: 0.7901 (0.7838)  grad: 0.0438 (0.0442)  time: 0.4781  data: 0.1711  max mem: 3956
+train: [15]  [320/400]  eta: 0:00:29  lr: 0.000054  loss: 0.7889 (0.7839)  grad: 0.0442 (0.0442)  time: 0.3242  data: 0.0041  max mem: 3956
+train: [15]  [340/400]  eta: 0:00:21  lr: 0.000053  loss: 0.7957 (0.7853)  grad: 0.0426 (0.0440)  time: 0.3515  data: 0.0027  max mem: 3956
+train: [15]  [360/400]  eta: 0:00:14  lr: 0.000052  loss: 0.7957 (0.7846)  grad: 0.0413 (0.0440)  time: 0.3555  data: 0.0039  max mem: 3956
+train: [15]  [380/400]  eta: 0:00:07  lr: 0.000051  loss: 0.7946 (0.7859)  grad: 0.0423 (0.0440)  time: 0.3383  data: 0.0040  max mem: 3956
+train: [15]  [399/400]  eta: 0:00:00  lr: 0.000050  loss: 0.7915 (0.7853)  grad: 0.0416 (0.0439)  time: 0.3398  data: 0.0041  max mem: 3956
+train: [15] Total time: 0:02:25 (0.3639 s / it)
+train: [15] Summary: lr: 0.000050  loss: 0.7915 (0.7853)  grad: 0.0416 (0.0439)
+eval (validation): [15]  [ 0/63]  eta: 0:03:07    time: 2.9737  data: 2.7693  max mem: 3956
+eval (validation): [15]  [20/63]  eta: 0:00:19    time: 0.3289  data: 0.0039  max mem: 3956
+eval (validation): [15]  [40/63]  eta: 0:00:09    time: 0.3920  data: 0.0047  max mem: 3956
+eval (validation): [15]  [60/63]  eta: 0:00:01    time: 0.3343  data: 0.0036  max mem: 3956
+eval (validation): [15]  [62/63]  eta: 0:00:00    time: 0.3256  data: 0.0039  max mem: 3956
+eval (validation): [15] Total time: 0:00:25 (0.3975 s / it)
+cv: [15]  best hparam: (26, 1.0) (044) ('044_lr2.6e+01_wd1.0e+00')  loss: 0.113  acc: 0.968  f1: 0.962
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+train: [16]  [  0/400]  eta: 0:21:48  lr: nan  time: 3.2721  data: 2.9934  max mem: 3956
+train: [16]  [ 20/400]  eta: 0:03:12  lr: 0.000048  loss: 0.7286 (0.7557)  grad: 0.0431 (0.0446)  time: 0.3690  data: 0.0046  max mem: 3956
+train: [16]  [ 40/400]  eta: 0:02:33  lr: 0.000047  loss: 0.7665 (0.7678)  grad: 0.0431 (0.0439)  time: 0.3388  data: 0.0032  max mem: 3956
+train: [16]  [ 60/400]  eta: 0:02:15  lr: 0.000046  loss: 0.7708 (0.7713)  grad: 0.0428 (0.0441)  time: 0.3416  data: 0.0045  max mem: 3956
+train: [16]  [ 80/400]  eta: 0:02:03  lr: 0.000045  loss: 0.7571 (0.7707)  grad: 0.0435 (0.0439)  time: 0.3456  data: 0.0042  max mem: 3956
+train: [16]  [100/400]  eta: 0:01:52  lr: 0.000044  loss: 0.7792 (0.7759)  grad: 0.0433 (0.0441)  time: 0.3379  data: 0.0041  max mem: 3956
+train: [16]  [120/400]  eta: 0:01:42  lr: 0.000043  loss: 0.7877 (0.7798)  grad: 0.0434 (0.0439)  time: 0.3278  data: 0.0041  max mem: 3956
+train: [16]  [140/400]  eta: 0:01:35  lr: 0.000042  loss: 0.7804 (0.7770)  grad: 0.0434 (0.0439)  time: 0.3538  data: 0.0041  max mem: 3956
+train: [16]  [160/400]  eta: 0:01:27  lr: 0.000041  loss: 0.7622 (0.7759)  grad: 0.0436 (0.0440)  time: 0.3610  data: 0.0042  max mem: 3956
+train: [16]  [180/400]  eta: 0:01:19  lr: 0.000040  loss: 0.7697 (0.7785)  grad: 0.0432 (0.0438)  time: 0.3480  data: 0.0045  max mem: 3956
+train: [16]  [200/400]  eta: 0:01:12  lr: 0.000039  loss: 0.7680 (0.7769)  grad: 0.0421 (0.0438)  time: 0.3411  data: 0.0043  max mem: 3956
+train: [16]  [220/400]  eta: 0:01:04  lr: 0.000038  loss: 0.7495 (0.7766)  grad: 0.0423 (0.0440)  time: 0.3271  data: 0.0042  max mem: 3956
+train: [16]  [240/400]  eta: 0:00:56  lr: 0.000036  loss: 0.7920 (0.7782)  grad: 0.0431 (0.0439)  time: 0.3310  data: 0.0041  max mem: 3956
+train: [16]  [260/400]  eta: 0:00:49  lr: 0.000035  loss: 0.7923 (0.7772)  grad: 0.0431 (0.0438)  time: 0.3535  data: 0.0044  max mem: 3956
+train: [16]  [280/400]  eta: 0:00:42  lr: 0.000034  loss: 0.7859 (0.7785)  grad: 0.0434 (0.0439)  time: 0.3461  data: 0.0040  max mem: 3956
+train: [16]  [300/400]  eta: 0:00:36  lr: 0.000033  loss: 0.7859 (0.7792)  grad: 0.0423 (0.0438)  time: 0.4823  data: 0.1709  max mem: 3956
+train: [16]  [320/400]  eta: 0:00:28  lr: 0.000032  loss: 0.7700 (0.7796)  grad: 0.0423 (0.0437)  time: 0.3445  data: 0.0042  max mem: 3956
+train: [16]  [340/400]  eta: 0:00:21  lr: 0.000031  loss: 0.7649 (0.7786)  grad: 0.0431 (0.0437)  time: 0.3202  data: 0.0030  max mem: 3956
+train: [16]  [360/400]  eta: 0:00:14  lr: 0.000031  loss: 0.7678 (0.7787)  grad: 0.0420 (0.0437)  time: 0.3538  data: 0.0040  max mem: 3956
+train: [16]  [380/400]  eta: 0:00:07  lr: 0.000030  loss: 0.7678 (0.7778)  grad: 0.0420 (0.0436)  time: 0.3451  data: 0.0039  max mem: 3956
+train: [16]  [399/400]  eta: 0:00:00  lr: 0.000029  loss: 0.7675 (0.7773)  grad: 0.0430 (0.0437)  time: 0.3230  data: 0.0040  max mem: 3956
+train: [16] Total time: 0:02:23 (0.3576 s / it)
+train: [16] Summary: lr: 0.000029  loss: 0.7675 (0.7773)  grad: 0.0430 (0.0437)
+eval (validation): [16]  [ 0/63]  eta: 0:03:09    time: 3.0058  data: 2.8027  max mem: 3956
+eval (validation): [16]  [20/63]  eta: 0:00:19    time: 0.3218  data: 0.0030  max mem: 3956
+eval (validation): [16]  [40/63]  eta: 0:00:08    time: 0.3072  data: 0.0028  max mem: 3956
+eval (validation): [16]  [60/63]  eta: 0:00:01    time: 0.3008  data: 0.0031  max mem: 3956
+eval (validation): [16]  [62/63]  eta: 0:00:00    time: 0.2990  data: 0.0031  max mem: 3956
+eval (validation): [16] Total time: 0:00:22 (0.3569 s / it)
+cv: [16]  best hparam: (43, 1.0) (047) ('047_lr4.3e+01_wd1.0e+00')  loss: 0.110  acc: 0.969  f1: 0.963
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+saving best checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-best.pth
+train: [17]  [  0/400]  eta: 0:21:19  lr: nan  time: 3.1987  data: 2.9249  max mem: 3956
+train: [17]  [ 20/400]  eta: 0:03:12  lr: 0.000028  loss: 0.7841 (0.7928)  grad: 0.0417 (0.0445)  time: 0.3724  data: 0.0030  max mem: 3956
+train: [17]  [ 40/400]  eta: 0:02:30  lr: 0.000027  loss: 0.7618 (0.7750)  grad: 0.0417 (0.0439)  time: 0.3271  data: 0.0039  max mem: 3956
+train: [17]  [ 60/400]  eta: 0:02:11  lr: 0.000026  loss: 0.7593 (0.7744)  grad: 0.0430 (0.0443)  time: 0.3227  data: 0.0041  max mem: 3956
+train: [17]  [ 80/400]  eta: 0:01:58  lr: 0.000025  loss: 0.7657 (0.7750)  grad: 0.0442 (0.0444)  time: 0.3235  data: 0.0040  max mem: 3956
+train: [17]  [100/400]  eta: 0:01:48  lr: 0.000024  loss: 0.7862 (0.7798)  grad: 0.0442 (0.0443)  time: 0.3144  data: 0.0040  max mem: 3956
+train: [17]  [120/400]  eta: 0:01:39  lr: 0.000023  loss: 0.7804 (0.7779)  grad: 0.0440 (0.0444)  time: 0.3288  data: 0.0040  max mem: 3956
+train: [17]  [140/400]  eta: 0:01:31  lr: 0.000023  loss: 0.7804 (0.7796)  grad: 0.0433 (0.0443)  time: 0.3294  data: 0.0042  max mem: 3956
+train: [17]  [160/400]  eta: 0:01:23  lr: 0.000022  loss: 0.7926 (0.7807)  grad: 0.0445 (0.0443)  time: 0.3126  data: 0.0039  max mem: 3956
+train: [17]  [180/400]  eta: 0:01:15  lr: 0.000021  loss: 0.7917 (0.7822)  grad: 0.0428 (0.0440)  time: 0.3289  data: 0.0041  max mem: 3956
+train: [17]  [200/400]  eta: 0:01:08  lr: 0.000020  loss: 0.7720 (0.7795)  grad: 0.0426 (0.0439)  time: 0.3364  data: 0.0041  max mem: 3956
+train: [17]  [220/400]  eta: 0:01:01  lr: 0.000019  loss: 0.7435 (0.7767)  grad: 0.0436 (0.0440)  time: 0.3297  data: 0.0042  max mem: 3956
+train: [17]  [240/400]  eta: 0:00:54  lr: 0.000019  loss: 0.7588 (0.7771)  grad: 0.0435 (0.0439)  time: 0.3327  data: 0.0042  max mem: 3956
+train: [17]  [260/400]  eta: 0:00:48  lr: 0.000018  loss: 0.7883 (0.7771)  grad: 0.0425 (0.0440)  time: 0.3569  data: 0.0042  max mem: 3956
+train: [17]  [280/400]  eta: 0:00:41  lr: 0.000017  loss: 0.7717 (0.7765)  grad: 0.0421 (0.0439)  time: 0.3271  data: 0.0041  max mem: 3956
+train: [17]  [300/400]  eta: 0:00:35  lr: 0.000016  loss: 0.7730 (0.7770)  grad: 0.0421 (0.0438)  time: 0.5107  data: 0.1733  max mem: 3956
+train: [17]  [320/400]  eta: 0:00:28  lr: 0.000016  loss: 0.7749 (0.7773)  grad: 0.0424 (0.0438)  time: 0.3338  data: 0.0041  max mem: 3956
+train: [17]  [340/400]  eta: 0:00:21  lr: 0.000015  loss: 0.7661 (0.7764)  grad: 0.0424 (0.0437)  time: 0.3348  data: 0.0038  max mem: 3956
+train: [17]  [360/400]  eta: 0:00:13  lr: 0.000014  loss: 0.7659 (0.7763)  grad: 0.0418 (0.0436)  time: 0.3298  data: 0.0040  max mem: 3956
+train: [17]  [380/400]  eta: 0:00:06  lr: 0.000014  loss: 0.7699 (0.7759)  grad: 0.0418 (0.0436)  time: 0.3252  data: 0.0040  max mem: 3956
+train: [17]  [399/400]  eta: 0:00:00  lr: 0.000013  loss: 0.7517 (0.7745)  grad: 0.0417 (0.0435)  time: 0.3309  data: 0.0037  max mem: 3956
+train: [17] Total time: 0:02:19 (0.3481 s / it)
+train: [17] Summary: lr: 0.000013  loss: 0.7517 (0.7745)  grad: 0.0417 (0.0435)
+eval (validation): [17]  [ 0/63]  eta: 0:03:23    time: 3.2313  data: 2.9642  max mem: 3956
+eval (validation): [17]  [20/63]  eta: 0:00:19    time: 0.3186  data: 0.0036  max mem: 3956
+eval (validation): [17]  [40/63]  eta: 0:00:09    time: 0.3413  data: 0.0037  max mem: 3956
+eval (validation): [17]  [60/63]  eta: 0:00:01    time: 0.3068  data: 0.0033  max mem: 3956
+eval (validation): [17]  [62/63]  eta: 0:00:00    time: 0.3051  data: 0.0028  max mem: 3956
+eval (validation): [17] Total time: 0:00:23 (0.3728 s / it)
+cv: [17]  best hparam: (31, 1.0) (045) ('045_lr3.1e+01_wd1.0e+00')  loss: 0.111  acc: 0.969  f1: 0.963
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+train: [18]  [  0/400]  eta: 0:20:58  lr: nan  time: 3.1460  data: 2.9365  max mem: 3956
+train: [18]  [ 20/400]  eta: 0:03:08  lr: 0.000012  loss: 0.7415 (0.7573)  grad: 0.0434 (0.0435)  time: 0.3627  data: 0.0035  max mem: 3956
+train: [18]  [ 40/400]  eta: 0:02:32  lr: 0.000012  loss: 0.7465 (0.7566)  grad: 0.0433 (0.0433)  time: 0.3485  data: 0.0032  max mem: 3956
+train: [18]  [ 60/400]  eta: 0:02:14  lr: 0.000011  loss: 0.7615 (0.7659)  grad: 0.0429 (0.0434)  time: 0.3375  data: 0.0040  max mem: 3956
+train: [18]  [ 80/400]  eta: 0:02:03  lr: 0.000011  loss: 0.7818 (0.7722)  grad: 0.0433 (0.0438)  time: 0.3522  data: 0.0039  max mem: 3956
+train: [18]  [100/400]  eta: 0:01:52  lr: 0.000010  loss: 0.7665 (0.7726)  grad: 0.0426 (0.0437)  time: 0.3366  data: 0.0041  max mem: 3956
+train: [18]  [120/400]  eta: 0:01:42  lr: 0.000009  loss: 0.7665 (0.7726)  grad: 0.0424 (0.0438)  time: 0.3170  data: 0.0041  max mem: 3956
+train: [18]  [140/400]  eta: 0:01:33  lr: 0.000009  loss: 0.7719 (0.7728)  grad: 0.0426 (0.0438)  time: 0.3320  data: 0.0042  max mem: 3956
+train: [18]  [160/400]  eta: 0:01:25  lr: 0.000008  loss: 0.7654 (0.7756)  grad: 0.0425 (0.0437)  time: 0.3196  data: 0.0040  max mem: 3956
+train: [18]  [180/400]  eta: 0:01:18  lr: 0.000008  loss: 0.7836 (0.7755)  grad: 0.0434 (0.0437)  time: 0.3532  data: 0.0044  max mem: 3956
+train: [18]  [200/400]  eta: 0:01:10  lr: 0.000007  loss: 0.7803 (0.7756)  grad: 0.0432 (0.0435)  time: 0.3164  data: 0.0040  max mem: 3956
+train: [18]  [220/400]  eta: 0:01:02  lr: 0.000007  loss: 0.7709 (0.7745)  grad: 0.0423 (0.0436)  time: 0.3342  data: 0.0042  max mem: 3956
+train: [18]  [240/400]  eta: 0:00:55  lr: 0.000006  loss: 0.7917 (0.7766)  grad: 0.0430 (0.0435)  time: 0.3360  data: 0.0040  max mem: 3956
+train: [18]  [260/400]  eta: 0:00:48  lr: 0.000006  loss: 0.7887 (0.7754)  grad: 0.0430 (0.0435)  time: 0.3320  data: 0.0038  max mem: 3956
+train: [18]  [280/400]  eta: 0:00:41  lr: 0.000006  loss: 0.7616 (0.7754)  grad: 0.0430 (0.0436)  time: 0.3318  data: 0.0039  max mem: 3956
+train: [18]  [300/400]  eta: 0:00:35  lr: 0.000005  loss: 0.7558 (0.7749)  grad: 0.0441 (0.0436)  time: 0.5152  data: 0.1823  max mem: 3956
+train: [18]  [320/400]  eta: 0:00:28  lr: 0.000005  loss: 0.7524 (0.7732)  grad: 0.0424 (0.0436)  time: 0.3382  data: 0.0045  max mem: 3956
+train: [18]  [340/400]  eta: 0:00:21  lr: 0.000004  loss: 0.7544 (0.7735)  grad: 0.0417 (0.0436)  time: 0.3344  data: 0.0038  max mem: 3956
+train: [18]  [360/400]  eta: 0:00:14  lr: 0.000004  loss: 0.7776 (0.7745)  grad: 0.0429 (0.0436)  time: 0.3562  data: 0.0043  max mem: 3956
+train: [18]  [380/400]  eta: 0:00:07  lr: 0.000004  loss: 0.7952 (0.7751)  grad: 0.0426 (0.0436)  time: 0.3355  data: 0.0038  max mem: 3956
+train: [18]  [399/400]  eta: 0:00:00  lr: 0.000003  loss: 0.7952 (0.7751)  grad: 0.0431 (0.0436)  time: 0.3264  data: 0.0040  max mem: 3956
+train: [18] Total time: 0:02:21 (0.3534 s / it)
+train: [18] Summary: lr: 0.000003  loss: 0.7952 (0.7751)  grad: 0.0431 (0.0436)
+eval (validation): [18]  [ 0/63]  eta: 0:03:14    time: 3.0883  data: 2.8733  max mem: 3956
+eval (validation): [18]  [20/63]  eta: 0:00:20    time: 0.3449  data: 0.0037  max mem: 3956
+eval (validation): [18]  [40/63]  eta: 0:00:09    time: 0.3335  data: 0.0030  max mem: 3956
+eval (validation): [18]  [60/63]  eta: 0:00:01    time: 0.2946  data: 0.0031  max mem: 3956
+eval (validation): [18]  [62/63]  eta: 0:00:00    time: 0.2927  data: 0.0024  max mem: 3956
+eval (validation): [18] Total time: 0:00:23 (0.3718 s / it)
+cv: [18]  best hparam: (31, 1.0) (045) ('045_lr3.1e+01_wd1.0e+00')  loss: 0.111  acc: 0.968  f1: 0.963
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+train: [19]  [  0/400]  eta: 0:21:23  lr: nan  time: 3.2096  data: 2.9818  max mem: 3956
+train: [19]  [ 20/400]  eta: 0:03:01  lr: 0.000003  loss: 0.7778 (0.7916)  grad: 0.0418 (0.0418)  time: 0.3418  data: 0.0035  max mem: 3956
+train: [19]  [ 40/400]  eta: 0:02:29  lr: 0.000003  loss: 0.7922 (0.7855)  grad: 0.0431 (0.0432)  time: 0.3487  data: 0.0029  max mem: 3956
+train: [19]  [ 60/400]  eta: 0:02:11  lr: 0.000002  loss: 0.7604 (0.7734)  grad: 0.0439 (0.0437)  time: 0.3297  data: 0.0040  max mem: 3956
+train: [19]  [ 80/400]  eta: 0:02:00  lr: 0.000002  loss: 0.7658 (0.7765)  grad: 0.0434 (0.0437)  time: 0.3420  data: 0.0042  max mem: 3956
+train: [19]  [100/400]  eta: 0:01:49  lr: 0.000002  loss: 0.7780 (0.7757)  grad: 0.0442 (0.0439)  time: 0.3268  data: 0.0041  max mem: 3956
+train: [19]  [120/400]  eta: 0:01:40  lr: 0.000002  loss: 0.7733 (0.7767)  grad: 0.0450 (0.0440)  time: 0.3268  data: 0.0040  max mem: 3956
+train: [19]  [140/400]  eta: 0:01:32  lr: 0.000001  loss: 0.7816 (0.7766)  grad: 0.0437 (0.0439)  time: 0.3448  data: 0.0042  max mem: 3956
+train: [19]  [160/400]  eta: 0:01:24  lr: 0.000001  loss: 0.7849 (0.7762)  grad: 0.0437 (0.0438)  time: 0.3226  data: 0.0038  max mem: 3956
+train: [19]  [180/400]  eta: 0:01:16  lr: 0.000001  loss: 0.7514 (0.7723)  grad: 0.0427 (0.0437)  time: 0.3236  data: 0.0042  max mem: 3956
+train: [19]  [200/400]  eta: 0:01:09  lr: 0.000001  loss: 0.7541 (0.7719)  grad: 0.0433 (0.0438)  time: 0.3490  data: 0.0044  max mem: 3956
+train: [19]  [220/400]  eta: 0:01:02  lr: 0.000001  loss: 0.7669 (0.7722)  grad: 0.0447 (0.0438)  time: 0.3405  data: 0.0045  max mem: 3956
+train: [19]  [240/400]  eta: 0:00:55  lr: 0.000001  loss: 0.7683 (0.7729)  grad: 0.0435 (0.0436)  time: 0.3357  data: 0.0040  max mem: 3956
+train: [19]  [260/400]  eta: 0:00:48  lr: 0.000000  loss: 0.7688 (0.7730)  grad: 0.0418 (0.0435)  time: 0.3329  data: 0.0043  max mem: 3956
+train: [19]  [280/400]  eta: 0:00:41  lr: 0.000000  loss: 0.7749 (0.7743)  grad: 0.0411 (0.0435)  time: 0.3124  data: 0.0038  max mem: 3956
+train: [19]  [300/400]  eta: 0:00:35  lr: 0.000000  loss: 0.7749 (0.7750)  grad: 0.0422 (0.0435)  time: 0.4755  data: 0.1688  max mem: 3956
+train: [19]  [320/400]  eta: 0:00:28  lr: 0.000000  loss: 0.7764 (0.7754)  grad: 0.0422 (0.0435)  time: 0.3179  data: 0.0039  max mem: 3956
+train: [19]  [340/400]  eta: 0:00:20  lr: 0.000000  loss: 0.7640 (0.7749)  grad: 0.0421 (0.0435)  time: 0.3079  data: 0.0030  max mem: 3956
+train: [19]  [360/400]  eta: 0:00:13  lr: 0.000000  loss: 0.7490 (0.7741)  grad: 0.0423 (0.0435)  time: 0.3257  data: 0.0041  max mem: 3956
+train: [19]  [380/400]  eta: 0:00:06  lr: 0.000000  loss: 0.7497 (0.7743)  grad: 0.0423 (0.0435)  time: 0.3231  data: 0.0040  max mem: 3956
+train: [19]  [399/400]  eta: 0:00:00  lr: 0.000000  loss: 0.7659 (0.7733)  grad: 0.0456 (0.0436)  time: 0.3139  data: 0.0038  max mem: 3956
+train: [19] Total time: 0:02:17 (0.3449 s / it)
+train: [19] Summary: lr: 0.000000  loss: 0.7659 (0.7733)  grad: 0.0456 (0.0436)
+eval (validation): [19]  [ 0/63]  eta: 0:03:06    time: 2.9555  data: 2.7500  max mem: 3956
+eval (validation): [19]  [20/63]  eta: 0:00:19    time: 0.3218  data: 0.0049  max mem: 3956
+eval (validation): [19]  [40/63]  eta: 0:00:08    time: 0.2897  data: 0.0028  max mem: 3956
+eval (validation): [19]  [60/63]  eta: 0:00:01    time: 0.2813  data: 0.0032  max mem: 3956
+eval (validation): [19]  [62/63]  eta: 0:00:00    time: 0.2778  data: 0.0032  max mem: 3956
+eval (validation): [19] Total time: 0:00:21 (0.3444 s / it)
+cv: [19]  best hparam: (22, 1.0) (043) ('043_lr2.2e+01_wd1.0e+00')  loss: 0.113  acc: 0.968  f1: 0.963
+saving checkpoint experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+evaluating last checkpoint: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-last.pth
+eval model info:
+{"score": 0.9682539682539683, "hparam": [22, 1.0], "hparam_id": 43, "epoch": 19, "is_best": false, "best_score": 0.9692460317460317}
+eval (train): [20]  [  0/297]  eta: 0:14:41    time: 2.9676  data: 2.7317  max mem: 3956
+eval (train): [20]  [ 20/297]  eta: 0:02:04    time: 0.3245  data: 0.0029  max mem: 3956
+eval (train): [20]  [ 40/297]  eta: 0:01:41    time: 0.3334  data: 0.0035  max mem: 3956
+eval (train): [20]  [ 60/297]  eta: 0:01:26    time: 0.3085  data: 0.0032  max mem: 3956
+eval (train): [20]  [ 80/297]  eta: 0:01:16    time: 0.3109  data: 0.0033  max mem: 3956
+eval (train): [20]  [100/297]  eta: 0:01:07    time: 0.2965  data: 0.0031  max mem: 3956
+eval (train): [20]  [120/297]  eta: 0:00:59    time: 0.3155  data: 0.0033  max mem: 3956
+eval (train): [20]  [140/297]  eta: 0:00:52    time: 0.3181  data: 0.0036  max mem: 3956
+eval (train): [20]  [160/297]  eta: 0:00:45    time: 0.3209  data: 0.0032  max mem: 3956
+eval (train): [20]  [180/297]  eta: 0:00:38    time: 0.3153  data: 0.0031  max mem: 3956
+eval (train): [20]  [200/297]  eta: 0:00:32    time: 0.3356  data: 0.0036  max mem: 3956
+eval (train): [20]  [220/297]  eta: 0:00:25    time: 0.3341  data: 0.0036  max mem: 3956
+eval (train): [20]  [240/297]  eta: 0:00:18    time: 0.3053  data: 0.0033  max mem: 3956
+eval (train): [20]  [260/297]  eta: 0:00:12    time: 0.3222  data: 0.0036  max mem: 3956
+eval (train): [20]  [280/297]  eta: 0:00:05    time: 0.3157  data: 0.0036  max mem: 3956
+eval (train): [20]  [296/297]  eta: 0:00:00    time: 0.2879  data: 0.0034  max mem: 3956
+eval (train): [20] Total time: 0:01:37 (0.3276 s / it)
+eval (validation): [20]  [ 0/63]  eta: 0:03:16    time: 3.1202  data: 2.9159  max mem: 3956
+eval (validation): [20]  [20/63]  eta: 0:00:19    time: 0.3176  data: 0.0035  max mem: 3956
+eval (validation): [20]  [40/63]  eta: 0:00:08    time: 0.3193  data: 0.0031  max mem: 3956
+eval (validation): [20]  [60/63]  eta: 0:00:01    time: 0.2951  data: 0.0034  max mem: 3956
+eval (validation): [20]  [62/63]  eta: 0:00:00    time: 0.2932  data: 0.0033  max mem: 3956
+eval (validation): [20] Total time: 0:00:22 (0.3596 s / it)
+eval (test): [20]  [ 0/79]  eta: 0:04:46    time: 3.6329  data: 3.3887  max mem: 3956
+eval (test): [20]  [20/79]  eta: 0:00:26    time: 0.2967  data: 0.0027  max mem: 3956
+eval (test): [20]  [40/79]  eta: 0:00:14    time: 0.3064  data: 0.0031  max mem: 3956
+eval (test): [20]  [60/79]  eta: 0:00:06    time: 0.3029  data: 0.0028  max mem: 3956
+eval (test): [20]  [78/79]  eta: 0:00:00    time: 0.2919  data: 0.0033  max mem: 3956
+eval (test): [20] Total time: 0:00:27 (0.3453 s / it)
+evaluating best checkpoint: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/hcpya_task21__reg__linear/checkpoint-best.pth
+eval model info:
+{"score": 0.9692460317460317, "hparam": [43, 1.0], "hparam_id": 47, "epoch": 16, "is_best": true, "best_score": 0.9692460317460317}
+eval (train): [20]  [  0/297]  eta: 0:15:26    time: 3.1198  data: 2.8595  max mem: 3956
+eval (train): [20]  [ 20/297]  eta: 0:02:10    time: 0.3380  data: 0.0049  max mem: 3956
+eval (train): [20]  [ 40/297]  eta: 0:01:40    time: 0.3064  data: 0.0029  max mem: 3956
+eval (train): [20]  [ 60/297]  eta: 0:01:25    time: 0.3007  data: 0.0032  max mem: 3956
+eval (train): [20]  [ 80/297]  eta: 0:01:16    time: 0.3256  data: 0.0034  max mem: 3956
+eval (train): [20]  [100/297]  eta: 0:01:09    time: 0.3452  data: 0.0036  max mem: 3956
+eval (train): [20]  [120/297]  eta: 0:01:00    time: 0.3122  data: 0.0033  max mem: 3956
+eval (train): [20]  [140/297]  eta: 0:00:53    time: 0.3152  data: 0.0035  max mem: 3956
+eval (train): [20]  [160/297]  eta: 0:00:46    time: 0.3122  data: 0.0032  max mem: 3956
+eval (train): [20]  [180/297]  eta: 0:00:39    time: 0.3120  data: 0.0034  max mem: 3956
+eval (train): [20]  [200/297]  eta: 0:00:31    time: 0.2892  data: 0.0030  max mem: 3956
+eval (train): [20]  [220/297]  eta: 0:00:25    time: 0.3180  data: 0.0033  max mem: 3956
+eval (train): [20]  [240/297]  eta: 0:00:18    time: 0.3166  data: 0.0034  max mem: 3956
+eval (train): [20]  [260/297]  eta: 0:00:12    time: 0.3461  data: 0.0039  max mem: 3956
+eval (train): [20]  [280/297]  eta: 0:00:05    time: 0.3614  data: 0.0036  max mem: 3956
+eval (train): [20]  [296/297]  eta: 0:00:00    time: 0.3206  data: 0.0036  max mem: 3956
+eval (train): [20] Total time: 0:01:38 (0.3327 s / it)
+eval (validation): [20]  [ 0/63]  eta: 0:03:09    time: 3.0022  data: 2.7954  max mem: 3956
+eval (validation): [20]  [20/63]  eta: 0:00:19    time: 0.3182  data: 0.0035  max mem: 3956
+eval (validation): [20]  [40/63]  eta: 0:00:09    time: 0.3481  data: 0.0031  max mem: 3956
+eval (validation): [20]  [60/63]  eta: 0:00:01    time: 0.3062  data: 0.0033  max mem: 3956
+eval (validation): [20]  [62/63]  eta: 0:00:00    time: 0.3012  data: 0.0035  max mem: 3956
+eval (validation): [20] Total time: 0:00:23 (0.3703 s / it)
+eval (test): [20]  [ 0/79]  eta: 0:03:48    time: 2.8983  data: 2.6931  max mem: 3956
+eval (test): [20]  [20/79]  eta: 0:00:25    time: 0.3148  data: 0.0026  max mem: 3956
+eval (test): [20]  [40/79]  eta: 0:00:15    time: 0.3719  data: 0.0036  max mem: 3956
+eval (test): [20]  [60/79]  eta: 0:00:07    time: 0.3239  data: 0.0034  max mem: 3956
+eval (test): [20]  [78/79]  eta: 0:00:00    time: 0.2919  data: 0.0031  max mem: 3956
+eval (test): [20] Total time: 0:00:28 (0.3624 s / it)
+eval results:
+
+| model    | repr   | clf    | dataset      | ckpt   |   epoch |     lr |   wd |   hparam_id | hparam    | split      |     loss |     acc |    acc_std |      f1 |     f1_std |
+|:---------|:-------|:-------|:-------------|:-------|--------:|-------:|-----:|------------:|:----------|:-----------|---------:|--------:|-----------:|--------:|-----------:|
+| flat_mae | reg    | linear | hcpya_task21 | best   |      16 | 0.0129 | 0.05 |          47 | [43, 1.0] | train      | 0.045581 | 0.99326 | 0.00057597 | 0.99458 | 0.00049849 |
+| flat_mae | reg    | linear | hcpya_task21 | best   |      16 | 0.0129 | 0.05 |          47 | [43, 1.0] | validation | 0.11012  | 0.96925 | 0.0026725  | 0.96344 | 0.0036046  |
+| flat_mae | reg    | linear | hcpya_task21 | best   |      16 | 0.0129 | 0.05 |          47 | [43, 1.0] | test       | 0.13361  | 0.9627  | 0.0025978  | 0.95678 | 0.0034142  |
+
+
+done! total time: 1:02:02

decoders/crossreg_reg4/eval_v2/nsd_cococlip__patch__attn/config.yaml

@@ -0,0 +1,96 @@
+output_root: experiments/decoders/output
+name_prefix: eval_probe
+remote_root: null
+notes: decoder ablations crossreg_reg4; eval v2 (nsd_cococlip patch attn)
+model_kwargs:
+  ckpt_path: experiments/decoders/output/decoders/crossreg_reg4/pretrain/checkpoint-last.pth
+dataset_kwargs: {}
+classifier_kwargs:
+  embed_dim: null
+  dropout: 0.0
+  xavier_init: false
+  norm: false
+lr_scale_grid:
+- 0.02
+- 0.023
+- 0.028
+- 0.033
+- 0.038
+- 0.045
+- 0.053
+- 0.062
+- 0.074
+- 0.087
+- 0.1
+- 0.12
+- 0.14
+- 0.17
+- 0.2
+- 0.23
+- 0.27
+- 0.32
+- 0.38
+- 0.44
+- 0.52
+- 0.61
+- 0.72
+- 0.85
+- 1
+- 1.2
+- 1.4
+- 1.6
+- 1.9
+- 2.3
+- 2.7
+- 3.1
+- 3.7
+- 4.3
+- 5.1
+- 6
+- 7.1
+- 8.3
+- 9.8
+- 12
+- 14
+- 16
+- 19
+- 22
+- 26
+- 31
+- 36
+- 43
+- 50
+wd_scale_grid:
+- 1.0
+num_workers: 8
+prefetch_factor: null
+balanced_sampling: false
+epochs: 20
+steps_per_epoch: 200
+batch_size: 64
+accum_iter: 2
+lr: 0.0003
+warmup_epochs: 5
+no_decay: false
+weight_decay: 0.05
+clip_grad: 1.0
+metrics:
+- acc
+- f1
+cv_metric: acc
+early_stopping: true
+amp: true
+device: cuda
+seed: 4466
+debug: false
+wandb: false
+wandb_entity: null
+wandb_project: fMRI-fm-eval
+name: decoders/crossreg_reg4/eval_v2/nsd_cococlip__patch__attn
+model: flat_mae
+representation: patch
+classifier: attn
+dataset: nsd_cococlip
+distributed: false
+output_dir: experiments/decoders/output/decoders/crossreg_reg4/eval_v2/nsd_cococlip__patch__attn
+remote_dir: null

decoders/crossreg_reg4/eval_v2/nsd_cococlip__patch__attn/eval_log.json

@@ -0,0 +1 @@
+{"eval/epoch": 7, "eval/id_best": 24, "eval/lr_best": 0.0003, "eval/wd_best": 0.05, "eval/train/loss": 2.0318095684051514, "eval/train/acc": 0.3866744521958265, "eval/train/acc_std": 0.002387726062369834, "eval/train/f1": 0.3318689003317408, "eval/train/f1_std": 0.0024548955470957705, "eval/validation/loss": 2.3388521671295166, "eval/validation/acc": 0.29992617201919525, "eval/validation/acc_std": 0.005267338133834117, "eval/validation/f1": 0.23211007531295333, "eval/validation/f1_std": 0.00474455591644188, "eval/test/loss": 2.3151602745056152, "eval/test/acc": 0.3096474953617811, "eval/test/acc_std": 0.005315251640359488, "eval/test/f1": 0.2452521830180977, "eval/test/f1_std": 0.00516823939276546, "eval/testid/loss": 2.2194652557373047, "eval/testid/acc": 0.32716406400616926, "eval/testid/acc_std": 0.005763874544401807, "eval/testid/f1": 0.2692919280949952, "eval/testid/f1_std": 0.005573398022299934}

decoders/crossreg_reg4/eval_v2/nsd_cococlip__patch__attn/eval_log_best.json

@@ -0,0 +1 @@
+{"eval/best/epoch": 7, "eval/best/id_best": 24, "eval/best/lr_best": 0.0003, "eval/best/wd_best": 0.05, "eval/best/train/loss": 2.0318095684051514, "eval/best/train/acc": 0.3866744521958265, "eval/best/train/acc_std": 0.002387726062369834, "eval/best/train/f1": 0.3318689003317408, "eval/best/train/f1_std": 0.0024548955470957705, "eval/best/validation/loss": 2.3388521671295166, "eval/best/validation/acc": 0.29992617201919525, "eval/best/validation/acc_std": 0.005267338133834117, "eval/best/validation/f1": 0.23211007531295333, "eval/best/validation/f1_std": 0.00474455591644188, "eval/best/test/loss": 2.3151602745056152, "eval/best/test/acc": 0.3096474953617811, "eval/best/test/acc_std": 0.005315251640359488, "eval/best/test/f1": 0.2452521830180977, "eval/best/test/f1_std": 0.00516823939276546, "eval/best/testid/loss": 2.2194652557373047, "eval/best/testid/acc": 0.32716406400616926, "eval/best/testid/acc_std": 0.005763874544401807, "eval/best/testid/f1": 0.2692919280949952, "eval/best/testid/f1_std": 0.005573398022299934}