diffcse-roberta-base-trans / trainer_state.json
voidism's picture
first commit
fcdec15
{
"best_metric": 0.8359799918026352,
"best_model_checkpoint": "result/simcse-celectra-amlp-dmlp-bs128-lr2e-6-mask0.40-elew0.01-roberta-base",
"epoch": 1.0,
"global_step": 7813,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"electra_acc": 0.2448,
"electra_fix_acc": 0.006,
"electra_rep_acc": 0.9912,
"epoch": 0.0,
"learning_rate": 1.9997440163829513e-06,
"loss": 8.6888,
"neg_sim": 0.2041,
"pos_sim": 0.3651,
"step": 1
},
{
"epoch": 0.02,
"eval_avg_sts": 0.7368351609735859,
"eval_sickr_spearman": 0.6964049570574558,
"eval_stsb_spearman": 0.7772653648897162,
"step": 125
},
{
"epoch": 0.03,
"eval_avg_sts": 0.7530561409181189,
"eval_sickr_spearman": 0.707297882521504,
"eval_stsb_spearman": 0.7988143993147339,
"step": 250
},
{
"epoch": 0.05,
"eval_avg_sts": 0.7605110871922394,
"eval_sickr_spearman": 0.7124740982612865,
"eval_stsb_spearman": 0.8085480761231921,
"step": 375
},
{
"electra_acc": 0.7243,
"electra_fix_acc": 0.8981,
"electra_rep_acc": 0.2064,
"epoch": 0.06,
"learning_rate": 1.8720081914757456e-06,
"loss": 0.6062,
"neg_sim": 0.0055,
"pos_sim": 0.6593,
"step": 500
},
{
"epoch": 0.06,
"eval_avg_sts": 0.763939631028596,
"eval_sickr_spearman": 0.7155249377624325,
"eval_stsb_spearman": 0.8123543242947595,
"step": 500
},
{
"epoch": 0.08,
"eval_avg_sts": 0.7670128200581756,
"eval_sickr_spearman": 0.7176028112417467,
"eval_stsb_spearman": 0.8164228288746044,
"step": 625
},
{
"epoch": 0.1,
"eval_avg_sts": 0.7683647856842559,
"eval_sickr_spearman": 0.7182528161375087,
"eval_stsb_spearman": 0.8184767552310033,
"step": 750
},
{
"epoch": 0.11,
"eval_avg_sts": 0.7703405632893722,
"eval_sickr_spearman": 0.7201300637061159,
"eval_stsb_spearman": 0.8205510628726285,
"step": 875
},
{
"electra_acc": 0.7954,
"electra_fix_acc": 0.9411,
"electra_rep_acc": 0.3621,
"epoch": 0.13,
"learning_rate": 1.744016382951491e-06,
"loss": 0.0043,
"neg_sim": -0.0066,
"pos_sim": 0.7176,
"step": 1000
},
{
"epoch": 0.13,
"eval_avg_sts": 0.7716371834988085,
"eval_sickr_spearman": 0.7207027865596639,
"eval_stsb_spearman": 0.8225715804379532,
"step": 1000
},
{
"epoch": 0.14,
"eval_avg_sts": 0.7719569589534554,
"eval_sickr_spearman": 0.7195236710504597,
"eval_stsb_spearman": 0.8243902468564511,
"step": 1125
},
{
"epoch": 0.16,
"eval_avg_sts": 0.7725905596083784,
"eval_sickr_spearman": 0.7204897205936844,
"eval_stsb_spearman": 0.8246913986230725,
"step": 1250
},
{
"epoch": 0.18,
"eval_avg_sts": 0.7740128875580021,
"eval_sickr_spearman": 0.7209413090094056,
"eval_stsb_spearman": 0.8270844661065986,
"step": 1375
},
{
"electra_acc": 0.806,
"electra_fix_acc": 0.9446,
"electra_rep_acc": 0.3921,
"epoch": 0.19,
"learning_rate": 1.6160245744272365e-06,
"loss": 0.0039,
"neg_sim": -0.0067,
"pos_sim": 0.7288,
"step": 1500
},
{
"epoch": 0.19,
"eval_avg_sts": 0.7747104419219604,
"eval_sickr_spearman": 0.721527768757965,
"eval_stsb_spearman": 0.8278931150859556,
"step": 1500
},
{
"epoch": 0.21,
"eval_avg_sts": 0.7750669451240817,
"eval_sickr_spearman": 0.722242135329654,
"eval_stsb_spearman": 0.8278917549185095,
"step": 1625
},
{
"epoch": 0.22,
"eval_avg_sts": 0.7751917311744843,
"eval_sickr_spearman": 0.7214598527805315,
"eval_stsb_spearman": 0.8289236095684369,
"step": 1750
},
{
"epoch": 0.24,
"eval_avg_sts": 0.775110379872699,
"eval_sickr_spearman": 0.7211095139266419,
"eval_stsb_spearman": 0.8291112458187563,
"step": 1875
},
{
"electra_acc": 0.8111,
"electra_fix_acc": 0.9455,
"electra_rep_acc": 0.4077,
"epoch": 0.26,
"learning_rate": 1.4880327659029822e-06,
"loss": 0.0037,
"neg_sim": -0.0067,
"pos_sim": 0.7361,
"step": 2000
},
{
"epoch": 0.26,
"eval_avg_sts": 0.7752389174925427,
"eval_sickr_spearman": 0.7207011535022151,
"eval_stsb_spearman": 0.8297766814828704,
"step": 2000
},
{
"epoch": 0.27,
"eval_avg_sts": 0.7759291537103485,
"eval_sickr_spearman": 0.7218097593545081,
"eval_stsb_spearman": 0.8300485480661889,
"step": 2125
},
{
"epoch": 0.29,
"eval_avg_sts": 0.7759100192724719,
"eval_sickr_spearman": 0.7214468843831432,
"eval_stsb_spearman": 0.8303731541618006,
"step": 2250
},
{
"epoch": 0.3,
"eval_avg_sts": 0.77580052054564,
"eval_sickr_spearman": 0.7208077825474077,
"eval_stsb_spearman": 0.8307932585438723,
"step": 2375
},
{
"electra_acc": 0.8137,
"electra_fix_acc": 0.9463,
"electra_rep_acc": 0.4184,
"epoch": 0.32,
"learning_rate": 1.3600409573787276e-06,
"loss": 0.0036,
"neg_sim": -0.0067,
"pos_sim": 0.7422,
"step": 2500
},
{
"epoch": 0.32,
"eval_avg_sts": 0.7764057696201115,
"eval_sickr_spearman": 0.7207564372999703,
"eval_stsb_spearman": 0.8320551019402528,
"step": 2500
},
{
"epoch": 0.34,
"eval_avg_sts": 0.7771167451325478,
"eval_sickr_spearman": 0.7216867036726237,
"eval_stsb_spearman": 0.8325467865924719,
"step": 2625
},
{
"epoch": 0.35,
"eval_avg_sts": 0.777409082133846,
"eval_sickr_spearman": 0.7218441496231378,
"eval_stsb_spearman": 0.8329740146445542,
"step": 2750
},
{
"epoch": 0.37,
"eval_avg_sts": 0.7776325118235485,
"eval_sickr_spearman": 0.722264277667417,
"eval_stsb_spearman": 0.8330007459796801,
"step": 2875
},
{
"electra_acc": 0.8163,
"electra_fix_acc": 0.9464,
"electra_rep_acc": 0.4284,
"epoch": 0.38,
"learning_rate": 1.2320491488544733e-06,
"loss": 0.0034,
"neg_sim": -0.0068,
"pos_sim": 0.7471,
"step": 3000
},
{
"epoch": 0.38,
"eval_avg_sts": 0.7777138423021386,
"eval_sickr_spearman": 0.7227461256770441,
"eval_stsb_spearman": 0.832681558927233,
"step": 3000
},
{
"epoch": 0.4,
"eval_avg_sts": 0.7781702830872564,
"eval_sickr_spearman": 0.7235760070476919,
"eval_stsb_spearman": 0.832764559126821,
"step": 3125
},
{
"epoch": 0.42,
"eval_avg_sts": 0.7788742472264717,
"eval_sickr_spearman": 0.7235008864050426,
"eval_stsb_spearman": 0.8342476080479008,
"step": 3250
},
{
"epoch": 0.43,
"eval_avg_sts": 0.7786540883853343,
"eval_sickr_spearman": 0.7228896906392426,
"eval_stsb_spearman": 0.8344184861314262,
"step": 3375
},
{
"electra_acc": 0.8185,
"electra_fix_acc": 0.9468,
"electra_rep_acc": 0.4365,
"epoch": 0.45,
"learning_rate": 1.104057340330219e-06,
"loss": 0.0034,
"neg_sim": -0.0068,
"pos_sim": 0.7518,
"step": 3500
},
{
"epoch": 0.45,
"eval_avg_sts": 0.7785390972296262,
"eval_sickr_spearman": 0.7228380572051966,
"eval_stsb_spearman": 0.8342401372540559,
"step": 3500
},
{
"epoch": 0.46,
"eval_avg_sts": 0.778413022922875,
"eval_sickr_spearman": 0.7219714320419488,
"eval_stsb_spearman": 0.8348546138038011,
"step": 3625
},
{
"epoch": 0.48,
"eval_avg_sts": 0.778801601085773,
"eval_sickr_spearman": 0.722748959512029,
"eval_stsb_spearman": 0.8348542426595171,
"step": 3750
},
{
"epoch": 0.5,
"eval_avg_sts": 0.7782529644249387,
"eval_sickr_spearman": 0.7217000082877221,
"eval_stsb_spearman": 0.8348059205621555,
"step": 3875
},
{
"electra_acc": 0.8202,
"electra_fix_acc": 0.9468,
"electra_rep_acc": 0.4426,
"epoch": 0.51,
"learning_rate": 9.760655318059644e-07,
"loss": 0.0033,
"neg_sim": -0.0068,
"pos_sim": 0.7559,
"step": 4000
},
{
"epoch": 0.51,
"eval_avg_sts": 0.7782394657347366,
"eval_sickr_spearman": 0.7217301237883238,
"eval_stsb_spearman": 0.8347488076811493,
"step": 4000
},
{
"epoch": 0.53,
"eval_avg_sts": 0.7782895079814026,
"eval_sickr_spearman": 0.7220120663537655,
"eval_stsb_spearman": 0.8345669496090398,
"step": 4125
},
{
"epoch": 0.54,
"eval_avg_sts": 0.778080439519264,
"eval_sickr_spearman": 0.7216004398144409,
"eval_stsb_spearman": 0.834560439224087,
"step": 4250
},
{
"epoch": 0.56,
"eval_avg_sts": 0.7782284276808015,
"eval_sickr_spearman": 0.7216616794687745,
"eval_stsb_spearman": 0.8347951758928285,
"step": 4375
},
{
"electra_acc": 0.8217,
"electra_fix_acc": 0.9477,
"electra_rep_acc": 0.4458,
"epoch": 0.58,
"learning_rate": 8.480737232817099e-07,
"loss": 0.0032,
"neg_sim": -0.0068,
"pos_sim": 0.7596,
"step": 4500
},
{
"epoch": 0.58,
"eval_avg_sts": 0.7781400153641456,
"eval_sickr_spearman": 0.7214844927355691,
"eval_stsb_spearman": 0.8347955379927221,
"step": 4500
},
{
"epoch": 0.59,
"eval_avg_sts": 0.7783651714396398,
"eval_sickr_spearman": 0.721677865949959,
"eval_stsb_spearman": 0.8350524769293207,
"step": 4625
},
{
"epoch": 0.61,
"eval_avg_sts": 0.7779810461862591,
"eval_sickr_spearman": 0.7215376151337598,
"eval_stsb_spearman": 0.8344244772387583,
"step": 4750
},
{
"epoch": 0.62,
"eval_avg_sts": 0.7774863632137369,
"eval_sickr_spearman": 0.7212506225972007,
"eval_stsb_spearman": 0.833722103830273,
"step": 4875
},
{
"electra_acc": 0.8222,
"electra_fix_acc": 0.9476,
"electra_rep_acc": 0.4478,
"epoch": 0.64,
"learning_rate": 7.200819147574555e-07,
"loss": 0.0033,
"neg_sim": -0.0068,
"pos_sim": 0.7614,
"step": 5000
},
{
"epoch": 0.64,
"eval_avg_sts": 0.7780441928626822,
"eval_sickr_spearman": 0.7222703776172996,
"eval_stsb_spearman": 0.833818008108065,
"step": 5000
},
{
"epoch": 0.66,
"eval_avg_sts": 0.7772624437857663,
"eval_sickr_spearman": 0.7211198886445525,
"eval_stsb_spearman": 0.8334049989269801,
"step": 5125
},
{
"epoch": 0.67,
"eval_avg_sts": 0.777433289264948,
"eval_sickr_spearman": 0.7208792047952461,
"eval_stsb_spearman": 0.83398737373465,
"step": 5250
},
{
"epoch": 0.69,
"eval_avg_sts": 0.778471716959233,
"eval_sickr_spearman": 0.7219667730251094,
"eval_stsb_spearman": 0.8349766608933566,
"step": 5375
},
{
"electra_acc": 0.824,
"electra_fix_acc": 0.9479,
"electra_rep_acc": 0.4538,
"epoch": 0.7,
"learning_rate": 5.920901062332011e-07,
"loss": 0.0033,
"neg_sim": -0.0068,
"pos_sim": 0.7634,
"step": 5500
},
{
"epoch": 0.7,
"eval_avg_sts": 0.7783835165868276,
"eval_sickr_spearman": 0.7220774366828226,
"eval_stsb_spearman": 0.8346895964908326,
"step": 5500
},
{
"epoch": 0.72,
"eval_avg_sts": 0.7777215174417764,
"eval_sickr_spearman": 0.7211491395853283,
"eval_stsb_spearman": 0.8342938952982245,
"step": 5625
},
{
"epoch": 0.74,
"eval_avg_sts": 0.7783006770868415,
"eval_sickr_spearman": 0.7218946783418506,
"eval_stsb_spearman": 0.8347066758318323,
"step": 5750
},
{
"epoch": 0.75,
"eval_avg_sts": 0.7786525912567952,
"eval_sickr_spearman": 0.7218885303608666,
"eval_stsb_spearman": 0.8354166521527239,
"step": 5875
},
{
"electra_acc": 0.8245,
"electra_fix_acc": 0.9479,
"electra_rep_acc": 0.4557,
"epoch": 0.77,
"learning_rate": 4.640982977089466e-07,
"loss": 0.0033,
"neg_sim": -0.0068,
"pos_sim": 0.7639,
"step": 6000
},
{
"epoch": 0.77,
"eval_avg_sts": 0.7786598191346372,
"eval_sickr_spearman": 0.7214591323140099,
"eval_stsb_spearman": 0.8358605059552645,
"step": 6000
},
{
"epoch": 0.78,
"eval_avg_sts": 0.7785391387872032,
"eval_sickr_spearman": 0.7212359798167284,
"eval_stsb_spearman": 0.8358422977576779,
"step": 6125
},
{
"epoch": 0.8,
"eval_avg_sts": 0.7780962443354171,
"eval_sickr_spearman": 0.7207366965172793,
"eval_stsb_spearman": 0.835455792153555,
"step": 6250
},
{
"epoch": 0.82,
"eval_avg_sts": 0.7787170644410477,
"eval_sickr_spearman": 0.7214541370794603,
"eval_stsb_spearman": 0.8359799918026352,
"step": 6375
},
{
"electra_acc": 0.8249,
"electra_fix_acc": 0.9477,
"electra_rep_acc": 0.458,
"epoch": 0.83,
"learning_rate": 3.361064891846921e-07,
"loss": 0.0034,
"neg_sim": -0.0068,
"pos_sim": 0.765,
"step": 6500
},
{
"epoch": 0.83,
"eval_avg_sts": 0.7787359930993909,
"eval_sickr_spearman": 0.7215321395881958,
"eval_stsb_spearman": 0.835939846610586,
"step": 6500
},
{
"epoch": 0.85,
"eval_avg_sts": 0.778662230635321,
"eval_sickr_spearman": 0.7217551960232744,
"eval_stsb_spearman": 0.8355692652473675,
"step": 6625
},
{
"epoch": 0.86,
"eval_avg_sts": 0.7788844288300487,
"eval_sickr_spearman": 0.7221871330355174,
"eval_stsb_spearman": 0.8355817246245801,
"step": 6750
},
{
"epoch": 0.88,
"eval_avg_sts": 0.7789522767655956,
"eval_sickr_spearman": 0.722305728507958,
"eval_stsb_spearman": 0.8355988250232333,
"step": 6875
},
{
"electra_acc": 0.8251,
"electra_fix_acc": 0.9476,
"electra_rep_acc": 0.4589,
"epoch": 0.9,
"learning_rate": 2.0811468066043772e-07,
"loss": 0.0034,
"neg_sim": -0.0069,
"pos_sim": 0.7655,
"step": 7000
},
{
"epoch": 0.9,
"eval_avg_sts": 0.7787617595098588,
"eval_sickr_spearman": 0.721966628931805,
"eval_stsb_spearman": 0.8355568900879128,
"step": 7000
},
{
"epoch": 0.91,
"eval_avg_sts": 0.7786478565950568,
"eval_sickr_spearman": 0.7216480386359659,
"eval_stsb_spearman": 0.8356476745541477,
"step": 7125
},
{
"epoch": 0.93,
"eval_avg_sts": 0.7787981847358625,
"eval_sickr_spearman": 0.7219965523080009,
"eval_stsb_spearman": 0.8355998171637242,
"step": 7250
},
{
"epoch": 0.94,
"eval_avg_sts": 0.7788565954562792,
"eval_sickr_spearman": 0.7221892050558691,
"eval_stsb_spearman": 0.8355239858566892,
"step": 7375
},
{
"electra_acc": 0.8254,
"electra_fix_acc": 0.9477,
"electra_rep_acc": 0.4607,
"epoch": 0.96,
"learning_rate": 8.012287213618328e-08,
"loss": 0.0032,
"neg_sim": -0.0069,
"pos_sim": 0.7665,
"step": 7500
},
{
"epoch": 0.96,
"eval_avg_sts": 0.7788488501171182,
"eval_sickr_spearman": 0.7221623076390639,
"eval_stsb_spearman": 0.8355353925951726,
"step": 7500
},
{
"epoch": 0.98,
"eval_avg_sts": 0.7788989946596425,
"eval_sickr_spearman": 0.7222944411991201,
"eval_stsb_spearman": 0.8355035481201649,
"step": 7625
},
{
"epoch": 0.99,
"eval_avg_sts": 0.7788451325075824,
"eval_sickr_spearman": 0.7221886767137534,
"eval_stsb_spearman": 0.8355015883014113,
"step": 7750
},
{
"epoch": 1.0,
"step": 7813,
"train_runtime": 7033.0789,
"train_samples_per_second": 1.111
}
],
"max_steps": 7813,
"num_train_epochs": 1,
"total_flos": 285611882411596800,
"trial_name": null,
"trial_params": null
}