deberta-v3-small-finetuned-mnli / trainer_state.json
mrm8488's picture
Initial commit from mrm8488
15d4c2c
raw
history blame
27.6 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 3.0,
"global_step": 73632,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.04,
"learning_rate": 3e-05,
"loss": 0.7773,
"step": 1000
},
{
"epoch": 0.04,
"eval_accuracy": 0.7983698420784513,
"eval_loss": 0.5241270065307617,
"eval_runtime": 42.0562,
"eval_samples_per_second": 233.378,
"eval_steps_per_second": 29.175,
"step": 1000
},
{
"epoch": 0.08,
"learning_rate": 2.9586958916180196e-05,
"loss": 0.546,
"step": 2000
},
{
"epoch": 0.08,
"eval_accuracy": 0.8193581253183903,
"eval_loss": 0.46290943026542664,
"eval_runtime": 42.1932,
"eval_samples_per_second": 232.621,
"eval_steps_per_second": 29.081,
"step": 2000
},
{
"epoch": 0.12,
"learning_rate": 2.9173917832360394e-05,
"loss": 0.5032,
"step": 3000
},
{
"epoch": 0.12,
"eval_accuracy": 0.8274070300560367,
"eval_loss": 0.4703749418258667,
"eval_runtime": 42.1306,
"eval_samples_per_second": 232.966,
"eval_steps_per_second": 29.124,
"step": 3000
},
{
"epoch": 0.16,
"learning_rate": 2.876087674854059e-05,
"loss": 0.4711,
"step": 4000
},
{
"epoch": 0.16,
"eval_accuracy": 0.8354559347936832,
"eval_loss": 0.4382944405078888,
"eval_runtime": 41.9141,
"eval_samples_per_second": 234.169,
"eval_steps_per_second": 29.274,
"step": 4000
},
{
"epoch": 0.2,
"learning_rate": 2.8347835664720783e-05,
"loss": 0.473,
"step": 5000
},
{
"epoch": 0.2,
"eval_accuracy": 0.8304635761589404,
"eval_loss": 0.4652259051799774,
"eval_runtime": 42.5911,
"eval_samples_per_second": 230.447,
"eval_steps_per_second": 28.809,
"step": 5000
},
{
"epoch": 0.24,
"learning_rate": 2.793479458090098e-05,
"loss": 0.4619,
"step": 6000
},
{
"epoch": 0.24,
"eval_accuracy": 0.8386143657666837,
"eval_loss": 0.42338472604751587,
"eval_runtime": 42.1336,
"eval_samples_per_second": 232.949,
"eval_steps_per_second": 29.122,
"step": 6000
},
{
"epoch": 0.29,
"learning_rate": 2.7521753497081176e-05,
"loss": 0.4542,
"step": 7000
},
{
"epoch": 0.29,
"eval_accuracy": 0.8349465104431992,
"eval_loss": 0.48245278000831604,
"eval_runtime": 42.4819,
"eval_samples_per_second": 231.04,
"eval_steps_per_second": 28.883,
"step": 7000
},
{
"epoch": 0.33,
"learning_rate": 2.7108712413261374e-05,
"loss": 0.4468,
"step": 8000
},
{
"epoch": 0.33,
"eval_accuracy": 0.8513499745287825,
"eval_loss": 0.39848873019218445,
"eval_runtime": 42.353,
"eval_samples_per_second": 231.743,
"eval_steps_per_second": 28.971,
"step": 8000
},
{
"epoch": 0.37,
"learning_rate": 2.669567132944157e-05,
"loss": 0.4288,
"step": 9000
},
{
"epoch": 0.37,
"eval_accuracy": 0.8493122771268466,
"eval_loss": 0.40836581587791443,
"eval_runtime": 42.666,
"eval_samples_per_second": 230.042,
"eval_steps_per_second": 28.758,
"step": 9000
},
{
"epoch": 0.41,
"learning_rate": 2.6282630245621764e-05,
"loss": 0.4354,
"step": 10000
},
{
"epoch": 0.41,
"eval_accuracy": 0.8532857870606215,
"eval_loss": 0.3850448429584503,
"eval_runtime": 42.6387,
"eval_samples_per_second": 230.19,
"eval_steps_per_second": 28.777,
"step": 10000
},
{
"epoch": 0.45,
"learning_rate": 2.5869589161801962e-05,
"loss": 0.423,
"step": 11000
},
{
"epoch": 0.45,
"eval_accuracy": 0.8509424350483953,
"eval_loss": 0.3855280578136444,
"eval_runtime": 42.4952,
"eval_samples_per_second": 230.967,
"eval_steps_per_second": 28.874,
"step": 11000
},
{
"epoch": 0.49,
"learning_rate": 2.5456548077982157e-05,
"loss": 0.4167,
"step": 12000
},
{
"epoch": 0.49,
"eval_accuracy": 0.8513499745287825,
"eval_loss": 0.41219788789749146,
"eval_runtime": 42.4026,
"eval_samples_per_second": 231.472,
"eval_steps_per_second": 28.937,
"step": 12000
},
{
"epoch": 0.53,
"learning_rate": 2.504350699416235e-05,
"loss": 0.4129,
"step": 13000
},
{
"epoch": 0.53,
"eval_accuracy": 0.8550178298522669,
"eval_loss": 0.40088921785354614,
"eval_runtime": 42.5886,
"eval_samples_per_second": 230.461,
"eval_steps_per_second": 28.811,
"step": 13000
},
{
"epoch": 0.57,
"learning_rate": 2.463046591034255e-05,
"loss": 0.4135,
"step": 14000
},
{
"epoch": 0.57,
"eval_accuracy": 0.8544065206316862,
"eval_loss": 0.4136492609977722,
"eval_runtime": 42.4803,
"eval_samples_per_second": 231.048,
"eval_steps_per_second": 28.884,
"step": 14000
},
{
"epoch": 0.61,
"learning_rate": 2.4217424826522744e-05,
"loss": 0.4074,
"step": 15000
},
{
"epoch": 0.61,
"eval_accuracy": 0.8595007641365258,
"eval_loss": 0.38692933320999146,
"eval_runtime": 42.276,
"eval_samples_per_second": 232.165,
"eval_steps_per_second": 29.024,
"step": 15000
},
{
"epoch": 0.65,
"learning_rate": 2.3804383742702942e-05,
"loss": 0.415,
"step": 16000
},
{
"epoch": 0.65,
"eval_accuracy": 0.8516556291390729,
"eval_loss": 0.39110422134399414,
"eval_runtime": 42.4226,
"eval_samples_per_second": 231.363,
"eval_steps_per_second": 28.923,
"step": 16000
},
{
"epoch": 0.69,
"learning_rate": 2.3391342658883137e-05,
"loss": 0.4095,
"step": 17000
},
{
"epoch": 0.69,
"eval_accuracy": 0.8592969943963321,
"eval_loss": 0.38802793622016907,
"eval_runtime": 42.6242,
"eval_samples_per_second": 230.268,
"eval_steps_per_second": 28.786,
"step": 17000
},
{
"epoch": 0.73,
"learning_rate": 2.2978301575063332e-05,
"loss": 0.4001,
"step": 18000
},
{
"epoch": 0.73,
"eval_accuracy": 0.8586856851757514,
"eval_loss": 0.3907186985015869,
"eval_runtime": 42.4988,
"eval_samples_per_second": 230.948,
"eval_steps_per_second": 28.871,
"step": 18000
},
{
"epoch": 0.77,
"learning_rate": 2.256526049124353e-05,
"loss": 0.4069,
"step": 19000
},
{
"epoch": 0.77,
"eval_accuracy": 0.8629648497198166,
"eval_loss": 0.3686215281486511,
"eval_runtime": 42.3356,
"eval_samples_per_second": 231.838,
"eval_steps_per_second": 28.983,
"step": 19000
},
{
"epoch": 0.81,
"learning_rate": 2.2152219407423725e-05,
"loss": 0.3927,
"step": 20000
},
{
"epoch": 0.81,
"eval_accuracy": 0.8592969943963321,
"eval_loss": 0.4008384943008423,
"eval_runtime": 42.4047,
"eval_samples_per_second": 231.46,
"eval_steps_per_second": 28.935,
"step": 20000
},
{
"epoch": 0.86,
"learning_rate": 2.173917832360392e-05,
"loss": 0.3958,
"step": 21000
},
{
"epoch": 0.86,
"eval_accuracy": 0.8638818135506877,
"eval_loss": 0.37160658836364746,
"eval_runtime": 42.5674,
"eval_samples_per_second": 230.575,
"eval_steps_per_second": 28.825,
"step": 21000
},
{
"epoch": 0.9,
"learning_rate": 2.1326137239784118e-05,
"loss": 0.4016,
"step": 22000
},
{
"epoch": 0.9,
"eval_accuracy": 0.8678553234844626,
"eval_loss": 0.35941794514656067,
"eval_runtime": 42.472,
"eval_samples_per_second": 231.093,
"eval_steps_per_second": 28.89,
"step": 22000
},
{
"epoch": 0.94,
"learning_rate": 2.0913096155964313e-05,
"loss": 0.3945,
"step": 23000
},
{
"epoch": 0.94,
"eval_accuracy": 0.8678553234844626,
"eval_loss": 0.359527587890625,
"eval_runtime": 42.614,
"eval_samples_per_second": 230.324,
"eval_steps_per_second": 28.793,
"step": 23000
},
{
"epoch": 0.98,
"learning_rate": 2.050005507214451e-05,
"loss": 0.3932,
"step": 24000
},
{
"epoch": 0.98,
"eval_accuracy": 0.8644931227712684,
"eval_loss": 0.35774311423301697,
"eval_runtime": 42.4615,
"eval_samples_per_second": 231.151,
"eval_steps_per_second": 28.897,
"step": 24000
},
{
"epoch": 1.02,
"learning_rate": 2.0087013988324705e-05,
"loss": 0.345,
"step": 25000
},
{
"epoch": 1.02,
"eval_accuracy": 0.8698930208863984,
"eval_loss": 0.40802302956581116,
"eval_runtime": 42.6011,
"eval_samples_per_second": 230.393,
"eval_steps_per_second": 28.802,
"step": 25000
},
{
"epoch": 1.06,
"learning_rate": 1.96739729045049e-05,
"loss": 0.2885,
"step": 26000
},
{
"epoch": 1.06,
"eval_accuracy": 0.8674477840040754,
"eval_loss": 0.39192140102386475,
"eval_runtime": 42.4976,
"eval_samples_per_second": 230.954,
"eval_steps_per_second": 28.872,
"step": 26000
},
{
"epoch": 1.1,
"learning_rate": 1.92609318206851e-05,
"loss": 0.2858,
"step": 27000
},
{
"epoch": 1.1,
"eval_accuracy": 0.8651044319918492,
"eval_loss": 0.4346281588077545,
"eval_runtime": 42.3673,
"eval_samples_per_second": 231.664,
"eval_steps_per_second": 28.961,
"step": 27000
},
{
"epoch": 1.14,
"learning_rate": 1.8847890736865293e-05,
"loss": 0.2872,
"step": 28000
},
{
"epoch": 1.14,
"eval_accuracy": 0.8674477840040754,
"eval_loss": 0.41050681471824646,
"eval_runtime": 42.4953,
"eval_samples_per_second": 230.967,
"eval_steps_per_second": 28.874,
"step": 28000
},
{
"epoch": 1.18,
"learning_rate": 1.8434849653045488e-05,
"loss": 0.3002,
"step": 29000
},
{
"epoch": 1.18,
"eval_accuracy": 0.8708099847172694,
"eval_loss": 0.4133119583129883,
"eval_runtime": 42.4588,
"eval_samples_per_second": 231.165,
"eval_steps_per_second": 28.899,
"step": 29000
},
{
"epoch": 1.22,
"learning_rate": 1.8021808569225686e-05,
"loss": 0.2954,
"step": 30000
},
{
"epoch": 1.22,
"eval_accuracy": 0.8667345899133979,
"eval_loss": 0.406183123588562,
"eval_runtime": 42.2388,
"eval_samples_per_second": 232.369,
"eval_steps_per_second": 29.049,
"step": 30000
},
{
"epoch": 1.26,
"learning_rate": 1.760876748540588e-05,
"loss": 0.2912,
"step": 31000
},
{
"epoch": 1.26,
"eval_accuracy": 0.8708099847172694,
"eval_loss": 0.397215336561203,
"eval_runtime": 42.5529,
"eval_samples_per_second": 230.654,
"eval_steps_per_second": 28.835,
"step": 31000
},
{
"epoch": 1.3,
"learning_rate": 1.7195726401586076e-05,
"loss": 0.2958,
"step": 32000
},
{
"epoch": 1.3,
"eval_accuracy": 0.8731533367294957,
"eval_loss": 0.3713410496711731,
"eval_runtime": 42.8662,
"eval_samples_per_second": 228.968,
"eval_steps_per_second": 28.624,
"step": 32000
},
{
"epoch": 1.34,
"learning_rate": 1.6782685317766274e-05,
"loss": 0.293,
"step": 33000
},
{
"epoch": 1.34,
"eval_accuracy": 0.871523178807947,
"eval_loss": 0.3716830015182495,
"eval_runtime": 42.5402,
"eval_samples_per_second": 230.723,
"eval_steps_per_second": 28.843,
"step": 33000
},
{
"epoch": 1.39,
"learning_rate": 1.636964423394647e-05,
"loss": 0.3001,
"step": 34000
},
{
"epoch": 1.39,
"eval_accuracy": 0.8716250636780438,
"eval_loss": 0.3826219141483307,
"eval_runtime": 42.576,
"eval_samples_per_second": 230.529,
"eval_steps_per_second": 28.819,
"step": 34000
},
{
"epoch": 1.43,
"learning_rate": 1.5956603150126667e-05,
"loss": 0.2864,
"step": 35000
},
{
"epoch": 1.43,
"eval_accuracy": 0.8693835965359145,
"eval_loss": 0.41551369428634644,
"eval_runtime": 42.2749,
"eval_samples_per_second": 232.171,
"eval_steps_per_second": 29.024,
"step": 35000
},
{
"epoch": 1.47,
"learning_rate": 1.554356206630686e-05,
"loss": 0.2827,
"step": 36000
},
{
"epoch": 1.47,
"eval_accuracy": 0.866632705043301,
"eval_loss": 0.4223748743534088,
"eval_runtime": 42.4098,
"eval_samples_per_second": 231.432,
"eval_steps_per_second": 28.932,
"step": 36000
},
{
"epoch": 1.51,
"learning_rate": 1.5130520982487058e-05,
"loss": 0.2836,
"step": 37000
},
{
"epoch": 1.51,
"eval_accuracy": 0.8743759551706571,
"eval_loss": 0.3832014799118042,
"eval_runtime": 42.6737,
"eval_samples_per_second": 230.001,
"eval_steps_per_second": 28.753,
"step": 37000
},
{
"epoch": 1.55,
"learning_rate": 1.4717479898667254e-05,
"loss": 0.2844,
"step": 38000
},
{
"epoch": 1.55,
"eval_accuracy": 0.8698930208863984,
"eval_loss": 0.4178868234157562,
"eval_runtime": 42.5391,
"eval_samples_per_second": 230.729,
"eval_steps_per_second": 28.844,
"step": 38000
},
{
"epoch": 1.59,
"learning_rate": 1.430443881484745e-05,
"loss": 0.2866,
"step": 39000
},
{
"epoch": 1.59,
"eval_accuracy": 0.8680590932246561,
"eval_loss": 0.3968764543533325,
"eval_runtime": 42.6173,
"eval_samples_per_second": 230.306,
"eval_steps_per_second": 28.791,
"step": 39000
},
{
"epoch": 1.63,
"learning_rate": 1.3891397731027647e-05,
"loss": 0.2883,
"step": 40000
},
{
"epoch": 1.63,
"eval_accuracy": 0.8682628629648497,
"eval_loss": 0.4000142514705658,
"eval_runtime": 42.4575,
"eval_samples_per_second": 231.172,
"eval_steps_per_second": 28.899,
"step": 40000
},
{
"epoch": 1.67,
"learning_rate": 1.3478356647207844e-05,
"loss": 0.2832,
"step": 41000
},
{
"epoch": 1.67,
"eval_accuracy": 0.8687722873153336,
"eval_loss": 0.3853473365306854,
"eval_runtime": 42.6575,
"eval_samples_per_second": 230.088,
"eval_steps_per_second": 28.764,
"step": 41000
},
{
"epoch": 1.71,
"learning_rate": 1.3065315563388038e-05,
"loss": 0.2876,
"step": 42000
},
{
"epoch": 1.71,
"eval_accuracy": 0.8676515537442689,
"eval_loss": 0.39242836833000183,
"eval_runtime": 42.1832,
"eval_samples_per_second": 232.675,
"eval_steps_per_second": 29.087,
"step": 42000
},
{
"epoch": 1.75,
"learning_rate": 1.2652274479568235e-05,
"loss": 0.2855,
"step": 43000
},
{
"epoch": 1.75,
"eval_accuracy": 0.8719307182883341,
"eval_loss": 0.4176536798477173,
"eval_runtime": 42.3591,
"eval_samples_per_second": 231.709,
"eval_steps_per_second": 28.967,
"step": 43000
},
{
"epoch": 1.79,
"learning_rate": 1.2239233395748431e-05,
"loss": 0.2845,
"step": 44000
},
{
"epoch": 1.79,
"eval_accuracy": 0.8724401426388181,
"eval_loss": 0.38765597343444824,
"eval_runtime": 42.3828,
"eval_samples_per_second": 231.58,
"eval_steps_per_second": 28.95,
"step": 44000
},
{
"epoch": 1.83,
"learning_rate": 1.1826192311928628e-05,
"loss": 0.2882,
"step": 45000
},
{
"epoch": 1.83,
"eval_accuracy": 0.8713194090677534,
"eval_loss": 0.3961141109466553,
"eval_runtime": 42.7036,
"eval_samples_per_second": 229.84,
"eval_steps_per_second": 28.733,
"step": 45000
},
{
"epoch": 1.87,
"learning_rate": 1.1413151228108823e-05,
"loss": 0.2773,
"step": 46000
},
{
"epoch": 1.87,
"eval_accuracy": 0.87396841569027,
"eval_loss": 0.37906670570373535,
"eval_runtime": 42.5306,
"eval_samples_per_second": 230.775,
"eval_steps_per_second": 28.85,
"step": 46000
},
{
"epoch": 1.91,
"learning_rate": 1.1000110144289019e-05,
"loss": 0.2767,
"step": 47000
},
{
"epoch": 1.91,
"eval_accuracy": 0.8779419256240448,
"eval_loss": 0.3877304494380951,
"eval_runtime": 42.632,
"eval_samples_per_second": 230.226,
"eval_steps_per_second": 28.781,
"step": 47000
},
{
"epoch": 1.96,
"learning_rate": 1.0587069060469215e-05,
"loss": 0.2772,
"step": 48000
},
{
"epoch": 1.96,
"eval_accuracy": 0.8689760570555273,
"eval_loss": 0.402159720659256,
"eval_runtime": 42.3751,
"eval_samples_per_second": 231.622,
"eval_steps_per_second": 28.956,
"step": 48000
},
{
"epoch": 2.0,
"learning_rate": 1.017402797664941e-05,
"loss": 0.2816,
"step": 49000
},
{
"epoch": 2.0,
"eval_accuracy": 0.8731533367294957,
"eval_loss": 0.3836565613746643,
"eval_runtime": 42.6247,
"eval_samples_per_second": 230.265,
"eval_steps_per_second": 28.786,
"step": 49000
},
{
"epoch": 2.04,
"learning_rate": 9.760986892829607e-06,
"loss": 0.2068,
"step": 50000
},
{
"epoch": 2.04,
"eval_accuracy": 0.872032603158431,
"eval_loss": 0.4643520712852478,
"eval_runtime": 42.4953,
"eval_samples_per_second": 230.967,
"eval_steps_per_second": 28.874,
"step": 50000
},
{
"epoch": 2.08,
"learning_rate": 9.347945809009803e-06,
"loss": 0.1914,
"step": 51000
},
{
"epoch": 2.08,
"eval_accuracy": 0.8743759551706571,
"eval_loss": 0.49194175004959106,
"eval_runtime": 42.669,
"eval_samples_per_second": 230.027,
"eval_steps_per_second": 28.756,
"step": 51000
},
{
"epoch": 2.12,
"learning_rate": 8.93490472519e-06,
"loss": 0.2,
"step": 52000
},
{
"epoch": 2.12,
"eval_accuracy": 0.8701986754966887,
"eval_loss": 0.4869604706764221,
"eval_runtime": 42.5176,
"eval_samples_per_second": 230.845,
"eval_steps_per_second": 28.859,
"step": 52000
},
{
"epoch": 2.16,
"learning_rate": 8.521863641370194e-06,
"loss": 0.1904,
"step": 53000
},
{
"epoch": 2.16,
"eval_accuracy": 0.8736627610799796,
"eval_loss": 0.5038197636604309,
"eval_runtime": 42.6855,
"eval_samples_per_second": 229.937,
"eval_steps_per_second": 28.745,
"step": 53000
},
{
"epoch": 2.2,
"learning_rate": 8.10882255755039e-06,
"loss": 0.1915,
"step": 54000
},
{
"epoch": 2.2,
"eval_accuracy": 0.8711156393275599,
"eval_loss": 0.5232452750205994,
"eval_runtime": 42.4057,
"eval_samples_per_second": 231.455,
"eval_steps_per_second": 28.935,
"step": 54000
},
{
"epoch": 2.24,
"learning_rate": 7.695781473730587e-06,
"loss": 0.1956,
"step": 55000
},
{
"epoch": 2.24,
"eval_accuracy": 0.8746816097809476,
"eval_loss": 0.5192070603370667,
"eval_runtime": 42.5451,
"eval_samples_per_second": 230.697,
"eval_steps_per_second": 28.84,
"step": 55000
},
{
"epoch": 2.28,
"learning_rate": 7.282740389910783e-06,
"loss": 0.1911,
"step": 56000
},
{
"epoch": 2.28,
"eval_accuracy": 0.8761079979623025,
"eval_loss": 0.5215316414833069,
"eval_runtime": 42.8008,
"eval_samples_per_second": 229.318,
"eval_steps_per_second": 28.668,
"step": 56000
},
{
"epoch": 2.32,
"learning_rate": 6.869699306090979e-06,
"loss": 0.2053,
"step": 57000
},
{
"epoch": 2.32,
"eval_accuracy": 0.8737646459500764,
"eval_loss": 0.460406094789505,
"eval_runtime": 42.6406,
"eval_samples_per_second": 230.18,
"eval_steps_per_second": 28.775,
"step": 57000
},
{
"epoch": 2.36,
"learning_rate": 6.456658222271175e-06,
"loss": 0.2008,
"step": 58000
},
{
"epoch": 2.36,
"eval_accuracy": 0.871523178807947,
"eval_loss": 0.5162080526351929,
"eval_runtime": 42.396,
"eval_samples_per_second": 231.508,
"eval_steps_per_second": 28.941,
"step": 58000
},
{
"epoch": 2.4,
"learning_rate": 6.043617138451371e-06,
"loss": 0.1971,
"step": 59000
},
{
"epoch": 2.4,
"eval_accuracy": 0.875394803871625,
"eval_loss": 0.4885903000831604,
"eval_runtime": 42.2887,
"eval_samples_per_second": 232.095,
"eval_steps_per_second": 29.015,
"step": 59000
},
{
"epoch": 2.44,
"learning_rate": 5.630576054631567e-06,
"loss": 0.192,
"step": 60000
},
{
"epoch": 2.44,
"eval_accuracy": 0.872542027508915,
"eval_loss": 0.49207794666290283,
"eval_runtime": 42.6387,
"eval_samples_per_second": 230.19,
"eval_steps_per_second": 28.777,
"step": 60000
},
{
"epoch": 2.49,
"learning_rate": 5.217534970811764e-06,
"loss": 0.1937,
"step": 61000
},
{
"epoch": 2.49,
"eval_accuracy": 0.8763117677024962,
"eval_loss": 0.4916786253452301,
"eval_runtime": 42.3165,
"eval_samples_per_second": 231.943,
"eval_steps_per_second": 28.996,
"step": 61000
},
{
"epoch": 2.53,
"learning_rate": 4.80449388699196e-06,
"loss": 0.1931,
"step": 62000
},
{
"epoch": 2.53,
"eval_accuracy": 0.8778400407539481,
"eval_loss": 0.47893819212913513,
"eval_runtime": 42.4257,
"eval_samples_per_second": 231.346,
"eval_steps_per_second": 28.921,
"step": 62000
},
{
"epoch": 2.57,
"learning_rate": 4.391452803172156e-06,
"loss": 0.1964,
"step": 63000
},
{
"epoch": 2.57,
"eval_accuracy": 0.8721344880285278,
"eval_loss": 0.49971821904182434,
"eval_runtime": 42.5737,
"eval_samples_per_second": 230.542,
"eval_steps_per_second": 28.821,
"step": 63000
},
{
"epoch": 2.61,
"learning_rate": 3.978411719352352e-06,
"loss": 0.2008,
"step": 64000
},
{
"epoch": 2.61,
"eval_accuracy": 0.8755985736118187,
"eval_loss": 0.4747646749019623,
"eval_runtime": 42.3736,
"eval_samples_per_second": 231.63,
"eval_steps_per_second": 28.957,
"step": 64000
},
{
"epoch": 2.65,
"learning_rate": 3.5653706355325475e-06,
"loss": 0.1962,
"step": 65000
},
{
"epoch": 2.65,
"eval_accuracy": 0.876413652572593,
"eval_loss": 0.4839693605899811,
"eval_runtime": 42.5306,
"eval_samples_per_second": 230.775,
"eval_steps_per_second": 28.85,
"step": 65000
},
{
"epoch": 2.69,
"learning_rate": 3.1523295517127435e-06,
"loss": 0.2029,
"step": 66000
},
{
"epoch": 2.69,
"eval_accuracy": 0.8767193071828834,
"eval_loss": 0.48892590403556824,
"eval_runtime": 42.2422,
"eval_samples_per_second": 232.351,
"eval_steps_per_second": 29.047,
"step": 66000
},
{
"epoch": 2.73,
"learning_rate": 2.73928846789294e-06,
"loss": 0.1927,
"step": 67000
},
{
"epoch": 2.73,
"eval_accuracy": 0.8758023433520122,
"eval_loss": 0.4820311963558197,
"eval_runtime": 42.6573,
"eval_samples_per_second": 230.089,
"eval_steps_per_second": 28.764,
"step": 67000
},
{
"epoch": 2.77,
"learning_rate": 2.326247384073136e-06,
"loss": 0.1926,
"step": 68000
},
{
"epoch": 2.77,
"eval_accuracy": 0.8762098828323994,
"eval_loss": 0.48573482036590576,
"eval_runtime": 42.7141,
"eval_samples_per_second": 229.783,
"eval_steps_per_second": 28.726,
"step": 68000
},
{
"epoch": 2.81,
"learning_rate": 1.913206300253332e-06,
"loss": 0.1919,
"step": 69000
},
{
"epoch": 2.81,
"eval_accuracy": 0.8748853795211411,
"eval_loss": 0.4835805296897888,
"eval_runtime": 42.3433,
"eval_samples_per_second": 231.796,
"eval_steps_per_second": 28.977,
"step": 69000
},
{
"epoch": 2.85,
"learning_rate": 1.500165216433528e-06,
"loss": 0.1911,
"step": 70000
},
{
"epoch": 2.85,
"eval_accuracy": 0.8741721854304636,
"eval_loss": 0.48588985204696655,
"eval_runtime": 42.2277,
"eval_samples_per_second": 232.43,
"eval_steps_per_second": 29.057,
"step": 70000
},
{
"epoch": 2.89,
"learning_rate": 1.087124132613724e-06,
"loss": 0.1897,
"step": 71000
},
{
"epoch": 2.89,
"eval_accuracy": 0.8766174223127865,
"eval_loss": 0.48527196049690247,
"eval_runtime": 42.5077,
"eval_samples_per_second": 230.899,
"eval_steps_per_second": 28.865,
"step": 71000
},
{
"epoch": 2.93,
"learning_rate": 6.740830487939201e-07,
"loss": 0.186,
"step": 72000
},
{
"epoch": 2.93,
"eval_accuracy": 0.8768211920529801,
"eval_loss": 0.4946127235889435,
"eval_runtime": 42.345,
"eval_samples_per_second": 231.786,
"eval_steps_per_second": 28.976,
"step": 72000
},
{
"epoch": 2.97,
"learning_rate": 2.610419649741161e-07,
"loss": 0.2011,
"step": 73000
},
{
"epoch": 2.97,
"eval_accuracy": 0.8767193071828834,
"eval_loss": 0.4851147532463074,
"eval_runtime": 42.2673,
"eval_samples_per_second": 232.213,
"eval_steps_per_second": 29.03,
"step": 73000
},
{
"epoch": 3.0,
"step": 73632,
"total_flos": 7.803448964124365e+16,
"train_loss": 0.3087343405972663,
"train_runtime": 16507.3954,
"train_samples_per_second": 71.368,
"train_steps_per_second": 4.461
}
],
"max_steps": 73632,
"num_train_epochs": 3,
"total_flos": 7.803448964124365e+16,
"trial_name": null,
"trial_params": null
}