qaconv-roberta-large-squad2 / trainer_state.json
Jason Wu
init 2ca8142
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 4.999350804660539,
"global_step": 4570,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.05,
"learning_rate": 1.5e-06,
"loss": 2.6965,
"step": 50
},
{
"epoch": 0.11,
"learning_rate": 3e-06,
"loss": 1.85,
"step": 100
},
{
"epoch": 0.16,
"learning_rate": 4.5e-06,
"loss": 1.7546,
"step": 150
},
{
"epoch": 0.22,
"learning_rate": 6e-06,
"loss": 1.6255,
"step": 200
},
{
"epoch": 0.27,
"learning_rate": 7.5e-06,
"loss": 1.5579,
"step": 250
},
{
"epoch": 0.33,
"learning_rate": 9e-06,
"loss": 1.5321,
"step": 300
},
{
"epoch": 0.38,
"learning_rate": 1.05e-05,
"loss": 1.4919,
"step": 350
},
{
"epoch": 0.44,
"learning_rate": 1.2e-05,
"loss": 1.4036,
"step": 400
},
{
"epoch": 0.49,
"learning_rate": 1.3500000000000001e-05,
"loss": 1.4189,
"step": 450
},
{
"epoch": 0.55,
"learning_rate": 1.5e-05,
"loss": 1.4814,
"step": 500
},
{
"epoch": 0.6,
"learning_rate": 1.65e-05,
"loss": 1.5028,
"step": 550
},
{
"epoch": 0.66,
"learning_rate": 1.8e-05,
"loss": 1.4442,
"step": 600
},
{
"epoch": 0.71,
"learning_rate": 1.95e-05,
"loss": 1.4614,
"step": 650
},
{
"epoch": 0.77,
"learning_rate": 2.1e-05,
"loss": 1.3753,
"step": 700
},
{
"epoch": 0.82,
"learning_rate": 2.25e-05,
"loss": 1.3651,
"step": 750
},
{
"epoch": 0.87,
"learning_rate": 2.4e-05,
"loss": 1.4083,
"step": 800
},
{
"epoch": 0.93,
"learning_rate": 2.55e-05,
"loss": 1.4364,
"step": 850
},
{
"epoch": 0.98,
"learning_rate": 2.7000000000000002e-05,
"loss": 1.448,
"step": 900
},
{
"HasAns_exact": 58.954531394989175,
"HasAns_f1": 74.09900890720559,
"HasAns_total": 3233,
"NoAns_exact": 71.16564417177914,
"NoAns_f1": 71.16564417177914,
"NoAns_total": 163,
"best_exact": 60.924617196702,
"best_exact_thresh": 0.0,
"best_f1": 75.3421954643685,
"best_f1_thresh": 0.0,
"epoch": 1.0,
"exact": 59.540636042402824,
"f1": 73.9582143100694,
"step": 914,
"total": 3396
},
{
"epoch": 1.04,
"learning_rate": 2.8499999999999998e-05,
"loss": 1.3679,
"step": 950
},
{
"epoch": 1.09,
"learning_rate": 3e-05,
"loss": 1.2567,
"step": 1000
},
{
"epoch": 1.15,
"learning_rate": 2.957983193277311e-05,
"loss": 1.2319,
"step": 1050
},
{
"epoch": 1.2,
"learning_rate": 2.9159663865546218e-05,
"loss": 1.2453,
"step": 1100
},
{
"epoch": 1.26,
"learning_rate": 2.8739495798319328e-05,
"loss": 1.3484,
"step": 1150
},
{
"epoch": 1.31,
"learning_rate": 2.831932773109244e-05,
"loss": 1.2464,
"step": 1200
},
{
"epoch": 1.37,
"learning_rate": 2.789915966386555e-05,
"loss": 1.2314,
"step": 1250
},
{
"epoch": 1.42,
"learning_rate": 2.7478991596638655e-05,
"loss": 1.3351,
"step": 1300
},
{
"epoch": 1.48,
"learning_rate": 2.7058823529411766e-05,
"loss": 1.2563,
"step": 1350
},
{
"epoch": 1.53,
"learning_rate": 2.6638655462184873e-05,
"loss": 1.3376,
"step": 1400
},
{
"epoch": 1.59,
"learning_rate": 2.6218487394957983e-05,
"loss": 1.2481,
"step": 1450
},
{
"epoch": 1.64,
"learning_rate": 2.5798319327731093e-05,
"loss": 1.1875,
"step": 1500
},
{
"epoch": 1.7,
"learning_rate": 2.5378151260504203e-05,
"loss": 1.2709,
"step": 1550
},
{
"epoch": 1.75,
"learning_rate": 2.4957983193277314e-05,
"loss": 1.1849,
"step": 1600
},
{
"epoch": 1.8,
"learning_rate": 2.453781512605042e-05,
"loss": 1.2105,
"step": 1650
},
{
"epoch": 1.86,
"learning_rate": 2.411764705882353e-05,
"loss": 1.2137,
"step": 1700
},
{
"epoch": 1.91,
"learning_rate": 2.3697478991596638e-05,
"loss": 1.2807,
"step": 1750
},
{
"epoch": 1.97,
"learning_rate": 2.3277310924369748e-05,
"loss": 1.2816,
"step": 1800
},
{
"HasAns_exact": 61.58366841942468,
"HasAns_f1": 76.76180559350382,
"HasAns_total": 3233,
"NoAns_exact": 54.60122699386503,
"NoAns_f1": 54.60122699386503,
"NoAns_total": 163,
"best_exact": 63.42756183745583,
"best_exact_thresh": 0.0,
"best_f1": 77.87718418250819,
"best_f1_thresh": 0.0,
"epoch": 2.0,
"exact": 61.24852767962309,
"f1": 75.69815002467547,
"step": 1828,
"total": 3396
},
{
"epoch": 2.02,
"learning_rate": 2.2857142857142858e-05,
"loss": 1.0327,
"step": 1850
},
{
"epoch": 2.08,
"learning_rate": 2.2436974789915968e-05,
"loss": 0.896,
"step": 1900
},
{
"epoch": 2.13,
"learning_rate": 2.2016806722689075e-05,
"loss": 0.9569,
"step": 1950
},
{
"epoch": 2.19,
"learning_rate": 2.1596638655462185e-05,
"loss": 0.8276,
"step": 2000
},
{
"epoch": 2.24,
"learning_rate": 2.1176470588235296e-05,
"loss": 0.9038,
"step": 2050
},
{
"epoch": 2.3,
"learning_rate": 2.0756302521008403e-05,
"loss": 0.9296,
"step": 2100
},
{
"epoch": 2.35,
"learning_rate": 2.0336134453781513e-05,
"loss": 0.868,
"step": 2150
},
{
"epoch": 2.41,
"learning_rate": 1.9915966386554623e-05,
"loss": 0.884,
"step": 2200
},
{
"epoch": 2.46,
"learning_rate": 1.9495798319327733e-05,
"loss": 0.8594,
"step": 2250
},
{
"epoch": 2.52,
"learning_rate": 1.907563025210084e-05,
"loss": 0.8973,
"step": 2300
},
{
"epoch": 2.57,
"learning_rate": 1.865546218487395e-05,
"loss": 0.9248,
"step": 2350
},
{
"epoch": 2.63,
"learning_rate": 1.8235294117647057e-05,
"loss": 0.8629,
"step": 2400
},
{
"epoch": 2.68,
"learning_rate": 1.7815126050420167e-05,
"loss": 0.9535,
"step": 2450
},
{
"epoch": 2.73,
"learning_rate": 1.7394957983193278e-05,
"loss": 0.9778,
"step": 2500
},
{
"epoch": 2.79,
"learning_rate": 1.6974789915966388e-05,
"loss": 0.8858,
"step": 2550
},
{
"epoch": 2.84,
"learning_rate": 1.6554621848739498e-05,
"loss": 0.8855,
"step": 2600
},
{
"epoch": 2.9,
"learning_rate": 1.6134453781512605e-05,
"loss": 0.9669,
"step": 2650
},
{
"epoch": 2.95,
"learning_rate": 1.5714285714285715e-05,
"loss": 0.8746,
"step": 2700
},
{
"HasAns_exact": 60.81039282400248,
"HasAns_f1": 76.48517908547677,
"HasAns_total": 3233,
"NoAns_exact": 57.668711656441715,
"NoAns_f1": 57.668711656441715,
"NoAns_total": 163,
"best_exact": 62.69140164899882,
"best_exact_thresh": 0.0,
"best_f1": 77.61383509521386,
"best_f1_thresh": 0.0,
"epoch": 3.0,
"exact": 60.65959952885748,
"f1": 75.58203297507256,
"step": 2742,
"total": 3396
},
{
"epoch": 3.01,
"learning_rate": 1.5294117647058822e-05,
"loss": 0.8392,
"step": 2750
},
{
"epoch": 3.06,
"learning_rate": 1.4873949579831934e-05,
"loss": 0.6246,
"step": 2800
},
{
"epoch": 3.12,
"learning_rate": 1.4453781512605043e-05,
"loss": 0.6114,
"step": 2850
},
{
"epoch": 3.17,
"learning_rate": 1.4033613445378151e-05,
"loss": 0.6607,
"step": 2900
},
{
"epoch": 3.23,
"learning_rate": 1.3613445378151261e-05,
"loss": 0.6362,
"step": 2950
},
{
"epoch": 3.28,
"learning_rate": 1.319327731092437e-05,
"loss": 0.6954,
"step": 3000
},
{
"epoch": 3.34,
"learning_rate": 1.2773109243697479e-05,
"loss": 0.5845,
"step": 3050
},
{
"epoch": 3.39,
"learning_rate": 1.2352941176470587e-05,
"loss": 0.6107,
"step": 3100
},
{
"epoch": 3.45,
"learning_rate": 1.1932773109243697e-05,
"loss": 0.5863,
"step": 3150
},
{
"epoch": 3.5,
"learning_rate": 1.1512605042016808e-05,
"loss": 0.6415,
"step": 3200
},
{
"epoch": 3.56,
"learning_rate": 1.1092436974789916e-05,
"loss": 0.6494,
"step": 3250
},
{
"epoch": 3.61,
"learning_rate": 1.0672268907563026e-05,
"loss": 0.6521,
"step": 3300
},
{
"epoch": 3.66,
"learning_rate": 1.0252100840336135e-05,
"loss": 0.6182,
"step": 3350
},
{
"epoch": 3.72,
"learning_rate": 9.831932773109244e-06,
"loss": 0.6293,
"step": 3400
},
{
"epoch": 3.77,
"learning_rate": 9.411764705882354e-06,
"loss": 0.5783,
"step": 3450
},
{
"epoch": 3.83,
"learning_rate": 8.991596638655462e-06,
"loss": 0.5962,
"step": 3500
},
{
"epoch": 3.88,
"learning_rate": 8.571428571428571e-06,
"loss": 0.6096,
"step": 3550
},
{
"epoch": 3.94,
"learning_rate": 8.15126050420168e-06,
"loss": 0.6828,
"step": 3600
},
{
"epoch": 3.99,
"learning_rate": 7.73109243697479e-06,
"loss": 0.6435,
"step": 3650
},
{
"HasAns_exact": 61.24342715743891,
"HasAns_f1": 76.51996301360617,
"HasAns_total": 3233,
"NoAns_exact": 60.736196319018404,
"NoAns_f1": 60.736196319018404,
"NoAns_total": 163,
"best_exact": 63.10365135453475,
"best_exact_thresh": 0.0,
"best_f1": 77.64694947673401,
"best_f1_thresh": 0.0,
"epoch": 4.0,
"exact": 61.21908127208481,
"f1": 75.76237939428408,
"step": 3656,
"total": 3396
},
{
"epoch": 4.05,
"learning_rate": 7.310924369747899e-06,
"loss": 0.4823,
"step": 3700
},
{
"epoch": 4.1,
"learning_rate": 6.8907563025210085e-06,
"loss": 0.4437,
"step": 3750
},
{
"epoch": 4.16,
"learning_rate": 6.470588235294118e-06,
"loss": 0.4424,
"step": 3800
},
{
"epoch": 4.21,
"learning_rate": 6.050420168067227e-06,
"loss": 0.4736,
"step": 3850
},
{
"epoch": 4.27,
"learning_rate": 5.630252100840336e-06,
"loss": 0.4747,
"step": 3900
},
{
"epoch": 4.32,
"learning_rate": 5.210084033613445e-06,
"loss": 0.4094,
"step": 3950
},
{
"epoch": 4.38,
"learning_rate": 4.789915966386555e-06,
"loss": 0.4644,
"step": 4000
},
{
"epoch": 4.43,
"learning_rate": 4.369747899159664e-06,
"loss": 0.4356,
"step": 4050
},
{
"epoch": 4.49,
"learning_rate": 3.9495798319327735e-06,
"loss": 0.4324,
"step": 4100
},
{
"epoch": 4.54,
"learning_rate": 3.5294117647058825e-06,
"loss": 0.4429,
"step": 4150
},
{
"epoch": 4.59,
"learning_rate": 3.109243697478992e-06,
"loss": 0.4524,
"step": 4200
},
{
"epoch": 4.65,
"learning_rate": 2.689075630252101e-06,
"loss": 0.4388,
"step": 4250
},
{
"epoch": 4.7,
"learning_rate": 2.26890756302521e-06,
"loss": 0.4409,
"step": 4300
},
{
"epoch": 4.76,
"learning_rate": 1.8487394957983194e-06,
"loss": 0.4438,
"step": 4350
},
{
"epoch": 4.81,
"learning_rate": 1.4285714285714286e-06,
"loss": 0.3694,
"step": 4400
},
{
"epoch": 4.87,
"learning_rate": 1.0084033613445378e-06,
"loss": 0.4524,
"step": 4450
},
{
"epoch": 4.92,
"learning_rate": 5.882352941176471e-07,
"loss": 0.4175,
"step": 4500
},
{
"epoch": 4.98,
"learning_rate": 1.680672268907563e-07,
"loss": 0.4197,
"step": 4550
},
{
"HasAns_exact": 61.11970306217136,
"HasAns_f1": 76.44576049759735,
"HasAns_total": 3233,
"NoAns_exact": 65.03067484662577,
"NoAns_f1": 65.03067484662577,
"NoAns_total": 163,
"best_exact": 62.985865724381625,
"best_exact_thresh": 0.0,
"best_f1": 77.5763085066938,
"best_f1_thresh": 0.0,
"epoch": 5.0,
"exact": 61.30742049469965,
"f1": 75.89786327701185,
"step": 4570,
"total": 3396
},
{
"epoch": 5.0,
"step": 4570,
"total_flos": 1.5925721822134272e+17,
"train_runtime": 26138.3161,
"train_samples_per_second": 0.175
}
],
"max_steps": 4570,
"num_train_epochs": 5,
"total_flos": 1.5925721822134272e+17,
"trial_name": null,
"trial_params": null
}