{
"best_metric": 0.8983986742650762,
"best_model_checkpoint": "../out/bio/ner-scibert-qa4-querytype5-span/checkpoints/checkpoint-12150",
"epoch": 127.89473684210526,
"global_step": 12150,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.53,
"learning_rate": 9.982456140350879e-06,
"loss": 0.497,
"step": 50
},
{
"epoch": 0.53,
"eval_loss": 0.16596195101737976,
"eval_micro_f1": 0.7603057598056351,
"eval_runtime": 3.8666,
"eval_samples_per_second": 259.4,
"step": 50
},
{
"epoch": 1.05,
"learning_rate": 9.964912280701755e-06,
"loss": 0.1446,
"step": 100
},
{
"epoch": 1.05,
"eval_loss": 0.12134452909231186,
"eval_micro_f1": 0.8320775021912038,
"eval_runtime": 2.9702,
"eval_samples_per_second": 337.691,
"step": 100
},
{
"epoch": 1.58,
"learning_rate": 9.947368421052632e-06,
"loss": 0.1055,
"step": 150
},
{
"epoch": 1.58,
"eval_loss": 0.1103413924574852,
"eval_micro_f1": 0.8564287692859414,
"eval_runtime": 3.8528,
"eval_samples_per_second": 260.329,
"step": 150
},
{
"epoch": 2.11,
"learning_rate": 9.929824561403509e-06,
"loss": 0.0844,
"step": 200
},
{
"epoch": 2.11,
"eval_loss": 0.10704845190048218,
"eval_micro_f1": 0.8708003272791329,
"eval_runtime": 3.2958,
"eval_samples_per_second": 304.328,
"step": 200
},
{
"epoch": 2.63,
"learning_rate": 9.912280701754386e-06,
"loss": 0.0663,
"step": 250
},
{
"epoch": 2.63,
"eval_loss": 0.10774940997362137,
"eval_micro_f1": 0.877275885531736,
"eval_runtime": 4.1403,
"eval_samples_per_second": 242.253,
"step": 250
},
{
"epoch": 3.16,
"learning_rate": 9.894736842105264e-06,
"loss": 0.058,
"step": 300
},
{
"epoch": 3.16,
"eval_loss": 0.10937967151403427,
"eval_micro_f1": 0.8814672866610472,
"eval_runtime": 3.0516,
"eval_samples_per_second": 328.685,
"step": 300
},
{
"epoch": 3.68,
"learning_rate": 9.87719298245614e-06,
"loss": 0.0467,
"step": 350
},
{
"epoch": 3.68,
"eval_loss": 0.11442398279905319,
"eval_micro_f1": 0.8797269619571783,
"eval_runtime": 7.9918,
"eval_samples_per_second": 125.504,
"step": 350
},
{
"epoch": 4.21,
"learning_rate": 9.859649122807018e-06,
"loss": 0.0436,
"step": 400
},
{
"epoch": 4.21,
"eval_loss": 0.12020551413297653,
"eval_micro_f1": 0.8807765923354797,
"eval_runtime": 3.1599,
"eval_samples_per_second": 317.415,
"step": 400
},
{
"epoch": 4.74,
"learning_rate": 9.842105263157896e-06,
"loss": 0.0324,
"step": 450
},
{
"epoch": 4.74,
"eval_loss": 0.12562759220600128,
"eval_micro_f1": 0.8849678561541727,
"eval_runtime": 3.8469,
"eval_samples_per_second": 260.73,
"step": 450
},
{
"epoch": 5.26,
"learning_rate": 9.824561403508772e-06,
"loss": 0.0318,
"step": 500
},
{
"epoch": 5.26,
"eval_loss": 0.1307975947856903,
"eval_micro_f1": 0.8806091917762102,
"eval_runtime": 3.3602,
"eval_samples_per_second": 298.498,
"step": 500
},
{
"epoch": 5.79,
"learning_rate": 9.80701754385965e-06,
"loss": 0.0248,
"step": 550
},
{
"epoch": 5.79,
"eval_loss": 0.13458630442619324,
"eval_micro_f1": 0.8805767133192626,
"eval_runtime": 3.9422,
"eval_samples_per_second": 254.429,
"step": 550
},
{
"epoch": 6.32,
"learning_rate": 9.789473684210527e-06,
"loss": 0.0178,
"step": 600
},
{
"epoch": 6.32,
"eval_loss": 0.15008904039859772,
"eval_micro_f1": 0.881859441117156,
"eval_runtime": 4.1673,
"eval_samples_per_second": 240.683,
"step": 600
},
{
"epoch": 6.84,
"learning_rate": 9.771929824561405e-06,
"loss": 0.021,
"step": 650
},
{
"epoch": 6.84,
"eval_loss": 0.14670833945274353,
"eval_micro_f1": 0.8866602550293828,
"eval_runtime": 4.0864,
"eval_samples_per_second": 245.451,
"step": 650
},
{
"epoch": 7.37,
"learning_rate": 9.754385964912281e-06,
"loss": 0.0134,
"step": 700
},
{
"epoch": 7.37,
"eval_loss": 0.1600644886493683,
"eval_micro_f1": 0.8886732283920349,
"eval_runtime": 3.7734,
"eval_samples_per_second": 265.81,
"step": 700
},
{
"epoch": 7.89,
"learning_rate": 9.736842105263159e-06,
"loss": 0.0154,
"step": 750
},
{
"epoch": 7.89,
"eval_loss": 0.15796250104904175,
"eval_micro_f1": 0.8849800737061796,
"eval_runtime": 9.0614,
"eval_samples_per_second": 110.689,
"step": 750
},
{
"epoch": 8.42,
"learning_rate": 9.719298245614037e-06,
"loss": 0.0122,
"step": 800
},
{
"epoch": 8.42,
"eval_loss": 0.16263359785079956,
"eval_micro_f1": 0.8857808852806778,
"eval_runtime": 3.0355,
"eval_samples_per_second": 330.427,
"step": 800
},
{
"epoch": 8.95,
"learning_rate": 9.701754385964913e-06,
"loss": 0.0102,
"step": 850
},
{
"epoch": 8.95,
"eval_loss": 0.1681569218635559,
"eval_micro_f1": 0.8865866952468597,
"eval_runtime": 3.9467,
"eval_samples_per_second": 254.139,
"step": 850
},
{
"epoch": 9.47,
"learning_rate": 9.68421052631579e-06,
"loss": 0.0085,
"step": 900
},
{
"epoch": 9.47,
"eval_loss": 0.17396388947963715,
"eval_micro_f1": 0.8904203992240877,
"eval_runtime": 3.2397,
"eval_samples_per_second": 309.6,
"step": 900
},
{
"epoch": 10.0,
"learning_rate": 9.666666666666667e-06,
"loss": 0.0082,
"step": 950
},
{
"epoch": 10.0,
"eval_loss": 0.17073805630207062,
"eval_micro_f1": 0.8865724861271382,
"eval_runtime": 4.0124,
"eval_samples_per_second": 249.975,
"step": 950
},
{
"epoch": 10.53,
"learning_rate": 9.649122807017545e-06,
"loss": 0.0067,
"step": 1000
},
{
"epoch": 10.53,
"eval_loss": 0.18167999386787415,
"eval_micro_f1": 0.8881885443065939,
"eval_runtime": 3.029,
"eval_samples_per_second": 331.129,
"step": 1000
},
{
"epoch": 11.05,
"learning_rate": 9.631578947368422e-06,
"loss": 0.0063,
"step": 1050
},
{
"epoch": 11.05,
"eval_loss": 0.19336147606372833,
"eval_micro_f1": 0.8878196310642488,
"eval_runtime": 4.1983,
"eval_samples_per_second": 238.906,
"step": 1050
},
{
"epoch": 11.58,
"learning_rate": 9.614035087719298e-06,
"loss": 0.0049,
"step": 1100
},
{
"epoch": 11.58,
"eval_loss": 0.19511640071868896,
"eval_micro_f1": 0.8885232409642093,
"eval_runtime": 7.209,
"eval_samples_per_second": 139.132,
"step": 1100
},
{
"epoch": 12.11,
"learning_rate": 9.596491228070176e-06,
"loss": 0.005,
"step": 1150
},
{
"epoch": 12.11,
"eval_loss": 0.19420750439167023,
"eval_micro_f1": 0.888308115043098,
"eval_runtime": 4.1643,
"eval_samples_per_second": 240.857,
"step": 1150
},
{
"epoch": 12.63,
"learning_rate": 9.578947368421054e-06,
"loss": 0.0045,
"step": 1200
},
{
"epoch": 12.63,
"eval_loss": 0.20499350130558014,
"eval_micro_f1": 0.8858688110376867,
"eval_runtime": 3.8458,
"eval_samples_per_second": 260.802,
"step": 1200
},
{
"epoch": 13.16,
"learning_rate": 9.56140350877193e-06,
"loss": 0.0038,
"step": 1250
},
{
"epoch": 13.16,
"eval_loss": 0.20660339295864105,
"eval_micro_f1": 0.8856629332755052,
"eval_runtime": 8.3149,
"eval_samples_per_second": 120.627,
"step": 1250
},
{
"epoch": 13.68,
"learning_rate": 9.543859649122808e-06,
"loss": 0.004,
"step": 1300
},
{
"epoch": 13.68,
"eval_loss": 0.20800620317459106,
"eval_micro_f1": 0.8840419652645282,
"eval_runtime": 3.2669,
"eval_samples_per_second": 307.02,
"step": 1300
},
{
"epoch": 14.21,
"learning_rate": 9.526315789473684e-06,
"loss": 0.0044,
"step": 1350
},
{
"epoch": 14.21,
"eval_loss": 0.20232698321342468,
"eval_micro_f1": 0.8879736403564538,
"eval_runtime": 3.93,
"eval_samples_per_second": 255.213,
"step": 1350
},
{
"epoch": 14.74,
"learning_rate": 9.508771929824562e-06,
"loss": 0.0031,
"step": 1400
},
{
"epoch": 14.74,
"eval_loss": 0.20720337331295013,
"eval_micro_f1": 0.8866318962881453,
"eval_runtime": 3.3011,
"eval_samples_per_second": 303.841,
"step": 1400
},
{
"epoch": 15.26,
"learning_rate": 9.49122807017544e-06,
"loss": 0.003,
"step": 1450
},
{
"epoch": 15.26,
"eval_loss": 0.21012167632579803,
"eval_micro_f1": 0.887170279847774,
"eval_runtime": 8.5032,
"eval_samples_per_second": 117.955,
"step": 1450
},
{
"epoch": 15.79,
"learning_rate": 9.473684210526315e-06,
"loss": 0.0027,
"step": 1500
},
{
"epoch": 15.79,
"eval_loss": 0.21568599343299866,
"eval_micro_f1": 0.8833127651654995,
"eval_runtime": 4.0273,
"eval_samples_per_second": 249.049,
"step": 1500
},
{
"epoch": 16.32,
"learning_rate": 9.456140350877195e-06,
"loss": 0.0022,
"step": 1550
},
{
"epoch": 16.32,
"eval_loss": 0.21815350651741028,
"eval_micro_f1": 0.8888275095121845,
"eval_runtime": 4.0059,
"eval_samples_per_second": 250.383,
"step": 1550
},
{
"epoch": 16.84,
"learning_rate": 9.438596491228071e-06,
"loss": 0.0032,
"step": 1600
},
{
"epoch": 16.84,
"eval_loss": 0.22826983034610748,
"eval_micro_f1": 0.8860584239771193,
"eval_runtime": 4.1145,
"eval_samples_per_second": 243.773,
"step": 1600
},
{
"epoch": 17.37,
"learning_rate": 9.421052631578949e-06,
"loss": 0.0025,
"step": 1650
},
{
"epoch": 17.37,
"eval_loss": 0.2320602834224701,
"eval_micro_f1": 0.8888272797880848,
"eval_runtime": 4.0759,
"eval_samples_per_second": 246.078,
"step": 1650
},
{
"epoch": 17.89,
"learning_rate": 9.403508771929825e-06,
"loss": 0.0029,
"step": 1700
},
{
"epoch": 17.89,
"eval_loss": 0.2191556990146637,
"eval_micro_f1": 0.8870923535694059,
"eval_runtime": 3.809,
"eval_samples_per_second": 263.323,
"step": 1700
},
{
"epoch": 18.42,
"learning_rate": 9.385964912280703e-06,
"loss": 0.0016,
"step": 1750
},
{
"epoch": 18.42,
"eval_loss": 0.2177244871854782,
"eval_micro_f1": 0.8863077338534753,
"eval_runtime": 3.9224,
"eval_samples_per_second": 255.708,
"step": 1750
},
{
"epoch": 18.95,
"learning_rate": 9.36842105263158e-06,
"loss": 0.0028,
"step": 1800
},
{
"epoch": 18.95,
"eval_loss": 0.21616131067276,
"eval_micro_f1": 0.8853046589979684,
"eval_runtime": 4.0177,
"eval_samples_per_second": 249.643,
"step": 1800
},
{
"epoch": 19.47,
"learning_rate": 9.350877192982457e-06,
"loss": 0.0019,
"step": 1850
},
{
"epoch": 19.47,
"eval_loss": 0.20461300015449524,
"eval_micro_f1": 0.8882208248619471,
"eval_runtime": 4.2755,
"eval_samples_per_second": 234.592,
"step": 1850
},
{
"epoch": 20.0,
"learning_rate": 9.333333333333334e-06,
"loss": 0.0019,
"step": 1900
},
{
"epoch": 20.0,
"eval_loss": 0.2104501575231552,
"eval_micro_f1": 0.8884625919908908,
"eval_runtime": 3.4773,
"eval_samples_per_second": 288.446,
"step": 1900
},
{
"epoch": 20.53,
"learning_rate": 9.315789473684212e-06,
"loss": 0.0018,
"step": 1950
},
{
"epoch": 20.53,
"eval_loss": 0.22620131075382233,
"eval_micro_f1": 0.8873531439365204,
"eval_runtime": 4.2745,
"eval_samples_per_second": 234.646,
"step": 1950
},
{
"epoch": 21.05,
"learning_rate": 9.298245614035088e-06,
"loss": 0.0021,
"step": 2000
},
{
"epoch": 21.05,
"eval_loss": 0.22461262345314026,
"eval_micro_f1": 0.8881224991548825,
"eval_runtime": 3.2192,
"eval_samples_per_second": 311.564,
"step": 2000
},
{
"epoch": 21.58,
"learning_rate": 9.280701754385966e-06,
"loss": 0.0018,
"step": 2050
},
{
"epoch": 21.58,
"eval_loss": 0.22409117221832275,
"eval_micro_f1": 0.8847725410416855,
"eval_runtime": 4.0098,
"eval_samples_per_second": 250.135,
"step": 2050
},
{
"epoch": 22.11,
"learning_rate": 9.263157894736842e-06,
"loss": 0.0018,
"step": 2100
},
{
"epoch": 22.11,
"eval_loss": 0.22294463217258453,
"eval_micro_f1": 0.8847900889698287,
"eval_runtime": 4.1446,
"eval_samples_per_second": 242.002,
"step": 2100
},
{
"epoch": 22.63,
"learning_rate": 9.24561403508772e-06,
"loss": 0.0013,
"step": 2150
},
{
"epoch": 22.63,
"eval_loss": 0.22384154796600342,
"eval_micro_f1": 0.8887362632360413,
"eval_runtime": 3.8498,
"eval_samples_per_second": 260.531,
"step": 2150
},
{
"epoch": 23.16,
"learning_rate": 9.228070175438598e-06,
"loss": 0.0016,
"step": 2200
},
{
"epoch": 23.16,
"eval_loss": 0.2359953075647354,
"eval_micro_f1": 0.8928472792567275,
"eval_runtime": 7.8584,
"eval_samples_per_second": 127.634,
"step": 2200
},
{
"epoch": 23.68,
"learning_rate": 9.210526315789474e-06,
"loss": 0.0025,
"step": 2250
},
{
"epoch": 23.68,
"eval_loss": 0.23350147902965546,
"eval_micro_f1": 0.889812889312646,
"eval_runtime": 4.2266,
"eval_samples_per_second": 237.307,
"step": 2250
},
{
"epoch": 24.21,
"learning_rate": 9.192982456140351e-06,
"loss": 0.0013,
"step": 2300
},
{
"epoch": 24.21,
"eval_loss": 0.2257278561592102,
"eval_micro_f1": 0.8870110394344386,
"eval_runtime": 3.1056,
"eval_samples_per_second": 322.962,
"step": 2300
},
{
"epoch": 24.74,
"learning_rate": 9.17543859649123e-06,
"loss": 0.0016,
"step": 2350
},
{
"epoch": 24.74,
"eval_loss": 0.23105472326278687,
"eval_micro_f1": 0.8878048775485645,
"eval_runtime": 4.2866,
"eval_samples_per_second": 233.983,
"step": 2350
},
{
"epoch": 25.26,
"learning_rate": 9.157894736842105e-06,
"loss": 0.0013,
"step": 2400
},
{
"epoch": 25.26,
"eval_loss": 0.2281590700149536,
"eval_micro_f1": 0.8916026339674028,
"eval_runtime": 3.4214,
"eval_samples_per_second": 293.154,
"step": 2400
},
{
"epoch": 25.79,
"learning_rate": 9.140350877192983e-06,
"loss": 0.0011,
"step": 2450
},
{
"epoch": 25.79,
"eval_loss": 0.2338750958442688,
"eval_micro_f1": 0.8898037596205473,
"eval_runtime": 8.6143,
"eval_samples_per_second": 116.434,
"step": 2450
},
{
"epoch": 26.32,
"learning_rate": 9.12280701754386e-06,
"loss": 0.0011,
"step": 2500
},
{
"epoch": 26.32,
"eval_loss": 0.23937590420246124,
"eval_micro_f1": 0.8897453539388568,
"eval_runtime": 7.6161,
"eval_samples_per_second": 131.695,
"step": 2500
},
{
"epoch": 26.84,
"learning_rate": 9.105263157894739e-06,
"loss": 0.0011,
"step": 2550
},
{
"epoch": 26.84,
"eval_loss": 0.2461203783750534,
"eval_micro_f1": 0.8883066735207185,
"eval_runtime": 4.018,
"eval_samples_per_second": 249.627,
"step": 2550
},
{
"epoch": 27.37,
"learning_rate": 9.087719298245615e-06,
"loss": 0.0012,
"step": 2600
},
{
"epoch": 27.37,
"eval_loss": 0.24809321761131287,
"eval_micro_f1": 0.8906206507661238,
"eval_runtime": 4.1771,
"eval_samples_per_second": 240.119,
"step": 2600
},
{
"epoch": 27.89,
"learning_rate": 9.070175438596493e-06,
"loss": 0.0011,
"step": 2650
},
{
"epoch": 27.89,
"eval_loss": 0.2448018193244934,
"eval_micro_f1": 0.8903526823596716,
"eval_runtime": 3.9644,
"eval_samples_per_second": 253.001,
"step": 2650
},
{
"epoch": 28.42,
"learning_rate": 9.05263157894737e-06,
"loss": 0.0011,
"step": 2700
},
{
"epoch": 28.42,
"eval_loss": 0.25840333104133606,
"eval_micro_f1": 0.8891343361006115,
"eval_runtime": 3.8391,
"eval_samples_per_second": 261.258,
"step": 2700
},
{
"epoch": 28.95,
"learning_rate": 9.035087719298246e-06,
"loss": 0.0019,
"step": 2750
},
{
"epoch": 28.95,
"eval_loss": 0.24923892319202423,
"eval_micro_f1": 0.889258950374031,
"eval_runtime": 3.973,
"eval_samples_per_second": 252.452,
"step": 2750
},
{
"epoch": 29.47,
"learning_rate": 9.017543859649124e-06,
"loss": 0.0006,
"step": 2800
},
{
"epoch": 29.47,
"eval_loss": 0.24881286919116974,
"eval_micro_f1": 0.8902725884024797,
"eval_runtime": 3.9882,
"eval_samples_per_second": 251.489,
"step": 2800
},
{
"epoch": 30.0,
"learning_rate": 9e-06,
"loss": 0.0009,
"step": 2850
},
{
"epoch": 30.0,
"eval_loss": 0.24882127344608307,
"eval_micro_f1": 0.889040907968823,
"eval_runtime": 7.8761,
"eval_samples_per_second": 127.347,
"step": 2850
},
{
"epoch": 30.53,
"learning_rate": 8.982456140350878e-06,
"loss": 0.0009,
"step": 2900
},
{
"epoch": 30.53,
"eval_loss": 0.2666034996509552,
"eval_micro_f1": 0.887988826315445,
"eval_runtime": 3.6896,
"eval_samples_per_second": 271.845,
"step": 2900
},
{
"epoch": 31.05,
"learning_rate": 8.964912280701756e-06,
"loss": 0.0008,
"step": 2950
},
{
"epoch": 31.05,
"eval_loss": 0.25441551208496094,
"eval_micro_f1": 0.8864879644888763,
"eval_runtime": 4.0487,
"eval_samples_per_second": 247.733,
"step": 2950
},
{
"epoch": 31.58,
"learning_rate": 8.947368421052632e-06,
"loss": 0.0009,
"step": 3000
},
{
"epoch": 31.58,
"eval_loss": 0.2586291432380676,
"eval_micro_f1": 0.8888272797880848,
"eval_runtime": 3.0865,
"eval_samples_per_second": 324.968,
"step": 3000
},
{
"epoch": 32.11,
"learning_rate": 8.92982456140351e-06,
"loss": 0.0011,
"step": 3050
},
{
"epoch": 32.11,
"eval_loss": 0.25473150610923767,
"eval_micro_f1": 0.8907840435162748,
"eval_runtime": 3.82,
"eval_samples_per_second": 262.568,
"step": 3050
},
{
"epoch": 32.63,
"learning_rate": 8.912280701754387e-06,
"loss": 0.0008,
"step": 3100
},
{
"epoch": 32.63,
"eval_loss": 0.25877171754837036,
"eval_micro_f1": 0.8904867251634705,
"eval_runtime": 3.1409,
"eval_samples_per_second": 319.338,
"step": 3100
},
{
"epoch": 33.16,
"learning_rate": 8.894736842105264e-06,
"loss": 0.0011,
"step": 3150
},
{
"epoch": 33.16,
"eval_loss": 0.25028061866760254,
"eval_micro_f1": 0.8921419688410055,
"eval_runtime": 4.2532,
"eval_samples_per_second": 235.82,
"step": 3150
},
{
"epoch": 33.68,
"learning_rate": 8.877192982456141e-06,
"loss": 0.0007,
"step": 3200
},
{
"epoch": 33.68,
"eval_loss": 0.25727859139442444,
"eval_micro_f1": 0.889716018248036,
"eval_runtime": 3.3044,
"eval_samples_per_second": 303.532,
"step": 3200
},
{
"epoch": 34.21,
"learning_rate": 8.859649122807017e-06,
"loss": 0.001,
"step": 3250
},
{
"epoch": 34.21,
"eval_loss": 0.2515270709991455,
"eval_micro_f1": 0.8864879644888763,
"eval_runtime": 4.1808,
"eval_samples_per_second": 239.905,
"step": 3250
},
{
"epoch": 34.74,
"learning_rate": 8.842105263157895e-06,
"loss": 0.0008,
"step": 3300
},
{
"epoch": 34.74,
"eval_loss": 0.25637340545654297,
"eval_micro_f1": 0.8926193113754515,
"eval_runtime": 4.1437,
"eval_samples_per_second": 242.051,
"step": 3300
},
{
"epoch": 35.26,
"learning_rate": 8.824561403508773e-06,
"loss": 0.0008,
"step": 3350
},
{
"epoch": 35.26,
"eval_loss": 0.26558125019073486,
"eval_micro_f1": 0.8904987456686294,
"eval_runtime": 3.5766,
"eval_samples_per_second": 280.432,
"step": 3350
},
{
"epoch": 35.79,
"learning_rate": 8.807017543859649e-06,
"loss": 0.0011,
"step": 3400
},
{
"epoch": 35.79,
"eval_loss": 0.24883043766021729,
"eval_micro_f1": 0.8895916398132595,
"eval_runtime": 4.0716,
"eval_samples_per_second": 246.339,
"step": 3400
},
{
"epoch": 36.32,
"learning_rate": 8.789473684210527e-06,
"loss": 0.0015,
"step": 3450
},
{
"epoch": 36.32,
"eval_loss": 0.23010390996932983,
"eval_micro_f1": 0.8911676232351213,
"eval_runtime": 4.0619,
"eval_samples_per_second": 246.929,
"step": 3450
},
{
"epoch": 36.84,
"learning_rate": 8.771929824561405e-06,
"loss": 0.0009,
"step": 3500
},
{
"epoch": 36.84,
"eval_loss": 0.24908015131950378,
"eval_micro_f1": 0.8912087907085681,
"eval_runtime": 4.0376,
"eval_samples_per_second": 248.415,
"step": 3500
},
{
"epoch": 37.37,
"learning_rate": 8.75438596491228e-06,
"loss": 0.001,
"step": 3550
},
{
"epoch": 37.37,
"eval_loss": 0.24616041779518127,
"eval_micro_f1": 0.8888582834826856,
"eval_runtime": 4.0021,
"eval_samples_per_second": 250.617,
"step": 3550
},
{
"epoch": 37.89,
"learning_rate": 8.736842105263158e-06,
"loss": 0.001,
"step": 3600
},
{
"epoch": 37.89,
"eval_loss": 0.26039624214172363,
"eval_micro_f1": 0.8903976716627039,
"eval_runtime": 4.0496,
"eval_samples_per_second": 247.68,
"step": 3600
},
{
"epoch": 38.42,
"learning_rate": 8.719298245614036e-06,
"loss": 0.0006,
"step": 3650
},
{
"epoch": 38.42,
"eval_loss": 0.24825327098369598,
"eval_micro_f1": 0.8922780978784502,
"eval_runtime": 4.0936,
"eval_samples_per_second": 245.015,
"step": 3650
},
{
"epoch": 38.95,
"learning_rate": 8.701754385964914e-06,
"loss": 0.0005,
"step": 3700
},
{
"epoch": 38.95,
"eval_loss": 0.2580792307853699,
"eval_micro_f1": 0.8930206738441455,
"eval_runtime": 7.9334,
"eval_samples_per_second": 126.428,
"step": 3700
},
{
"epoch": 39.47,
"learning_rate": 8.68421052631579e-06,
"loss": 0.0004,
"step": 3750
},
{
"epoch": 39.47,
"eval_loss": 0.2521957755088806,
"eval_micro_f1": 0.8948311504301512,
"eval_runtime": 7.7435,
"eval_samples_per_second": 129.529,
"step": 3750
},
{
"epoch": 40.0,
"learning_rate": 8.666666666666668e-06,
"loss": 0.0006,
"step": 3800
},
{
"epoch": 40.0,
"eval_loss": 0.25524747371673584,
"eval_micro_f1": 0.8941500339113264,
"eval_runtime": 3.1595,
"eval_samples_per_second": 317.453,
"step": 3800
},
{
"epoch": 40.53,
"learning_rate": 8.649122807017546e-06,
"loss": 0.0003,
"step": 3850
},
{
"epoch": 40.53,
"eval_loss": 0.26724773645401,
"eval_micro_f1": 0.8946349552519649,
"eval_runtime": 4.1388,
"eval_samples_per_second": 242.341,
"step": 3850
},
{
"epoch": 41.05,
"learning_rate": 8.631578947368422e-06,
"loss": 0.0004,
"step": 3900
},
{
"epoch": 41.05,
"eval_loss": 0.27384868264198303,
"eval_micro_f1": 0.8935934684356875,
"eval_runtime": 3.0974,
"eval_samples_per_second": 323.824,
"step": 3900
},
{
"epoch": 41.58,
"learning_rate": 8.6140350877193e-06,
"loss": 0.0003,
"step": 3950
},
{
"epoch": 41.58,
"eval_loss": 0.2772686183452606,
"eval_micro_f1": 0.88907330517057,
"eval_runtime": 4.0147,
"eval_samples_per_second": 249.83,
"step": 3950
},
{
"epoch": 42.11,
"learning_rate": 8.596491228070176e-06,
"loss": 0.0006,
"step": 4000
},
{
"epoch": 42.11,
"eval_loss": 0.29643380641937256,
"eval_micro_f1": 0.885829508608378,
"eval_runtime": 3.3429,
"eval_samples_per_second": 300.038,
"step": 4000
},
{
"epoch": 42.63,
"learning_rate": 8.578947368421053e-06,
"loss": 0.0008,
"step": 4050
},
{
"epoch": 42.63,
"eval_loss": 0.26611000299453735,
"eval_micro_f1": 0.8879096662582328,
"eval_runtime": 4.0238,
"eval_samples_per_second": 249.27,
"step": 4050
},
{
"epoch": 43.16,
"learning_rate": 8.561403508771931e-06,
"loss": 0.0004,
"step": 4100
},
{
"epoch": 43.16,
"eval_loss": 0.2743793725967407,
"eval_micro_f1": 0.8882442321285712,
"eval_runtime": 3.1772,
"eval_samples_per_second": 315.686,
"step": 4100
},
{
"epoch": 43.68,
"learning_rate": 8.543859649122807e-06,
"loss": 0.0008,
"step": 4150
},
{
"epoch": 43.68,
"eval_loss": 0.2628384828567505,
"eval_micro_f1": 0.8910398225086031,
"eval_runtime": 4.2593,
"eval_samples_per_second": 235.485,
"step": 4150
},
{
"epoch": 44.21,
"learning_rate": 8.526315789473685e-06,
"loss": 0.0006,
"step": 4200
},
{
"epoch": 44.21,
"eval_loss": 0.26010385155677795,
"eval_micro_f1": 0.8914301455455033,
"eval_runtime": 3.0812,
"eval_samples_per_second": 325.522,
"step": 4200
},
{
"epoch": 44.74,
"learning_rate": 8.508771929824563e-06,
"loss": 0.0007,
"step": 4250
},
{
"epoch": 44.74,
"eval_loss": 0.2640901803970337,
"eval_micro_f1": 0.893069580354643,
"eval_runtime": 8.5126,
"eval_samples_per_second": 117.826,
"step": 4250
},
{
"epoch": 45.26,
"learning_rate": 8.491228070175439e-06,
"loss": 0.0012,
"step": 4300
},
{
"epoch": 45.26,
"eval_loss": 0.25755465030670166,
"eval_micro_f1": 0.8900480434256571,
"eval_runtime": 3.9392,
"eval_samples_per_second": 254.619,
"step": 4300
},
{
"epoch": 45.79,
"learning_rate": 8.473684210526317e-06,
"loss": 0.001,
"step": 4350
},
{
"epoch": 45.79,
"eval_loss": 0.2633231282234192,
"eval_micro_f1": 0.88768418076193,
"eval_runtime": 3.922,
"eval_samples_per_second": 255.734,
"step": 4350
},
{
"epoch": 46.32,
"learning_rate": 8.456140350877193e-06,
"loss": 0.0006,
"step": 4400
},
{
"epoch": 46.32,
"eval_loss": 0.2707374095916748,
"eval_micro_f1": 0.8882082690250439,
"eval_runtime": 4.0637,
"eval_samples_per_second": 246.818,
"step": 4400
},
{
"epoch": 46.84,
"learning_rate": 8.43859649122807e-06,
"loss": 0.0007,
"step": 4450
},
{
"epoch": 46.84,
"eval_loss": 0.2625438868999481,
"eval_micro_f1": 0.885142543359445,
"eval_runtime": 4.0886,
"eval_samples_per_second": 245.319,
"step": 4450
},
{
"epoch": 47.37,
"learning_rate": 8.421052631578948e-06,
"loss": 0.0011,
"step": 4500
},
{
"epoch": 47.37,
"eval_loss": 0.27610084414482117,
"eval_micro_f1": 0.8910726638596154,
"eval_runtime": 4.0841,
"eval_samples_per_second": 245.585,
"step": 4500
},
{
"epoch": 47.89,
"learning_rate": 8.403508771929824e-06,
"loss": 0.0007,
"step": 4550
},
{
"epoch": 47.89,
"eval_loss": 0.26576119661331177,
"eval_micro_f1": 0.8919966763205791,
"eval_runtime": 3.9294,
"eval_samples_per_second": 255.252,
"step": 4550
},
{
"epoch": 48.42,
"learning_rate": 8.385964912280704e-06,
"loss": 0.0009,
"step": 4600
},
{
"epoch": 48.42,
"eval_loss": 0.2603236734867096,
"eval_micro_f1": 0.891277386849905,
"eval_runtime": 3.9123,
"eval_samples_per_second": 256.369,
"step": 4600
},
{
"epoch": 48.95,
"learning_rate": 8.36842105263158e-06,
"loss": 0.0009,
"step": 4650
},
{
"epoch": 48.95,
"eval_loss": 0.25420811772346497,
"eval_micro_f1": 0.8894707823002013,
"eval_runtime": 4.0944,
"eval_samples_per_second": 244.97,
"step": 4650
},
{
"epoch": 49.47,
"learning_rate": 8.350877192982458e-06,
"loss": 0.0001,
"step": 4700
},
{
"epoch": 49.47,
"eval_loss": 0.2677100598812103,
"eval_micro_f1": 0.8913433655083149,
"eval_runtime": 3.6777,
"eval_samples_per_second": 272.725,
"step": 4700
},
{
"epoch": 50.0,
"learning_rate": 8.333333333333334e-06,
"loss": 0.0005,
"step": 4750
},
{
"epoch": 50.0,
"eval_loss": 0.2767828404903412,
"eval_micro_f1": 0.8868238552556532,
"eval_runtime": 3.9956,
"eval_samples_per_second": 251.023,
"step": 4750
},
{
"epoch": 50.53,
"learning_rate": 8.315789473684212e-06,
"loss": 0.0007,
"step": 4800
},
{
"epoch": 50.53,
"eval_loss": 0.2751113176345825,
"eval_micro_f1": 0.8862474079309921,
"eval_runtime": 3.8895,
"eval_samples_per_second": 257.871,
"step": 4800
},
{
"epoch": 51.05,
"learning_rate": 8.29824561403509e-06,
"loss": 0.0004,
"step": 4850
},
{
"epoch": 51.05,
"eval_loss": 0.2764758765697479,
"eval_micro_f1": 0.8864820395326969,
"eval_runtime": 7.8242,
"eval_samples_per_second": 128.192,
"step": 4850
},
{
"epoch": 51.58,
"learning_rate": 8.280701754385965e-06,
"loss": 0.0006,
"step": 4900
},
{
"epoch": 51.58,
"eval_loss": 0.2669425308704376,
"eval_micro_f1": 0.8887368271992756,
"eval_runtime": 3.978,
"eval_samples_per_second": 252.135,
"step": 4900
},
{
"epoch": 52.11,
"learning_rate": 8.263157894736843e-06,
"loss": 0.0003,
"step": 4950
},
{
"epoch": 52.11,
"eval_loss": 0.2705802917480469,
"eval_micro_f1": 0.8861521615140425,
"eval_runtime": 4.0428,
"eval_samples_per_second": 248.098,
"step": 4950
},
{
"epoch": 52.63,
"learning_rate": 8.24561403508772e-06,
"loss": 0.0009,
"step": 5000
},
{
"epoch": 52.63,
"eval_loss": 0.27194783091545105,
"eval_micro_f1": 0.8902976841745119,
"eval_runtime": 3.2158,
"eval_samples_per_second": 311.894,
"step": 5000
},
{
"epoch": 53.16,
"learning_rate": 8.228070175438597e-06,
"loss": 0.0006,
"step": 5050
},
{
"epoch": 53.16,
"eval_loss": 0.2671356499195099,
"eval_micro_f1": 0.8891352544886696,
"eval_runtime": 4.0034,
"eval_samples_per_second": 250.538,
"step": 5050
},
{
"epoch": 53.68,
"learning_rate": 8.210526315789475e-06,
"loss": 0.0003,
"step": 5100
},
{
"epoch": 53.68,
"eval_loss": 0.26829880475997925,
"eval_micro_f1": 0.8904690272586036,
"eval_runtime": 7.7338,
"eval_samples_per_second": 129.69,
"step": 5100
},
{
"epoch": 54.21,
"learning_rate": 8.192982456140351e-06,
"loss": 0.0003,
"step": 5150
},
{
"epoch": 54.21,
"eval_loss": 0.2754873037338257,
"eval_micro_f1": 0.88907330517057,
"eval_runtime": 3.966,
"eval_samples_per_second": 252.898,
"step": 5150
},
{
"epoch": 54.74,
"learning_rate": 8.175438596491229e-06,
"loss": 0.0003,
"step": 5200
},
{
"epoch": 54.74,
"eval_loss": 0.2686787545681,
"eval_micro_f1": 0.8927837869843981,
"eval_runtime": 4.1014,
"eval_samples_per_second": 244.551,
"step": 5200
},
{
"epoch": 55.26,
"learning_rate": 8.157894736842106e-06,
"loss": 0.0002,
"step": 5250
},
{
"epoch": 55.26,
"eval_loss": 0.28408950567245483,
"eval_micro_f1": 0.8936347242258694,
"eval_runtime": 4.228,
"eval_samples_per_second": 237.226,
"step": 5250
},
{
"epoch": 55.79,
"learning_rate": 8.140350877192983e-06,
"loss": 0.0004,
"step": 5300
},
{
"epoch": 55.79,
"eval_loss": 0.27808433771133423,
"eval_micro_f1": 0.8849435723046118,
"eval_runtime": 4.0129,
"eval_samples_per_second": 249.945,
"step": 5300
},
{
"epoch": 56.32,
"learning_rate": 8.12280701754386e-06,
"loss": 0.0005,
"step": 5350
},
{
"epoch": 56.32,
"eval_loss": 0.2645488381385803,
"eval_micro_f1": 0.8912568301009216,
"eval_runtime": 7.527,
"eval_samples_per_second": 133.253,
"step": 5350
},
{
"epoch": 56.84,
"learning_rate": 8.105263157894736e-06,
"loss": 0.0007,
"step": 5400
},
{
"epoch": 56.84,
"eval_loss": 0.26811301708221436,
"eval_micro_f1": 0.8901494184261513,
"eval_runtime": 4.0185,
"eval_samples_per_second": 249.597,
"step": 5400
},
{
"epoch": 57.37,
"learning_rate": 8.087719298245614e-06,
"loss": 0.0007,
"step": 5450
},
{
"epoch": 57.37,
"eval_loss": 0.24789077043533325,
"eval_micro_f1": 0.8911788948006819,
"eval_runtime": 3.9915,
"eval_samples_per_second": 251.281,
"step": 5450
},
{
"epoch": 57.89,
"learning_rate": 8.070175438596492e-06,
"loss": 0.0007,
"step": 5500
},
{
"epoch": 57.89,
"eval_loss": 0.2630859911441803,
"eval_micro_f1": 0.8869613102528972,
"eval_runtime": 3.8412,
"eval_samples_per_second": 261.116,
"step": 5500
},
{
"epoch": 58.42,
"learning_rate": 8.052631578947368e-06,
"loss": 0.0002,
"step": 5550
},
{
"epoch": 58.42,
"eval_loss": 0.2711201310157776,
"eval_micro_f1": 0.8895933833729507,
"eval_runtime": 4.3155,
"eval_samples_per_second": 232.419,
"step": 5550
},
{
"epoch": 58.95,
"learning_rate": 8.035087719298247e-06,
"loss": 0.0007,
"step": 5600
},
{
"epoch": 58.95,
"eval_loss": 0.27410438656806946,
"eval_micro_f1": 0.8887356950740042,
"eval_runtime": 3.0382,
"eval_samples_per_second": 330.131,
"step": 5600
},
{
"epoch": 59.47,
"learning_rate": 8.017543859649124e-06,
"loss": 0.0006,
"step": 5650
},
{
"epoch": 59.47,
"eval_loss": 0.27450627088546753,
"eval_micro_f1": 0.8883643427377516,
"eval_runtime": 7.9142,
"eval_samples_per_second": 126.733,
"step": 5650
},
{
"epoch": 60.0,
"learning_rate": 8.000000000000001e-06,
"loss": 0.0004,
"step": 5700
},
{
"epoch": 60.0,
"eval_loss": 0.2895042598247528,
"eval_micro_f1": 0.8930115771256787,
"eval_runtime": 3.9822,
"eval_samples_per_second": 251.868,
"step": 5700
},
{
"epoch": 60.53,
"learning_rate": 7.982456140350877e-06,
"loss": 0.0006,
"step": 5750
},
{
"epoch": 60.53,
"eval_loss": 0.2659127712249756,
"eval_micro_f1": 0.895068492650483,
"eval_runtime": 4.0269,
"eval_samples_per_second": 249.073,
"step": 5750
},
{
"epoch": 61.05,
"learning_rate": 7.964912280701755e-06,
"loss": 0.001,
"step": 5800
},
{
"epoch": 61.05,
"eval_loss": 0.25278377532958984,
"eval_micro_f1": 0.8918733721186797,
"eval_runtime": 3.8241,
"eval_samples_per_second": 262.284,
"step": 5800
},
{
"epoch": 61.58,
"learning_rate": 7.947368421052633e-06,
"loss": 0.0011,
"step": 5850
},
{
"epoch": 61.58,
"eval_loss": 0.2676132321357727,
"eval_micro_f1": 0.8920883416859995,
"eval_runtime": 4.6141,
"eval_samples_per_second": 217.379,
"step": 5850
},
{
"epoch": 62.11,
"learning_rate": 7.929824561403509e-06,
"loss": 0.0002,
"step": 5900
},
{
"epoch": 62.11,
"eval_loss": 0.2659474015235901,
"eval_micro_f1": 0.8916862631600105,
"eval_runtime": 3.9094,
"eval_samples_per_second": 256.559,
"step": 5900
},
{
"epoch": 62.63,
"learning_rate": 7.912280701754387e-06,
"loss": 0.0002,
"step": 5950
},
{
"epoch": 62.63,
"eval_loss": 0.2626572549343109,
"eval_micro_f1": 0.8894059671975905,
"eval_runtime": 3.9364,
"eval_samples_per_second": 254.803,
"step": 5950
},
{
"epoch": 63.16,
"learning_rate": 7.894736842105265e-06,
"loss": 0.0005,
"step": 6000
},
{
"epoch": 63.16,
"eval_loss": 0.2726729214191437,
"eval_micro_f1": 0.8897139708969101,
"eval_runtime": 3.9043,
"eval_samples_per_second": 256.893,
"step": 6000
},
{
"epoch": 63.68,
"learning_rate": 7.87719298245614e-06,
"loss": 0.0007,
"step": 6050
},
{
"epoch": 63.68,
"eval_loss": 0.26887959241867065,
"eval_micro_f1": 0.8911218164302537,
"eval_runtime": 3.9588,
"eval_samples_per_second": 253.362,
"step": 6050
},
{
"epoch": 64.21,
"learning_rate": 7.859649122807018e-06,
"loss": 0.0001,
"step": 6100
},
{
"epoch": 64.21,
"eval_loss": 0.2804659605026245,
"eval_micro_f1": 0.889628396615675,
"eval_runtime": 7.0827,
"eval_samples_per_second": 141.613,
"step": 6100
},
{
"epoch": 64.74,
"learning_rate": 7.842105263157895e-06,
"loss": 0.0002,
"step": 6150
},
{
"epoch": 64.74,
"eval_loss": 0.280333548784256,
"eval_micro_f1": 0.8922312676106984,
"eval_runtime": 3.9784,
"eval_samples_per_second": 252.111,
"step": 6150
},
{
"epoch": 65.26,
"learning_rate": 7.824561403508772e-06,
"loss": 0.0002,
"step": 6200
},
{
"epoch": 65.26,
"eval_loss": 0.29168689250946045,
"eval_micro_f1": 0.8890115953031794,
"eval_runtime": 3.2032,
"eval_samples_per_second": 313.124,
"step": 6200
},
{
"epoch": 65.79,
"learning_rate": 7.80701754385965e-06,
"loss": 0.0002,
"step": 6250
},
{
"epoch": 65.79,
"eval_loss": 0.2908705174922943,
"eval_micro_f1": 0.8887349309116997,
"eval_runtime": 4.3807,
"eval_samples_per_second": 228.961,
"step": 6250
},
{
"epoch": 66.32,
"learning_rate": 7.789473684210526e-06,
"loss": 0.0003,
"step": 6300
},
{
"epoch": 66.32,
"eval_loss": 0.28455570340156555,
"eval_micro_f1": 0.8916059893335442,
"eval_runtime": 3.0692,
"eval_samples_per_second": 326.791,
"step": 6300
},
{
"epoch": 66.84,
"learning_rate": 7.771929824561404e-06,
"loss": 0.0003,
"step": 6350
},
{
"epoch": 66.84,
"eval_loss": 0.2848065197467804,
"eval_micro_f1": 0.8897756978622271,
"eval_runtime": 4.0213,
"eval_samples_per_second": 249.424,
"step": 6350
},
{
"epoch": 67.37,
"learning_rate": 7.754385964912282e-06,
"loss": 0.0004,
"step": 6400
},
{
"epoch": 67.37,
"eval_loss": 0.2870292067527771,
"eval_micro_f1": 0.8889809439902276,
"eval_runtime": 4.0404,
"eval_samples_per_second": 248.246,
"step": 6400
},
{
"epoch": 67.89,
"learning_rate": 7.736842105263158e-06,
"loss": 0.0015,
"step": 6450
},
{
"epoch": 67.89,
"eval_loss": 0.2784861922264099,
"eval_micro_f1": 0.8912922901264898,
"eval_runtime": 3.9428,
"eval_samples_per_second": 254.389,
"step": 6450
},
{
"epoch": 68.42,
"learning_rate": 7.719298245614036e-06,
"loss": 0.0005,
"step": 6500
},
{
"epoch": 68.42,
"eval_loss": 0.28143617510795593,
"eval_micro_f1": 0.8934177907418446,
"eval_runtime": 3.9587,
"eval_samples_per_second": 253.364,
"step": 6500
},
{
"epoch": 68.95,
"learning_rate": 7.701754385964912e-06,
"loss": 0.0003,
"step": 6550
},
{
"epoch": 68.95,
"eval_loss": 0.2798239290714264,
"eval_micro_f1": 0.8936757796709776,
"eval_runtime": 3.1753,
"eval_samples_per_second": 315.876,
"step": 6550
},
{
"epoch": 69.47,
"learning_rate": 7.68421052631579e-06,
"loss": 0.0001,
"step": 6600
},
{
"epoch": 69.47,
"eval_loss": 0.28019773960113525,
"eval_micro_f1": 0.8973865194447462,
"eval_runtime": 4.0594,
"eval_samples_per_second": 247.081,
"step": 6600
},
{
"epoch": 70.0,
"learning_rate": 7.666666666666667e-06,
"loss": 0.0005,
"step": 6650
},
{
"epoch": 70.0,
"eval_loss": 0.2975628674030304,
"eval_micro_f1": 0.8937831051591087,
"eval_runtime": 9.2455,
"eval_samples_per_second": 108.485,
"step": 6650
},
{
"epoch": 70.53,
"learning_rate": 7.649122807017545e-06,
"loss": 0.0003,
"step": 6700
},
{
"epoch": 70.53,
"eval_loss": 0.2941923141479492,
"eval_micro_f1": 0.8903672898670463,
"eval_runtime": 3.9883,
"eval_samples_per_second": 251.483,
"step": 6700
},
{
"epoch": 71.05,
"learning_rate": 7.631578947368423e-06,
"loss": 0.0012,
"step": 6750
},
{
"epoch": 71.05,
"eval_loss": 0.2895115315914154,
"eval_micro_f1": 0.8915230870256033,
"eval_runtime": 4.0799,
"eval_samples_per_second": 245.842,
"step": 6750
},
{
"epoch": 71.58,
"learning_rate": 7.614035087719299e-06,
"loss": 0.0012,
"step": 6800
},
{
"epoch": 71.58,
"eval_loss": 0.267494261264801,
"eval_micro_f1": 0.89286705473151,
"eval_runtime": 3.1513,
"eval_samples_per_second": 318.281,
"step": 6800
},
{
"epoch": 72.11,
"learning_rate": 7.596491228070177e-06,
"loss": 0.0008,
"step": 6850
},
{
"epoch": 72.11,
"eval_loss": 0.26893123984336853,
"eval_micro_f1": 0.8880340701138721,
"eval_runtime": 3.9688,
"eval_samples_per_second": 252.723,
"step": 6850
},
{
"epoch": 72.63,
"learning_rate": 7.578947368421054e-06,
"loss": 0.0001,
"step": 6900
},
{
"epoch": 72.63,
"eval_loss": 0.2776913046836853,
"eval_micro_f1": 0.8887043184366316,
"eval_runtime": 6.9318,
"eval_samples_per_second": 144.695,
"step": 6900
},
{
"epoch": 73.16,
"learning_rate": 7.5614035087719305e-06,
"loss": 0.0003,
"step": 6950
},
{
"epoch": 73.16,
"eval_loss": 0.28223922848701477,
"eval_micro_f1": 0.8916228913991177,
"eval_runtime": 4.0849,
"eval_samples_per_second": 245.537,
"step": 6950
},
{
"epoch": 73.68,
"learning_rate": 7.5438596491228074e-06,
"loss": 0.0001,
"step": 7000
},
{
"epoch": 73.68,
"eval_loss": 0.2847548723220825,
"eval_micro_f1": 0.8945695725272507,
"eval_runtime": 3.2705,
"eval_samples_per_second": 306.684,
"step": 7000
},
{
"epoch": 74.21,
"learning_rate": 7.526315789473685e-06,
"loss": 0.0002,
"step": 7050
},
{
"epoch": 74.21,
"eval_loss": 0.2813137173652649,
"eval_micro_f1": 0.8936170207763768,
"eval_runtime": 4.1172,
"eval_samples_per_second": 243.614,
"step": 7050
},
{
"epoch": 74.74,
"learning_rate": 7.508771929824562e-06,
"loss": 0.0002,
"step": 7100
},
{
"epoch": 74.74,
"eval_loss": 0.2780473530292511,
"eval_micro_f1": 0.8961056827253051,
"eval_runtime": 3.3844,
"eval_samples_per_second": 296.361,
"step": 7100
},
{
"epoch": 75.26,
"learning_rate": 7.491228070175439e-06,
"loss": 0.0002,
"step": 7150
},
{
"epoch": 75.26,
"eval_loss": 0.2878120541572571,
"eval_micro_f1": 0.8926741443550371,
"eval_runtime": 4.2487,
"eval_samples_per_second": 236.073,
"step": 7150
},
{
"epoch": 75.79,
"learning_rate": 7.473684210526316e-06,
"loss": 0.0002,
"step": 7200
},
{
"epoch": 75.79,
"eval_loss": 0.27698081731796265,
"eval_micro_f1": 0.8874897339646318,
"eval_runtime": 3.7419,
"eval_samples_per_second": 268.044,
"step": 7200
},
{
"epoch": 76.32,
"learning_rate": 7.456140350877194e-06,
"loss": 0.0005,
"step": 7250
},
{
"epoch": 76.32,
"eval_loss": 0.27582716941833496,
"eval_micro_f1": 0.8924820337728324,
"eval_runtime": 3.7947,
"eval_samples_per_second": 264.318,
"step": 7250
},
{
"epoch": 76.84,
"learning_rate": 7.438596491228071e-06,
"loss": 0.0001,
"step": 7300
},
{
"epoch": 76.84,
"eval_loss": 0.27750343084335327,
"eval_micro_f1": 0.8934460297062115,
"eval_runtime": 3.8719,
"eval_samples_per_second": 259.046,
"step": 7300
},
{
"epoch": 77.37,
"learning_rate": 7.421052631578948e-06,
"loss": 0.0001,
"step": 7350
},
{
"epoch": 77.37,
"eval_loss": 0.2822408676147461,
"eval_micro_f1": 0.8927240088881296,
"eval_runtime": 4.0645,
"eval_samples_per_second": 246.769,
"step": 7350
},
{
"epoch": 77.89,
"learning_rate": 7.4035087719298246e-06,
"loss": 0.0005,
"step": 7400
},
{
"epoch": 77.89,
"eval_loss": 0.28152230381965637,
"eval_micro_f1": 0.8936757796709776,
"eval_runtime": 3.7022,
"eval_samples_per_second": 270.923,
"step": 7400
},
{
"epoch": 78.42,
"learning_rate": 7.385964912280702e-06,
"loss": 0.0002,
"step": 7450
},
{
"epoch": 78.42,
"eval_loss": 0.2781830430030823,
"eval_micro_f1": 0.8965137103995797,
"eval_runtime": 4.0322,
"eval_samples_per_second": 248.749,
"step": 7450
},
{
"epoch": 78.95,
"learning_rate": 7.368421052631579e-06,
"loss": 0.0003,
"step": 7500
},
{
"epoch": 78.95,
"eval_loss": 0.2824631929397583,
"eval_micro_f1": 0.8959513605609658,
"eval_runtime": 3.2588,
"eval_samples_per_second": 307.784,
"step": 7500
},
{
"epoch": 79.47,
"learning_rate": 7.350877192982456e-06,
"loss": 0.0004,
"step": 7550
},
{
"epoch": 79.47,
"eval_loss": 0.2894085645675659,
"eval_micro_f1": 0.8889806020585798,
"eval_runtime": 4.5841,
"eval_samples_per_second": 218.802,
"step": 7550
},
{
"epoch": 80.0,
"learning_rate": 7.333333333333333e-06,
"loss": 0.0003,
"step": 7600
},
{
"epoch": 80.0,
"eval_loss": 0.2869914174079895,
"eval_micro_f1": 0.8926027392258261,
"eval_runtime": 3.1583,
"eval_samples_per_second": 317.574,
"step": 7600
},
{
"epoch": 80.53,
"learning_rate": 7.315789473684212e-06,
"loss": 0.0003,
"step": 7650
},
{
"epoch": 80.53,
"eval_loss": 0.2831559181213379,
"eval_micro_f1": 0.8920174953992206,
"eval_runtime": 3.836,
"eval_samples_per_second": 261.472,
"step": 7650
},
{
"epoch": 81.05,
"learning_rate": 7.298245614035089e-06,
"loss": 0.0002,
"step": 7700
},
{
"epoch": 81.05,
"eval_loss": 0.2863539457321167,
"eval_micro_f1": 0.8948891026820168,
"eval_runtime": 7.0889,
"eval_samples_per_second": 141.489,
"step": 7700
},
{
"epoch": 81.58,
"learning_rate": 7.280701754385966e-06,
"loss": 0.0004,
"step": 7750
},
{
"epoch": 81.58,
"eval_loss": 0.29815465211868286,
"eval_micro_f1": 0.8911676232351213,
"eval_runtime": 8.2897,
"eval_samples_per_second": 120.994,
"step": 7750
},
{
"epoch": 82.11,
"learning_rate": 7.263157894736843e-06,
"loss": 0.0007,
"step": 7800
},
{
"epoch": 82.11,
"eval_loss": 0.27996498346328735,
"eval_micro_f1": 0.8902472149305036,
"eval_runtime": 3.1418,
"eval_samples_per_second": 319.239,
"step": 7800
},
{
"epoch": 82.63,
"learning_rate": 7.24561403508772e-06,
"loss": 0.0004,
"step": 7850
},
{
"epoch": 82.63,
"eval_loss": 0.27985525131225586,
"eval_micro_f1": 0.8931476625357428,
"eval_runtime": 3.8998,
"eval_samples_per_second": 257.195,
"step": 7850
},
{
"epoch": 83.16,
"learning_rate": 7.228070175438597e-06,
"loss": 0.0005,
"step": 7900
},
{
"epoch": 83.16,
"eval_loss": 0.28243011236190796,
"eval_micro_f1": 0.8924686763550573,
"eval_runtime": 7.1825,
"eval_samples_per_second": 139.644,
"step": 7900
},
{
"epoch": 83.68,
"learning_rate": 7.210526315789474e-06,
"loss": 0.0003,
"step": 7950
},
{
"epoch": 83.68,
"eval_loss": 0.28672757744789124,
"eval_micro_f1": 0.8895004811290499,
"eval_runtime": 4.6325,
"eval_samples_per_second": 216.514,
"step": 7950
},
{
"epoch": 84.21,
"learning_rate": 7.192982456140352e-06,
"loss": 0.0005,
"step": 8000
},
{
"epoch": 84.21,
"eval_loss": 0.28343382477760315,
"eval_micro_f1": 0.8904203992240877,
"eval_runtime": 3.1489,
"eval_samples_per_second": 318.521,
"step": 8000
},
{
"epoch": 84.74,
"learning_rate": 7.175438596491229e-06,
"loss": 0.0008,
"step": 8050
},
{
"epoch": 84.74,
"eval_loss": 0.26917701959609985,
"eval_micro_f1": 0.892177010049188,
"eval_runtime": 8.5597,
"eval_samples_per_second": 117.177,
"step": 8050
},
{
"epoch": 85.26,
"learning_rate": 7.157894736842106e-06,
"loss": 0.0007,
"step": 8100
},
{
"epoch": 85.26,
"eval_loss": 0.27744022011756897,
"eval_micro_f1": 0.8899600600974766,
"eval_runtime": 3.1179,
"eval_samples_per_second": 321.691,
"step": 8100
},
{
"epoch": 85.79,
"learning_rate": 7.140350877192983e-06,
"loss": 0.0003,
"step": 8150
},
{
"epoch": 85.79,
"eval_loss": 0.27590543031692505,
"eval_micro_f1": 0.8898956612241085,
"eval_runtime": 4.1349,
"eval_samples_per_second": 242.567,
"step": 8150
},
{
"epoch": 86.32,
"learning_rate": 7.1228070175438605e-06,
"loss": 0.0004,
"step": 8200
},
{
"epoch": 86.32,
"eval_loss": 0.2814866602420807,
"eval_micro_f1": 0.8921339022411855,
"eval_runtime": 3.0821,
"eval_samples_per_second": 325.424,
"step": 8200
},
{
"epoch": 86.84,
"learning_rate": 7.1052631578947375e-06,
"loss": 0.0008,
"step": 8250
},
{
"epoch": 86.84,
"eval_loss": 0.29091206192970276,
"eval_micro_f1": 0.8922651928699199,
"eval_runtime": 8.1826,
"eval_samples_per_second": 122.578,
"step": 8250
},
{
"epoch": 87.37,
"learning_rate": 7.087719298245614e-06,
"loss": 0.0003,
"step": 8300
},
{
"epoch": 87.37,
"eval_loss": 0.2864170968532562,
"eval_micro_f1": 0.893517240879067,
"eval_runtime": 3.1741,
"eval_samples_per_second": 315.994,
"step": 8300
},
{
"epoch": 87.89,
"learning_rate": 7.070175438596491e-06,
"loss": 0.0002,
"step": 8350
},
{
"epoch": 87.89,
"eval_loss": 0.2824071943759918,
"eval_micro_f1": 0.8926708751517175,
"eval_runtime": 4.7943,
"eval_samples_per_second": 209.207,
"step": 8350
},
{
"epoch": 88.42,
"learning_rate": 7.052631578947369e-06,
"loss": 0.0008,
"step": 8400
},
{
"epoch": 88.42,
"eval_loss": 0.28318047523498535,
"eval_micro_f1": 0.8900767538857594,
"eval_runtime": 2.9559,
"eval_samples_per_second": 339.316,
"step": 8400
},
{
"epoch": 88.95,
"learning_rate": 7.035087719298246e-06,
"loss": 0.0001,
"step": 8450
},
{
"epoch": 88.95,
"eval_loss": 0.2841491103172302,
"eval_micro_f1": 0.8939101675901431,
"eval_runtime": 7.7315,
"eval_samples_per_second": 129.73,
"step": 8450
},
{
"epoch": 89.47,
"learning_rate": 7.017543859649123e-06,
"loss": 0.0002,
"step": 8500
},
{
"epoch": 89.47,
"eval_loss": 0.2948380410671234,
"eval_micro_f1": 0.891776087958634,
"eval_runtime": 3.2961,
"eval_samples_per_second": 304.299,
"step": 8500
},
{
"epoch": 90.0,
"learning_rate": 7e-06,
"loss": 0.0003,
"step": 8550
},
{
"epoch": 90.0,
"eval_loss": 0.31475064158439636,
"eval_micro_f1": 0.8908309053119371,
"eval_runtime": 4.0607,
"eval_samples_per_second": 247.004,
"step": 8550
},
{
"epoch": 90.53,
"learning_rate": 6.982456140350878e-06,
"loss": 0.0002,
"step": 8600
},
{
"epoch": 90.53,
"eval_loss": 0.29183682799339294,
"eval_micro_f1": 0.8886145399661801,
"eval_runtime": 3.4404,
"eval_samples_per_second": 291.538,
"step": 8600
},
{
"epoch": 91.05,
"learning_rate": 6.964912280701755e-06,
"loss": 0.0002,
"step": 8650
},
{
"epoch": 91.05,
"eval_loss": 0.293345183134079,
"eval_micro_f1": 0.8882175221583899,
"eval_runtime": 3.8539,
"eval_samples_per_second": 260.254,
"step": 8650
},
{
"epoch": 91.58,
"learning_rate": 6.947368421052632e-06,
"loss": 0.0002,
"step": 8700
},
{
"epoch": 91.58,
"eval_loss": 0.2861791253089905,
"eval_micro_f1": 0.8892876858608822,
"eval_runtime": 3.9934,
"eval_samples_per_second": 251.164,
"step": 8700
},
{
"epoch": 92.11,
"learning_rate": 6.92982456140351e-06,
"loss": 0.0001,
"step": 8750
},
{
"epoch": 92.11,
"eval_loss": 0.29430919885635376,
"eval_micro_f1": 0.8907911081563892,
"eval_runtime": 7.6515,
"eval_samples_per_second": 131.086,
"step": 8750
},
{
"epoch": 92.63,
"learning_rate": 6.912280701754387e-06,
"loss": 0.0001,
"step": 8800
},
{
"epoch": 92.63,
"eval_loss": 0.2997332215309143,
"eval_micro_f1": 0.8926397338661063,
"eval_runtime": 3.2351,
"eval_samples_per_second": 310.036,
"step": 8800
},
{
"epoch": 93.16,
"learning_rate": 6.894736842105264e-06,
"loss": 0.0,
"step": 8850
},
{
"epoch": 93.16,
"eval_loss": 0.30131086707115173,
"eval_micro_f1": 0.893069580354643,
"eval_runtime": 3.8296,
"eval_samples_per_second": 261.91,
"step": 8850
},
{
"epoch": 93.68,
"learning_rate": 6.877192982456141e-06,
"loss": 0.0,
"step": 8900
},
{
"epoch": 93.68,
"eval_loss": 0.30136561393737793,
"eval_micro_f1": 0.8927488277324576,
"eval_runtime": 3.1162,
"eval_samples_per_second": 321.87,
"step": 8900
},
{
"epoch": 94.21,
"learning_rate": 6.859649122807019e-06,
"loss": 0.0002,
"step": 8950
},
{
"epoch": 94.21,
"eval_loss": 0.30289435386657715,
"eval_micro_f1": 0.8931181900941578,
"eval_runtime": 3.7531,
"eval_samples_per_second": 267.243,
"step": 8950
},
{
"epoch": 94.74,
"learning_rate": 6.842105263157896e-06,
"loss": 0.0,
"step": 9000
},
{
"epoch": 94.74,
"eval_loss": 0.30314457416534424,
"eval_micro_f1": 0.8904690272586036,
"eval_runtime": 8.0675,
"eval_samples_per_second": 124.325,
"step": 9000
},
{
"epoch": 95.26,
"learning_rate": 6.824561403508773e-06,
"loss": 0.0002,
"step": 9050
},
{
"epoch": 95.26,
"eval_loss": 0.2991909980773926,
"eval_micro_f1": 0.8936639113454912,
"eval_runtime": 4.1382,
"eval_samples_per_second": 242.375,
"step": 9050
},
{
"epoch": 95.79,
"learning_rate": 6.8070175438596495e-06,
"loss": 0.0001,
"step": 9100
},
{
"epoch": 95.79,
"eval_loss": 0.30111491680145264,
"eval_micro_f1": 0.8924642459244121,
"eval_runtime": 7.7751,
"eval_samples_per_second": 129.001,
"step": 9100
},
{
"epoch": 96.32,
"learning_rate": 6.789473684210527e-06,
"loss": 0.0,
"step": 9150
},
{
"epoch": 96.32,
"eval_loss": 0.3051038086414337,
"eval_micro_f1": 0.892303432500861,
"eval_runtime": 3.1315,
"eval_samples_per_second": 320.296,
"step": 9150
},
{
"epoch": 96.84,
"learning_rate": 6.771929824561404e-06,
"loss": 0.0004,
"step": 9200
},
{
"epoch": 96.84,
"eval_loss": 0.30604293942451477,
"eval_micro_f1": 0.8915529828262037,
"eval_runtime": 3.9084,
"eval_samples_per_second": 256.627,
"step": 9200
},
{
"epoch": 97.37,
"learning_rate": 6.754385964912281e-06,
"loss": 0.0005,
"step": 9250
},
{
"epoch": 97.37,
"eval_loss": 0.28967851400375366,
"eval_micro_f1": 0.8916586763933765,
"eval_runtime": 3.0891,
"eval_samples_per_second": 324.693,
"step": 9250
},
{
"epoch": 97.89,
"learning_rate": 6.736842105263158e-06,
"loss": 0.0004,
"step": 9300
},
{
"epoch": 97.89,
"eval_loss": 0.28513702750205994,
"eval_micro_f1": 0.8911788948006819,
"eval_runtime": 4.1021,
"eval_samples_per_second": 244.506,
"step": 9300
},
{
"epoch": 98.42,
"learning_rate": 6.719298245614035e-06,
"loss": 0.0002,
"step": 9350
},
{
"epoch": 98.42,
"eval_loss": 0.29083964228630066,
"eval_micro_f1": 0.8888888883886673,
"eval_runtime": 3.1048,
"eval_samples_per_second": 323.047,
"step": 9350
},
{
"epoch": 98.95,
"learning_rate": 6.701754385964913e-06,
"loss": 0.0005,
"step": 9400
},
{
"epoch": 98.95,
"eval_loss": 0.29221805930137634,
"eval_micro_f1": 0.8874226799121446,
"eval_runtime": 8.322,
"eval_samples_per_second": 120.524,
"step": 9400
},
{
"epoch": 99.47,
"learning_rate": 6.68421052631579e-06,
"loss": 0.0008,
"step": 9450
},
{
"epoch": 99.47,
"eval_loss": 0.28144344687461853,
"eval_micro_f1": 0.8876772077877538,
"eval_runtime": 3.2697,
"eval_samples_per_second": 306.759,
"step": 9450
},
{
"epoch": 100.0,
"learning_rate": 6.666666666666667e-06,
"loss": 0.0005,
"step": 9500
},
{
"epoch": 100.0,
"eval_loss": 0.29380202293395996,
"eval_micro_f1": 0.8887966799976794,
"eval_runtime": 4.1414,
"eval_samples_per_second": 242.187,
"step": 9500
},
{
"epoch": 100.53,
"learning_rate": 6.649122807017544e-06,
"loss": 0.0001,
"step": 9550
},
{
"epoch": 100.53,
"eval_loss": 0.2875373065471649,
"eval_micro_f1": 0.8876048970097425,
"eval_runtime": 3.5444,
"eval_samples_per_second": 282.98,
"step": 9550
},
{
"epoch": 101.05,
"learning_rate": 6.631578947368421e-06,
"loss": 0.0002,
"step": 9600
},
{
"epoch": 101.05,
"eval_loss": 0.28619715571403503,
"eval_micro_f1": 0.8927639703908955,
"eval_runtime": 3.9021,
"eval_samples_per_second": 257.041,
"step": 9600
},
{
"epoch": 101.58,
"learning_rate": 6.614035087719298e-06,
"loss": 0.0002,
"step": 9650
},
{
"epoch": 101.58,
"eval_loss": 0.2913980185985565,
"eval_micro_f1": 0.8870834475856909,
"eval_runtime": 3.1769,
"eval_samples_per_second": 315.716,
"step": 9650
},
{
"epoch": 102.11,
"learning_rate": 6.596491228070177e-06,
"loss": 0.0005,
"step": 9700
},
{
"epoch": 102.11,
"eval_loss": 0.2946756184101105,
"eval_micro_f1": 0.8876606323586742,
"eval_runtime": 9.0019,
"eval_samples_per_second": 111.421,
"step": 9700
},
{
"epoch": 102.63,
"learning_rate": 6.578947368421054e-06,
"loss": 0.0002,
"step": 9750
},
{
"epoch": 102.63,
"eval_loss": 0.29088717699050903,
"eval_micro_f1": 0.8892534859641347,
"eval_runtime": 3.1417,
"eval_samples_per_second": 319.256,
"step": 9750
},
{
"epoch": 103.16,
"learning_rate": 6.561403508771931e-06,
"loss": 0.0,
"step": 9800
},
{
"epoch": 103.16,
"eval_loss": 0.2949836552143097,
"eval_micro_f1": 0.8900811660976494,
"eval_runtime": 7.9049,
"eval_samples_per_second": 126.883,
"step": 9800
},
{
"epoch": 103.68,
"learning_rate": 6.543859649122808e-06,
"loss": 0.0001,
"step": 9850
},
{
"epoch": 103.68,
"eval_loss": 0.2967689335346222,
"eval_micro_f1": 0.8890708140720921,
"eval_runtime": 7.05,
"eval_samples_per_second": 142.27,
"step": 9850
},
{
"epoch": 104.21,
"learning_rate": 6.526315789473685e-06,
"loss": 0.0001,
"step": 9900
},
{
"epoch": 104.21,
"eval_loss": 0.31163182854652405,
"eval_micro_f1": 0.8898981552938755,
"eval_runtime": 3.9986,
"eval_samples_per_second": 250.84,
"step": 9900
},
{
"epoch": 104.74,
"learning_rate": 6.508771929824562e-06,
"loss": 0.0002,
"step": 9950
},
{
"epoch": 104.74,
"eval_loss": 0.30857715010643005,
"eval_micro_f1": 0.8909965630736558,
"eval_runtime": 3.0819,
"eval_samples_per_second": 325.453,
"step": 9950
},
{
"epoch": 105.26,
"learning_rate": 6.491228070175439e-06,
"loss": 0.0007,
"step": 10000
},
{
"epoch": 105.26,
"eval_loss": 0.30797579884529114,
"eval_micro_f1": 0.8926119193020305,
"eval_runtime": 8.4523,
"eval_samples_per_second": 118.666,
"step": 10000
},
{
"epoch": 105.79,
"learning_rate": 6.473684210526316e-06,
"loss": 0.0008,
"step": 10050
},
{
"epoch": 105.79,
"eval_loss": 0.28551915287971497,
"eval_micro_f1": 0.8895269341457042,
"eval_runtime": 3.6854,
"eval_samples_per_second": 272.152,
"step": 10050
},
{
"epoch": 106.32,
"learning_rate": 6.456140350877193e-06,
"loss": 0.0003,
"step": 10100
},
{
"epoch": 106.32,
"eval_loss": 0.2874017655849457,
"eval_micro_f1": 0.8948965512238941,
"eval_runtime": 3.8385,
"eval_samples_per_second": 261.303,
"step": 10100
},
{
"epoch": 106.84,
"learning_rate": 6.438596491228071e-06,
"loss": 0.0001,
"step": 10150
},
{
"epoch": 106.84,
"eval_loss": 0.28866690397262573,
"eval_micro_f1": 0.8892240192420461,
"eval_runtime": 3.1417,
"eval_samples_per_second": 319.249,
"step": 10150
},
{
"epoch": 107.37,
"learning_rate": 6.421052631578948e-06,
"loss": 0.0001,
"step": 10200
},
{
"epoch": 107.37,
"eval_loss": 0.29645153880119324,
"eval_micro_f1": 0.8903563071074169,
"eval_runtime": 8.2788,
"eval_samples_per_second": 121.152,
"step": 10200
},
{
"epoch": 107.89,
"learning_rate": 6.403508771929825e-06,
"loss": 0.0004,
"step": 10250
},
{
"epoch": 107.89,
"eval_loss": 0.28964516520500183,
"eval_micro_f1": 0.8920547940203469,
"eval_runtime": 3.3488,
"eval_samples_per_second": 299.513,
"step": 10250
},
{
"epoch": 108.42,
"learning_rate": 6.385964912280702e-06,
"loss": 0.0,
"step": 10300
},
{
"epoch": 108.42,
"eval_loss": 0.29358983039855957,
"eval_micro_f1": 0.8941726218197285,
"eval_runtime": 4.0373,
"eval_samples_per_second": 248.435,
"step": 10300
},
{
"epoch": 108.95,
"learning_rate": 6.3684210526315795e-06,
"loss": 0.0001,
"step": 10350
},
{
"epoch": 108.95,
"eval_loss": 0.29186615347862244,
"eval_micro_f1": 0.8937919688022478,
"eval_runtime": 7.8983,
"eval_samples_per_second": 126.989,
"step": 10350
},
{
"epoch": 109.47,
"learning_rate": 6.3508771929824565e-06,
"loss": 0.0001,
"step": 10400
},
{
"epoch": 109.47,
"eval_loss": 0.2970679998397827,
"eval_micro_f1": 0.8943807809439072,
"eval_runtime": 8.6868,
"eval_samples_per_second": 115.462,
"step": 10400
},
{
"epoch": 110.0,
"learning_rate": 6.333333333333333e-06,
"loss": 0.0002,
"step": 10450
},
{
"epoch": 110.0,
"eval_loss": 0.2990773022174835,
"eval_micro_f1": 0.8954275447408084,
"eval_runtime": 4.0125,
"eval_samples_per_second": 249.967,
"step": 10450
},
{
"epoch": 110.53,
"learning_rate": 6.31578947368421e-06,
"loss": 0.0,
"step": 10500
},
{
"epoch": 110.53,
"eval_loss": 0.3070140480995178,
"eval_micro_f1": 0.8945112673001121,
"eval_runtime": 3.9494,
"eval_samples_per_second": 253.965,
"step": 10500
},
{
"epoch": 111.05,
"learning_rate": 6.298245614035088e-06,
"loss": 0.0003,
"step": 10550
},
{
"epoch": 111.05,
"eval_loss": 0.2980208396911621,
"eval_micro_f1": 0.8904109584039089,
"eval_runtime": 4.1953,
"eval_samples_per_second": 239.08,
"step": 10550
},
{
"epoch": 111.58,
"learning_rate": 6.280701754385965e-06,
"loss": 0.0005,
"step": 10600
},
{
"epoch": 111.58,
"eval_loss": 0.3040313124656677,
"eval_micro_f1": 0.8894444439442073,
"eval_runtime": 4.0291,
"eval_samples_per_second": 248.936,
"step": 10600
},
{
"epoch": 112.11,
"learning_rate": 6.263157894736842e-06,
"loss": 0.0002,
"step": 10650
},
{
"epoch": 112.11,
"eval_loss": 0.2912485897541046,
"eval_micro_f1": 0.8896816679959434,
"eval_runtime": 2.9919,
"eval_samples_per_second": 335.24,
"step": 10650
},
{
"epoch": 112.63,
"learning_rate": 6.245614035087721e-06,
"loss": 0.0006,
"step": 10700
},
{
"epoch": 112.63,
"eval_loss": 0.3086911141872406,
"eval_micro_f1": 0.8893187779991326,
"eval_runtime": 4.04,
"eval_samples_per_second": 248.266,
"step": 10700
},
{
"epoch": 113.16,
"learning_rate": 6.2280701754385975e-06,
"loss": 0.0,
"step": 10750
},
{
"epoch": 113.16,
"eval_loss": 0.29828786849975586,
"eval_micro_f1": 0.8922822941757889,
"eval_runtime": 3.1424,
"eval_samples_per_second": 319.186,
"step": 10750
},
{
"epoch": 113.68,
"learning_rate": 6.2105263157894745e-06,
"loss": 0.0,
"step": 10800
},
{
"epoch": 113.68,
"eval_loss": 0.3017314374446869,
"eval_micro_f1": 0.8927097656620789,
"eval_runtime": 7.965,
"eval_samples_per_second": 125.925,
"step": 10800
},
{
"epoch": 114.21,
"learning_rate": 6.192982456140351e-06,
"loss": 0.0001,
"step": 10850
},
{
"epoch": 114.21,
"eval_loss": 0.3105458915233612,
"eval_micro_f1": 0.892120204457516,
"eval_runtime": 2.9845,
"eval_samples_per_second": 336.071,
"step": 10850
},
{
"epoch": 114.74,
"learning_rate": 6.175438596491229e-06,
"loss": 0.0002,
"step": 10900
},
{
"epoch": 114.74,
"eval_loss": 0.30062082409858704,
"eval_micro_f1": 0.8916323726136417,
"eval_runtime": 8.4769,
"eval_samples_per_second": 118.321,
"step": 10900
},
{
"epoch": 115.26,
"learning_rate": 6.157894736842106e-06,
"loss": 0.0004,
"step": 10950
},
{
"epoch": 115.26,
"eval_loss": 0.30511754751205444,
"eval_micro_f1": 0.8859467654630745,
"eval_runtime": 3.2692,
"eval_samples_per_second": 306.806,
"step": 10950
},
{
"epoch": 115.79,
"learning_rate": 6.140350877192983e-06,
"loss": 0.0002,
"step": 11000
},
{
"epoch": 115.79,
"eval_loss": 0.30255937576293945,
"eval_micro_f1": 0.8904544821631056,
"eval_runtime": 3.9666,
"eval_samples_per_second": 252.859,
"step": 11000
},
{
"epoch": 116.32,
"learning_rate": 6.12280701754386e-06,
"loss": 0.0003,
"step": 11050
},
{
"epoch": 116.32,
"eval_loss": 0.29517313838005066,
"eval_micro_f1": 0.8859925910761354,
"eval_runtime": 7.5463,
"eval_samples_per_second": 132.912,
"step": 11050
},
{
"epoch": 116.84,
"learning_rate": 6.105263157894738e-06,
"loss": 0.0004,
"step": 11100
},
{
"epoch": 116.84,
"eval_loss": 0.29613345861434937,
"eval_micro_f1": 0.8932280842525581,
"eval_runtime": 3.9637,
"eval_samples_per_second": 253.043,
"step": 11100
},
{
"epoch": 117.37,
"learning_rate": 6.087719298245615e-06,
"loss": 0.0001,
"step": 11150
},
{
"epoch": 117.37,
"eval_loss": 0.2902396321296692,
"eval_micro_f1": 0.892036858253723,
"eval_runtime": 3.7992,
"eval_samples_per_second": 264.003,
"step": 11150
},
{
"epoch": 117.89,
"learning_rate": 6.070175438596492e-06,
"loss": 0.0,
"step": 11200
},
{
"epoch": 117.89,
"eval_loss": 0.29245704412460327,
"eval_micro_f1": 0.8922484876800918,
"eval_runtime": 3.2414,
"eval_samples_per_second": 309.433,
"step": 11200
},
{
"epoch": 118.42,
"learning_rate": 6.0526315789473685e-06,
"loss": 0.0,
"step": 11250
},
{
"epoch": 118.42,
"eval_loss": 0.29434511065483093,
"eval_micro_f1": 0.8918807523504421,
"eval_runtime": 7.7527,
"eval_samples_per_second": 129.374,
"step": 11250
},
{
"epoch": 118.95,
"learning_rate": 6.035087719298246e-06,
"loss": 0.0001,
"step": 11300
},
{
"epoch": 118.95,
"eval_loss": 0.3034317195415497,
"eval_micro_f1": 0.8932119200295573,
"eval_runtime": 3.6067,
"eval_samples_per_second": 278.094,
"step": 11300
},
{
"epoch": 119.47,
"learning_rate": 6.017543859649123e-06,
"loss": 0.0,
"step": 11350
},
{
"epoch": 119.47,
"eval_loss": 0.30254286527633667,
"eval_micro_f1": 0.8921676773817213,
"eval_runtime": 4.2931,
"eval_samples_per_second": 233.629,
"step": 11350
},
{
"epoch": 120.0,
"learning_rate": 6e-06,
"loss": 0.0002,
"step": 11400
},
{
"epoch": 120.0,
"eval_loss": 0.2985408902168274,
"eval_micro_f1": 0.886975535921856,
"eval_runtime": 4.074,
"eval_samples_per_second": 246.195,
"step": 11400
},
{
"epoch": 120.53,
"learning_rate": 5.982456140350877e-06,
"loss": 0.0002,
"step": 11450
},
{
"epoch": 120.53,
"eval_loss": 0.31120577454566956,
"eval_micro_f1": 0.8919559454080521,
"eval_runtime": 7.5617,
"eval_samples_per_second": 132.642,
"step": 11450
},
{
"epoch": 121.05,
"learning_rate": 5.964912280701755e-06,
"loss": 0.0001,
"step": 11500
},
{
"epoch": 121.05,
"eval_loss": 0.3012202978134155,
"eval_micro_f1": 0.8929705836732108,
"eval_runtime": 7.7225,
"eval_samples_per_second": 129.881,
"step": 11500
},
{
"epoch": 121.58,
"learning_rate": 5.947368421052632e-06,
"loss": 0.0001,
"step": 11550
},
{
"epoch": 121.58,
"eval_loss": 0.3262581527233124,
"eval_micro_f1": 0.8914728677168081,
"eval_runtime": 2.9754,
"eval_samples_per_second": 337.103,
"step": 11550
},
{
"epoch": 122.11,
"learning_rate": 5.929824561403509e-06,
"loss": 0.0002,
"step": 11600
},
{
"epoch": 122.11,
"eval_loss": 0.30825820565223694,
"eval_micro_f1": 0.8931771187278773,
"eval_runtime": 4.2577,
"eval_samples_per_second": 235.574,
"step": 11600
},
{
"epoch": 122.63,
"learning_rate": 5.912280701754386e-06,
"loss": 0.0001,
"step": 11650
},
{
"epoch": 122.63,
"eval_loss": 0.3162708282470703,
"eval_micro_f1": 0.8928274711142647,
"eval_runtime": 7.5376,
"eval_samples_per_second": 133.066,
"step": 11650
},
{
"epoch": 123.16,
"learning_rate": 5.8947368421052634e-06,
"loss": 0.0001,
"step": 11700
},
{
"epoch": 123.16,
"eval_loss": 0.31679707765579224,
"eval_micro_f1": 0.8918358884346897,
"eval_runtime": 4.5395,
"eval_samples_per_second": 220.947,
"step": 11700
},
{
"epoch": 123.68,
"learning_rate": 5.877192982456141e-06,
"loss": 0.0001,
"step": 11750
},
{
"epoch": 123.68,
"eval_loss": 0.30037057399749756,
"eval_micro_f1": 0.891358024191145,
"eval_runtime": 7.746,
"eval_samples_per_second": 129.486,
"step": 11750
},
{
"epoch": 124.21,
"learning_rate": 5.859649122807018e-06,
"loss": 0.0,
"step": 11800
},
{
"epoch": 124.21,
"eval_loss": 0.3091674745082855,
"eval_micro_f1": 0.8942254298162501,
"eval_runtime": 8.4498,
"eval_samples_per_second": 118.702,
"step": 11800
},
{
"epoch": 124.74,
"learning_rate": 5.842105263157896e-06,
"loss": 0.0,
"step": 11850
},
{
"epoch": 124.74,
"eval_loss": 0.305372416973114,
"eval_micro_f1": 0.8922907180213405,
"eval_runtime": 4.0476,
"eval_samples_per_second": 247.801,
"step": 11850
},
{
"epoch": 125.26,
"learning_rate": 5.824561403508773e-06,
"loss": 0.0,
"step": 11900
},
{
"epoch": 125.26,
"eval_loss": 0.30899912118911743,
"eval_micro_f1": 0.8909717431248454,
"eval_runtime": 3.249,
"eval_samples_per_second": 308.71,
"step": 11900
},
{
"epoch": 125.79,
"learning_rate": 5.80701754385965e-06,
"loss": 0.0,
"step": 11950
},
{
"epoch": 125.79,
"eval_loss": 0.31083518266677856,
"eval_micro_f1": 0.8919552913446573,
"eval_runtime": 3.8972,
"eval_samples_per_second": 257.362,
"step": 11950
},
{
"epoch": 126.32,
"learning_rate": 5.789473684210527e-06,
"loss": 0.0,
"step": 12000
},
{
"epoch": 126.32,
"eval_loss": 0.32175785303115845,
"eval_micro_f1": 0.892328955961398,
"eval_runtime": 3.2964,
"eval_samples_per_second": 304.271,
"step": 12000
},
{
"epoch": 126.84,
"learning_rate": 5.7719298245614045e-06,
"loss": 0.0,
"step": 12050
},
{
"epoch": 126.84,
"eval_loss": 0.32289454340934753,
"eval_micro_f1": 0.8932762662400555,
"eval_runtime": 4.0516,
"eval_samples_per_second": 247.558,
"step": 12050
},
{
"epoch": 127.37,
"learning_rate": 5.754385964912281e-06,
"loss": 0.0001,
"step": 12100
},
{
"epoch": 127.37,
"eval_loss": 0.3297015428543091,
"eval_micro_f1": 0.8932092759891377,
"eval_runtime": 3.8348,
"eval_samples_per_second": 261.55,
"step": 12100
},
{
"epoch": 127.89,
"learning_rate": 5.736842105263158e-06,
"loss": 0.0001,
"step": 12150
},
{
"epoch": 127.89,
"eval_loss": 0.3137093782424927,
"eval_micro_f1": 0.8983986742650762,
"eval_runtime": 4.0423,
"eval_samples_per_second": 248.125,
"step": 12150
}
],
"max_steps": 28500,
"num_train_epochs": 300,
"total_flos": 19120804616737800,
"trial_name": null,
"trial_params": null
}
{
"best_metric": 0.8983986742650762,
"best_model_checkpoint": "../out/bio/ner-scibert-qa4-querytype5-span/checkpoints/checkpoint-12150",
"epoch": 127.89473684210526,
"global_step": 12150,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.53,
"learning_rate": 9.982456140350879e-06,
"loss": 0.497,
"step": 50
},
{
"epoch": 0.53,
"eval_loss": 0.16596195101737976,
"eval_micro_f1": 0.7603057598056351,
"eval_runtime": 3.8666,
"eval_samples_per_second": 259.4,
"step": 50
},
{
"epoch": 1.05,
"learning_rate": 9.964912280701755e-06,
"loss": 0.1446,
"step": 100
},
{
"epoch": 1.05,
"eval_loss": 0.12134452909231186,
"eval_micro_f1": 0.8320775021912038,
"eval_runtime": 2.9702,
"eval_samples_per_second": 337.691,
"step": 100
},
{
"epoch": 1.58,
"learning_rate": 9.947368421052632e-06,
"loss": 0.1055,
"step": 150
},
{
"epoch": 1.58,
"eval_loss": 0.1103413924574852,
"eval_micro_f1": 0.8564287692859414,
"eval_runtime": 3.8528,
"eval_samples_per_second": 260.329,
"step": 150
},
{
"epoch": 2.11,
"learning_rate": 9.929824561403509e-06,
"loss": 0.0844,
"step": 200
},
{
"epoch": 2.11,
"eval_loss": 0.10704845190048218,
"eval_micro_f1": 0.8708003272791329,
"eval_runtime": 3.2958,
"eval_samples_per_second": 304.328,
"step": 200
},
{
"epoch": 2.63,
"learning_rate": 9.912280701754386e-06,
"loss": 0.0663,
"step": 250
},
{
"epoch": 2.63,
"eval_loss": 0.10774940997362137,
"eval_micro_f1": 0.877275885531736,
"eval_runtime": 4.1403,
"eval_samples_per_second": 242.253,
"step": 250
},
{
"epoch": 3.16,
"learning_rate": 9.894736842105264e-06,
"loss": 0.058,
"step": 300
},
{
"epoch": 3.16,
"eval_loss": 0.10937967151403427,
"eval_micro_f1": 0.8814672866610472,
"eval_runtime": 3.0516,
"eval_samples_per_second": 328.685,
"step": 300
},
{
"epoch": 3.68,
"learning_rate": 9.87719298245614e-06,
"loss": 0.0467,
"step": 350
},
{