t5-base-TEDxJP-1body-0context / trainer_state.json
tyoyo
add train result manually
4d42c1b
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 8.71313672922252,
"global_step": 6500,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.03,
"learning_rate": 2.680965147453083e-06,
"loss": 35.9761,
"step": 20
},
{
"epoch": 0.05,
"learning_rate": 5.361930294906166e-06,
"loss": 35.0955,
"step": 40
},
{
"epoch": 0.08,
"learning_rate": 8.04289544235925e-06,
"loss": 34.9335,
"step": 60
},
{
"epoch": 0.11,
"learning_rate": 1.0723860589812333e-05,
"loss": 34.6391,
"step": 80
},
{
"epoch": 0.13,
"learning_rate": 1.3404825737265417e-05,
"loss": 34.1356,
"step": 100
},
{
"epoch": 0.16,
"learning_rate": 1.60857908847185e-05,
"loss": 33.2089,
"step": 120
},
{
"epoch": 0.19,
"learning_rate": 1.8766756032171583e-05,
"loss": 32.6544,
"step": 140
},
{
"epoch": 0.21,
"learning_rate": 2.1447721179624665e-05,
"loss": 32.3698,
"step": 160
},
{
"epoch": 0.24,
"learning_rate": 2.4128686327077747e-05,
"loss": 31.6695,
"step": 180
},
{
"epoch": 0.27,
"learning_rate": 2.6809651474530833e-05,
"loss": 32.0937,
"step": 200
},
{
"epoch": 0.29,
"learning_rate": 2.9490616621983912e-05,
"loss": 31.3979,
"step": 220
},
{
"epoch": 0.32,
"learning_rate": 3.2171581769437e-05,
"loss": 30.5171,
"step": 240
},
{
"epoch": 0.35,
"learning_rate": 3.485254691689008e-05,
"loss": 31.0177,
"step": 260
},
{
"epoch": 0.38,
"learning_rate": 3.7533512064343166e-05,
"loss": 30.2928,
"step": 280
},
{
"epoch": 0.4,
"learning_rate": 4.021447721179625e-05,
"loss": 29.3157,
"step": 300
},
{
"epoch": 0.43,
"learning_rate": 4.289544235924933e-05,
"loss": 27.5923,
"step": 320
},
{
"epoch": 0.46,
"learning_rate": 4.557640750670242e-05,
"loss": 25.9267,
"step": 340
},
{
"epoch": 0.48,
"learning_rate": 4.8257372654155495e-05,
"loss": 23.5854,
"step": 360
},
{
"epoch": 0.51,
"learning_rate": 5.0938337801608584e-05,
"loss": 20.8788,
"step": 380
},
{
"epoch": 0.54,
"learning_rate": 5.3619302949061666e-05,
"loss": 18.1499,
"step": 400
},
{
"epoch": 0.56,
"learning_rate": 5.6300268096514755e-05,
"loss": 14.4047,
"step": 420
},
{
"epoch": 0.59,
"learning_rate": 5.8981233243967824e-05,
"loss": 7.4829,
"step": 440
},
{
"epoch": 0.62,
"learning_rate": 6.16621983914209e-05,
"loss": 1.5146,
"step": 460
},
{
"epoch": 0.64,
"learning_rate": 6.4343163538874e-05,
"loss": 0.9193,
"step": 480
},
{
"epoch": 0.67,
"learning_rate": 6.702412868632708e-05,
"loss": 0.8398,
"step": 500
},
{
"epoch": 0.7,
"learning_rate": 6.970509383378017e-05,
"loss": 0.7985,
"step": 520
},
{
"epoch": 0.72,
"learning_rate": 7.238605898123325e-05,
"loss": 0.724,
"step": 540
},
{
"epoch": 0.75,
"learning_rate": 7.506702412868633e-05,
"loss": 0.6716,
"step": 560
},
{
"epoch": 0.78,
"learning_rate": 7.774798927613941e-05,
"loss": 0.6661,
"step": 580
},
{
"epoch": 0.8,
"learning_rate": 8.04289544235925e-05,
"loss": 0.6585,
"step": 600
},
{
"epoch": 0.83,
"learning_rate": 8.310991957104558e-05,
"loss": 0.6667,
"step": 620
},
{
"epoch": 0.86,
"learning_rate": 8.579088471849866e-05,
"loss": 0.6471,
"step": 640
},
{
"epoch": 0.88,
"learning_rate": 8.847184986595176e-05,
"loss": 0.6125,
"step": 660
},
{
"epoch": 0.91,
"learning_rate": 9.115281501340484e-05,
"loss": 0.6344,
"step": 680
},
{
"epoch": 0.94,
"learning_rate": 9.383378016085791e-05,
"loss": 0.5928,
"step": 700
},
{
"epoch": 0.97,
"learning_rate": 9.651474530831099e-05,
"loss": 0.6335,
"step": 720
},
{
"epoch": 0.99,
"learning_rate": 9.919571045576409e-05,
"loss": 0.6122,
"step": 740
},
{
"epoch": 1.0,
"eval_cer": 0.22800093910360325,
"eval_deletions": 4828,
"eval_hits": 54224,
"eval_insertions": 3962,
"eval_loss": 0.4796110987663269,
"eval_mer": 0.23053781751099758,
"eval_runtime": 1011.4956,
"eval_samples_per_second": 6.204,
"eval_steps_per_second": 0.776,
"eval_substitutions": 7456,
"eval_wer": 0.24427136585072473,
"eval_wil": 0.3265158023477125,
"eval_wip": 0.6734841976522875,
"step": 746
},
{
"epoch": 1.02,
"learning_rate": 9.979148048853143e-05,
"loss": 0.5954,
"step": 760
},
{
"epoch": 1.05,
"learning_rate": 9.949359547214776e-05,
"loss": 0.5923,
"step": 780
},
{
"epoch": 1.07,
"learning_rate": 9.919571045576409e-05,
"loss": 0.5518,
"step": 800
},
{
"epoch": 1.1,
"learning_rate": 9.88978254393804e-05,
"loss": 0.5674,
"step": 820
},
{
"epoch": 1.13,
"learning_rate": 9.859994042299673e-05,
"loss": 0.5602,
"step": 840
},
{
"epoch": 1.15,
"learning_rate": 9.830205540661305e-05,
"loss": 0.5601,
"step": 860
},
{
"epoch": 1.18,
"learning_rate": 9.800417039022938e-05,
"loss": 0.5503,
"step": 880
},
{
"epoch": 1.21,
"learning_rate": 9.770628537384569e-05,
"loss": 0.5754,
"step": 900
},
{
"epoch": 1.23,
"learning_rate": 9.740840035746203e-05,
"loss": 0.5782,
"step": 920
},
{
"epoch": 1.26,
"learning_rate": 9.711051534107835e-05,
"loss": 0.5389,
"step": 940
},
{
"epoch": 1.29,
"learning_rate": 9.681263032469468e-05,
"loss": 0.571,
"step": 960
},
{
"epoch": 1.31,
"learning_rate": 9.651474530831099e-05,
"loss": 0.5699,
"step": 980
},
{
"epoch": 1.34,
"learning_rate": 9.621686029192732e-05,
"loss": 0.526,
"step": 1000
},
{
"epoch": 1.37,
"learning_rate": 9.591897527554365e-05,
"loss": 0.5505,
"step": 1020
},
{
"epoch": 1.39,
"learning_rate": 9.562109025915997e-05,
"loss": 0.5709,
"step": 1040
},
{
"epoch": 1.42,
"learning_rate": 9.532320524277629e-05,
"loss": 0.5532,
"step": 1060
},
{
"epoch": 1.45,
"learning_rate": 9.502532022639262e-05,
"loss": 0.5264,
"step": 1080
},
{
"epoch": 1.47,
"learning_rate": 9.472743521000894e-05,
"loss": 0.5521,
"step": 1100
},
{
"epoch": 1.5,
"learning_rate": 9.442955019362527e-05,
"loss": 0.5387,
"step": 1120
},
{
"epoch": 1.53,
"learning_rate": 9.413166517724159e-05,
"loss": 0.5352,
"step": 1140
},
{
"epoch": 1.55,
"learning_rate": 9.383378016085791e-05,
"loss": 0.5279,
"step": 1160
},
{
"epoch": 1.58,
"learning_rate": 9.353589514447424e-05,
"loss": 0.5198,
"step": 1180
},
{
"epoch": 1.61,
"learning_rate": 9.323801012809056e-05,
"loss": 0.5806,
"step": 1200
},
{
"epoch": 1.64,
"learning_rate": 9.294012511170688e-05,
"loss": 0.5378,
"step": 1220
},
{
"epoch": 1.66,
"learning_rate": 9.26422400953232e-05,
"loss": 0.5302,
"step": 1240
},
{
"epoch": 1.69,
"learning_rate": 9.234435507893954e-05,
"loss": 0.5466,
"step": 1260
},
{
"epoch": 1.72,
"learning_rate": 9.204647006255586e-05,
"loss": 0.5166,
"step": 1280
},
{
"epoch": 1.74,
"learning_rate": 9.174858504617218e-05,
"loss": 0.521,
"step": 1300
},
{
"epoch": 1.77,
"learning_rate": 9.14507000297885e-05,
"loss": 0.5241,
"step": 1320
},
{
"epoch": 1.8,
"learning_rate": 9.115281501340484e-05,
"loss": 0.5219,
"step": 1340
},
{
"epoch": 1.82,
"learning_rate": 9.085492999702116e-05,
"loss": 0.5093,
"step": 1360
},
{
"epoch": 1.85,
"learning_rate": 9.055704498063747e-05,
"loss": 0.5485,
"step": 1380
},
{
"epoch": 1.88,
"learning_rate": 9.02591599642538e-05,
"loss": 0.5173,
"step": 1400
},
{
"epoch": 1.9,
"learning_rate": 8.996127494787012e-05,
"loss": 0.5229,
"step": 1420
},
{
"epoch": 1.93,
"learning_rate": 8.966338993148646e-05,
"loss": 0.5474,
"step": 1440
},
{
"epoch": 1.96,
"learning_rate": 8.936550491510277e-05,
"loss": 0.5388,
"step": 1460
},
{
"epoch": 1.98,
"learning_rate": 8.90676198987191e-05,
"loss": 0.5345,
"step": 1480
},
{
"epoch": 2.0,
"eval_cer": 0.1926671660312811,
"eval_deletions": 4683,
"eval_hits": 54896,
"eval_insertions": 2609,
"eval_loss": 0.43797770142555237,
"eval_mer": 0.20575256449209312,
"eval_runtime": 945.3181,
"eval_samples_per_second": 6.638,
"eval_steps_per_second": 0.83,
"eval_substitutions": 6929,
"eval_wer": 0.21382390088410416,
"eval_wil": 0.296778046061947,
"eval_wip": 0.703221953938053,
"step": 1492
},
{
"epoch": 2.01,
"learning_rate": 8.876973488233542e-05,
"loss": 0.517,
"step": 1500
},
{
"epoch": 2.04,
"learning_rate": 8.847184986595176e-05,
"loss": 0.4945,
"step": 1520
},
{
"epoch": 2.06,
"learning_rate": 8.817396484956806e-05,
"loss": 0.4924,
"step": 1540
},
{
"epoch": 2.09,
"learning_rate": 8.78760798331844e-05,
"loss": 0.4789,
"step": 1560
},
{
"epoch": 2.12,
"learning_rate": 8.757819481680072e-05,
"loss": 0.487,
"step": 1580
},
{
"epoch": 2.14,
"learning_rate": 8.728030980041705e-05,
"loss": 0.5086,
"step": 1600
},
{
"epoch": 2.17,
"learning_rate": 8.698242478403336e-05,
"loss": 0.4746,
"step": 1620
},
{
"epoch": 2.2,
"learning_rate": 8.668453976764968e-05,
"loss": 0.4631,
"step": 1640
},
{
"epoch": 2.23,
"learning_rate": 8.638665475126602e-05,
"loss": 0.509,
"step": 1660
},
{
"epoch": 2.25,
"learning_rate": 8.608876973488234e-05,
"loss": 0.4867,
"step": 1680
},
{
"epoch": 2.28,
"learning_rate": 8.579088471849866e-05,
"loss": 0.4788,
"step": 1700
},
{
"epoch": 2.31,
"learning_rate": 8.549299970211498e-05,
"loss": 0.4859,
"step": 1720
},
{
"epoch": 2.33,
"learning_rate": 8.519511468573132e-05,
"loss": 0.4856,
"step": 1740
},
{
"epoch": 2.36,
"learning_rate": 8.489722966934764e-05,
"loss": 0.4725,
"step": 1760
},
{
"epoch": 2.39,
"learning_rate": 8.459934465296396e-05,
"loss": 0.4691,
"step": 1780
},
{
"epoch": 2.41,
"learning_rate": 8.430145963658028e-05,
"loss": 0.4485,
"step": 1800
},
{
"epoch": 2.44,
"learning_rate": 8.400357462019661e-05,
"loss": 0.484,
"step": 1820
},
{
"epoch": 2.47,
"learning_rate": 8.370568960381294e-05,
"loss": 0.484,
"step": 1840
},
{
"epoch": 2.49,
"learning_rate": 8.340780458742926e-05,
"loss": 0.477,
"step": 1860
},
{
"epoch": 2.52,
"learning_rate": 8.310991957104558e-05,
"loss": 0.5029,
"step": 1880
},
{
"epoch": 2.55,
"learning_rate": 8.28120345546619e-05,
"loss": 0.4957,
"step": 1900
},
{
"epoch": 2.57,
"learning_rate": 8.251414953827823e-05,
"loss": 0.4917,
"step": 1920
},
{
"epoch": 2.6,
"learning_rate": 8.221626452189454e-05,
"loss": 0.4959,
"step": 1940
},
{
"epoch": 2.63,
"learning_rate": 8.191837950551088e-05,
"loss": 0.4756,
"step": 1960
},
{
"epoch": 2.65,
"learning_rate": 8.16204944891272e-05,
"loss": 0.4749,
"step": 1980
},
{
"epoch": 2.68,
"learning_rate": 8.132260947274353e-05,
"loss": 0.4821,
"step": 2000
},
{
"epoch": 2.71,
"learning_rate": 8.102472445635984e-05,
"loss": 0.4749,
"step": 2020
},
{
"epoch": 2.73,
"learning_rate": 8.072683943997617e-05,
"loss": 0.461,
"step": 2040
},
{
"epoch": 2.76,
"learning_rate": 8.04289544235925e-05,
"loss": 0.4985,
"step": 2060
},
{
"epoch": 2.79,
"learning_rate": 8.013106940720883e-05,
"loss": 0.4886,
"step": 2080
},
{
"epoch": 2.82,
"learning_rate": 7.983318439082514e-05,
"loss": 0.476,
"step": 2100
},
{
"epoch": 2.84,
"learning_rate": 7.953529937444147e-05,
"loss": 0.4872,
"step": 2120
},
{
"epoch": 2.87,
"learning_rate": 7.92374143580578e-05,
"loss": 0.4823,
"step": 2140
},
{
"epoch": 2.9,
"learning_rate": 7.893952934167412e-05,
"loss": 0.4901,
"step": 2160
},
{
"epoch": 2.92,
"learning_rate": 7.864164432529044e-05,
"loss": 0.4831,
"step": 2180
},
{
"epoch": 2.95,
"learning_rate": 7.834375930890676e-05,
"loss": 0.4747,
"step": 2200
},
{
"epoch": 2.98,
"learning_rate": 7.804587429252309e-05,
"loss": 0.4642,
"step": 2220
},
{
"epoch": 3.0,
"eval_cer": 0.18081657294263642,
"eval_deletions": 4438,
"eval_hits": 55151,
"eval_insertions": 2260,
"eval_loss": 0.4211605191230774,
"eval_mer": 0.19801361098185202,
"eval_runtime": 929.7123,
"eval_samples_per_second": 6.749,
"eval_steps_per_second": 0.844,
"eval_substitutions": 6919,
"eval_wer": 0.20474228664220845,
"eval_wil": 0.28908227331086855,
"eval_wip": 0.7109177266891314,
"step": 2238
},
{
"epoch": 3.0,
"learning_rate": 7.774798927613941e-05,
"loss": 0.4761,
"step": 2240
},
{
"epoch": 3.03,
"learning_rate": 7.745010425975573e-05,
"loss": 0.4689,
"step": 2260
},
{
"epoch": 3.06,
"learning_rate": 7.715221924337206e-05,
"loss": 0.4719,
"step": 2280
},
{
"epoch": 3.08,
"learning_rate": 7.685433422698839e-05,
"loss": 0.4701,
"step": 2300
},
{
"epoch": 3.11,
"learning_rate": 7.655644921060471e-05,
"loss": 0.4208,
"step": 2320
},
{
"epoch": 3.14,
"learning_rate": 7.625856419422103e-05,
"loss": 0.4471,
"step": 2340
},
{
"epoch": 3.16,
"learning_rate": 7.596067917783735e-05,
"loss": 0.4465,
"step": 2360
},
{
"epoch": 3.19,
"learning_rate": 7.566279416145369e-05,
"loss": 0.4379,
"step": 2380
},
{
"epoch": 3.22,
"learning_rate": 7.536490914507001e-05,
"loss": 0.4697,
"step": 2400
},
{
"epoch": 3.24,
"learning_rate": 7.506702412868633e-05,
"loss": 0.4335,
"step": 2420
},
{
"epoch": 3.27,
"learning_rate": 7.476913911230265e-05,
"loss": 0.454,
"step": 2440
},
{
"epoch": 3.3,
"learning_rate": 7.447125409591897e-05,
"loss": 0.4643,
"step": 2460
},
{
"epoch": 3.32,
"learning_rate": 7.417336907953531e-05,
"loss": 0.4512,
"step": 2480
},
{
"epoch": 3.35,
"learning_rate": 7.387548406315163e-05,
"loss": 0.4301,
"step": 2500
},
{
"epoch": 3.38,
"learning_rate": 7.357759904676795e-05,
"loss": 0.4546,
"step": 2520
},
{
"epoch": 3.4,
"learning_rate": 7.327971403038427e-05,
"loss": 0.4316,
"step": 2540
},
{
"epoch": 3.43,
"learning_rate": 7.29818290140006e-05,
"loss": 0.4463,
"step": 2560
},
{
"epoch": 3.46,
"learning_rate": 7.268394399761693e-05,
"loss": 0.4456,
"step": 2580
},
{
"epoch": 3.49,
"learning_rate": 7.238605898123325e-05,
"loss": 0.4487,
"step": 2600
},
{
"epoch": 3.51,
"learning_rate": 7.208817396484957e-05,
"loss": 0.4441,
"step": 2620
},
{
"epoch": 3.54,
"learning_rate": 7.17902889484659e-05,
"loss": 0.4364,
"step": 2640
},
{
"epoch": 3.57,
"learning_rate": 7.149240393208221e-05,
"loss": 0.4475,
"step": 2660
},
{
"epoch": 3.59,
"learning_rate": 7.119451891569855e-05,
"loss": 0.4462,
"step": 2680
},
{
"epoch": 3.62,
"learning_rate": 7.089663389931487e-05,
"loss": 0.4608,
"step": 2700
},
{
"epoch": 3.65,
"learning_rate": 7.059874888293119e-05,
"loss": 0.4439,
"step": 2720
},
{
"epoch": 3.67,
"learning_rate": 7.030086386654751e-05,
"loss": 0.4347,
"step": 2740
},
{
"epoch": 3.7,
"learning_rate": 7.000297885016383e-05,
"loss": 0.4608,
"step": 2760
},
{
"epoch": 3.73,
"learning_rate": 6.970509383378017e-05,
"loss": 0.4532,
"step": 2780
},
{
"epoch": 3.75,
"learning_rate": 6.940720881739649e-05,
"loss": 0.4375,
"step": 2800
},
{
"epoch": 3.78,
"learning_rate": 6.910932380101281e-05,
"loss": 0.4493,
"step": 2820
},
{
"epoch": 3.81,
"learning_rate": 6.881143878462913e-05,
"loss": 0.4443,
"step": 2840
},
{
"epoch": 3.83,
"learning_rate": 6.851355376824547e-05,
"loss": 0.4402,
"step": 2860
},
{
"epoch": 3.86,
"learning_rate": 6.821566875186179e-05,
"loss": 0.4413,
"step": 2880
},
{
"epoch": 3.89,
"learning_rate": 6.791778373547811e-05,
"loss": 0.4388,
"step": 2900
},
{
"epoch": 3.91,
"learning_rate": 6.761989871909443e-05,
"loss": 0.4439,
"step": 2920
},
{
"epoch": 3.94,
"learning_rate": 6.732201370271076e-05,
"loss": 0.448,
"step": 2940
},
{
"epoch": 3.97,
"learning_rate": 6.702412868632708e-05,
"loss": 0.4476,
"step": 2960
},
{
"epoch": 3.99,
"learning_rate": 6.67262436699434e-05,
"loss": 0.4481,
"step": 2980
},
{
"epoch": 4.0,
"eval_cer": 0.1899560633671336,
"eval_deletions": 4153,
"eval_hits": 55367,
"eval_insertions": 2944,
"eval_loss": 0.4186785817146301,
"eval_mer": 0.20280193514945574,
"eval_runtime": 923.6349,
"eval_samples_per_second": 6.794,
"eval_steps_per_second": 0.85,
"eval_substitutions": 6988,
"eval_wer": 0.21177903410116075,
"eval_wil": 0.2941351331957437,
"eval_wip": 0.7058648668042563,
"step": 2984
},
{
"epoch": 4.02,
"learning_rate": 6.642835865355973e-05,
"loss": 0.4268,
"step": 3000
},
{
"epoch": 4.05,
"learning_rate": 6.613047363717605e-05,
"loss": 0.4343,
"step": 3020
},
{
"epoch": 4.08,
"learning_rate": 6.583258862079238e-05,
"loss": 0.4045,
"step": 3040
},
{
"epoch": 4.1,
"learning_rate": 6.55347036044087e-05,
"loss": 0.4079,
"step": 3060
},
{
"epoch": 4.13,
"learning_rate": 6.523681858802503e-05,
"loss": 0.4329,
"step": 3080
},
{
"epoch": 4.16,
"learning_rate": 6.493893357164135e-05,
"loss": 0.425,
"step": 3100
},
{
"epoch": 4.18,
"learning_rate": 6.464104855525768e-05,
"loss": 0.4117,
"step": 3120
},
{
"epoch": 4.21,
"learning_rate": 6.4343163538874e-05,
"loss": 0.429,
"step": 3140
},
{
"epoch": 4.24,
"learning_rate": 6.404527852249032e-05,
"loss": 0.4186,
"step": 3160
},
{
"epoch": 4.26,
"learning_rate": 6.374739350610664e-05,
"loss": 0.4209,
"step": 3180
},
{
"epoch": 4.29,
"learning_rate": 6.344950848972298e-05,
"loss": 0.4236,
"step": 3200
},
{
"epoch": 4.32,
"learning_rate": 6.31516234733393e-05,
"loss": 0.4057,
"step": 3220
},
{
"epoch": 4.34,
"learning_rate": 6.285373845695561e-05,
"loss": 0.4042,
"step": 3240
},
{
"epoch": 4.37,
"learning_rate": 6.255585344057194e-05,
"loss": 0.4413,
"step": 3260
},
{
"epoch": 4.4,
"learning_rate": 6.225796842418826e-05,
"loss": 0.4316,
"step": 3280
},
{
"epoch": 4.42,
"learning_rate": 6.19600834078046e-05,
"loss": 0.4273,
"step": 3300
},
{
"epoch": 4.45,
"learning_rate": 6.16621983914209e-05,
"loss": 0.4361,
"step": 3320
},
{
"epoch": 4.48,
"learning_rate": 6.136431337503724e-05,
"loss": 0.4485,
"step": 3340
},
{
"epoch": 4.5,
"learning_rate": 6.106642835865356e-05,
"loss": 0.4034,
"step": 3360
},
{
"epoch": 4.53,
"learning_rate": 6.076854334226989e-05,
"loss": 0.4119,
"step": 3380
},
{
"epoch": 4.56,
"learning_rate": 6.047065832588621e-05,
"loss": 0.3945,
"step": 3400
},
{
"epoch": 4.58,
"learning_rate": 6.017277330950254e-05,
"loss": 0.4258,
"step": 3420
},
{
"epoch": 4.61,
"learning_rate": 5.987488829311886e-05,
"loss": 0.4311,
"step": 3440
},
{
"epoch": 4.64,
"learning_rate": 5.957700327673519e-05,
"loss": 0.4134,
"step": 3460
},
{
"epoch": 4.66,
"learning_rate": 5.927911826035151e-05,
"loss": 0.4262,
"step": 3480
},
{
"epoch": 4.69,
"learning_rate": 5.8981233243967824e-05,
"loss": 0.4208,
"step": 3500
},
{
"epoch": 4.72,
"learning_rate": 5.868334822758416e-05,
"loss": 0.4111,
"step": 3520
},
{
"epoch": 4.75,
"learning_rate": 5.838546321120047e-05,
"loss": 0.4273,
"step": 3540
},
{
"epoch": 4.77,
"learning_rate": 5.808757819481681e-05,
"loss": 0.3966,
"step": 3560
},
{
"epoch": 4.8,
"learning_rate": 5.778969317843312e-05,
"loss": 0.4442,
"step": 3580
},
{
"epoch": 4.83,
"learning_rate": 5.749180816204946e-05,
"loss": 0.4085,
"step": 3600
},
{
"epoch": 4.85,
"learning_rate": 5.719392314566577e-05,
"loss": 0.4319,
"step": 3620
},
{
"epoch": 4.88,
"learning_rate": 5.6896038129282106e-05,
"loss": 0.4201,
"step": 3640
},
{
"epoch": 4.91,
"learning_rate": 5.659815311289842e-05,
"loss": 0.4209,
"step": 3660
},
{
"epoch": 4.93,
"learning_rate": 5.6300268096514755e-05,
"loss": 0.4329,
"step": 3680
},
{
"epoch": 4.96,
"learning_rate": 5.600238308013107e-05,
"loss": 0.4164,
"step": 3700
},
{
"epoch": 4.99,
"learning_rate": 5.5704498063747404e-05,
"loss": 0.4224,
"step": 3720
},
{
"epoch": 5.0,
"eval_cer": 0.180956320502644,
"eval_deletions": 4305,
"eval_hits": 55337,
"eval_insertions": 2397,
"eval_loss": 0.4172378480434418,
"eval_mer": 0.19690878746099702,
"eval_runtime": 915.2753,
"eval_samples_per_second": 6.856,
"eval_steps_per_second": 0.858,
"eval_substitutions": 6866,
"eval_wer": 0.20400553316894207,
"eval_wil": 0.28727036701839537,
"eval_wip": 0.7127296329816046,
"step": 3730
},
{
"epoch": 5.01,
"learning_rate": 5.540661304736372e-05,
"loss": 0.3878,
"step": 3740
},
{
"epoch": 5.04,
"learning_rate": 5.510872803098004e-05,
"loss": 0.4118,
"step": 3760
},
{
"epoch": 5.07,
"learning_rate": 5.481084301459637e-05,
"loss": 0.3936,
"step": 3780
},
{
"epoch": 5.09,
"learning_rate": 5.451295799821269e-05,
"loss": 0.4022,
"step": 3800
},
{
"epoch": 5.12,
"learning_rate": 5.421507298182902e-05,
"loss": 0.3973,
"step": 3820
},
{
"epoch": 5.15,
"learning_rate": 5.391718796544534e-05,
"loss": 0.4062,
"step": 3840
},
{
"epoch": 5.17,
"learning_rate": 5.3619302949061666e-05,
"loss": 0.4071,
"step": 3860
},
{
"epoch": 5.2,
"learning_rate": 5.332141793267799e-05,
"loss": 0.4052,
"step": 3880
},
{
"epoch": 5.23,
"learning_rate": 5.3023532916294315e-05,
"loss": 0.3924,
"step": 3900
},
{
"epoch": 5.25,
"learning_rate": 5.2725647899910637e-05,
"loss": 0.3916,
"step": 3920
},
{
"epoch": 5.28,
"learning_rate": 5.2427762883526964e-05,
"loss": 0.4163,
"step": 3940
},
{
"epoch": 5.31,
"learning_rate": 5.2129877867143286e-05,
"loss": 0.4034,
"step": 3960
},
{
"epoch": 5.34,
"learning_rate": 5.1831992850759614e-05,
"loss": 0.4115,
"step": 3980
},
{
"epoch": 5.36,
"learning_rate": 5.1534107834375935e-05,
"loss": 0.4087,
"step": 4000
},
{
"epoch": 5.39,
"learning_rate": 5.123622281799225e-05,
"loss": 0.3804,
"step": 4020
},
{
"epoch": 5.42,
"learning_rate": 5.0938337801608584e-05,
"loss": 0.4055,
"step": 4040
},
{
"epoch": 5.44,
"learning_rate": 5.06404527852249e-05,
"loss": 0.3821,
"step": 4060
},
{
"epoch": 5.47,
"learning_rate": 5.034256776884123e-05,
"loss": 0.3997,
"step": 4080
},
{
"epoch": 5.5,
"learning_rate": 5.004468275245755e-05,
"loss": 0.3917,
"step": 4100
},
{
"epoch": 5.52,
"learning_rate": 4.974679773607388e-05,
"loss": 0.4167,
"step": 4120
},
{
"epoch": 5.55,
"learning_rate": 4.94489127196902e-05,
"loss": 0.376,
"step": 4140
},
{
"epoch": 5.58,
"learning_rate": 4.9151027703306525e-05,
"loss": 0.4055,
"step": 4160
},
{
"epoch": 5.6,
"learning_rate": 4.8853142686922846e-05,
"loss": 0.4081,
"step": 4180
},
{
"epoch": 5.63,
"learning_rate": 4.8555257670539174e-05,
"loss": 0.4118,
"step": 4200
},
{
"epoch": 5.66,
"learning_rate": 4.8257372654155495e-05,
"loss": 0.4007,
"step": 4220
},
{
"epoch": 5.68,
"learning_rate": 4.795948763777182e-05,
"loss": 0.4243,
"step": 4240
},
{
"epoch": 5.71,
"learning_rate": 4.7661602621388144e-05,
"loss": 0.4045,
"step": 4260
},
{
"epoch": 5.74,
"learning_rate": 4.736371760500447e-05,
"loss": 0.3798,
"step": 4280
},
{
"epoch": 5.76,
"learning_rate": 4.706583258862079e-05,
"loss": 0.3839,
"step": 4300
},
{
"epoch": 5.79,
"learning_rate": 4.676794757223712e-05,
"loss": 0.3866,
"step": 4320
},
{
"epoch": 5.82,
"learning_rate": 4.647006255585344e-05,
"loss": 0.3841,
"step": 4340
},
{
"epoch": 5.84,
"learning_rate": 4.617217753946977e-05,
"loss": 0.4069,
"step": 4360
},
{
"epoch": 5.87,
"learning_rate": 4.587429252308609e-05,
"loss": 0.4176,
"step": 4380
},
{
"epoch": 5.9,
"learning_rate": 4.557640750670242e-05,
"loss": 0.4169,
"step": 4400
},
{
"epoch": 5.92,
"learning_rate": 4.5278522490318734e-05,
"loss": 0.391,
"step": 4420
},
{
"epoch": 5.95,
"learning_rate": 4.498063747393506e-05,
"loss": 0.3957,
"step": 4440
},
{
"epoch": 5.98,
"learning_rate": 4.468275245755138e-05,
"loss": 0.3927,
"step": 4460
},
{
"epoch": 6.0,
"eval_cer": 0.18605990139412165,
"eval_deletions": 3953,
"eval_hits": 55575,
"eval_insertions": 2665,
"eval_loss": 0.4182780683040619,
"eval_mer": 0.1965795903025747,
"eval_runtime": 936.5757,
"eval_samples_per_second": 6.7,
"eval_steps_per_second": 0.838,
"eval_substitutions": 6980,
"eval_wer": 0.20445660672400312,
"eval_wil": 0.28796021536777305,
"eval_wip": 0.712039784632227,
"step": 4476
},
{
"epoch": 6.01,
"learning_rate": 4.438486744116771e-05,
"loss": 0.3821,
"step": 4480
},
{
"epoch": 6.03,
"learning_rate": 4.408698242478403e-05,
"loss": 0.3741,
"step": 4500
},
{
"epoch": 6.06,
"learning_rate": 4.378909740840036e-05,
"loss": 0.3642,
"step": 4520
},
{
"epoch": 6.09,
"learning_rate": 4.349121239201668e-05,
"loss": 0.3825,
"step": 4540
},
{
"epoch": 6.11,
"learning_rate": 4.319332737563301e-05,
"loss": 0.3787,
"step": 4560
},
{
"epoch": 6.14,
"learning_rate": 4.289544235924933e-05,
"loss": 0.3808,
"step": 4580
},
{
"epoch": 6.17,
"learning_rate": 4.259755734286566e-05,
"loss": 0.3632,
"step": 4600
},
{
"epoch": 6.19,
"learning_rate": 4.229967232648198e-05,
"loss": 0.3658,
"step": 4620
},
{
"epoch": 6.22,
"learning_rate": 4.200178731009831e-05,
"loss": 0.3821,
"step": 4640
},
{
"epoch": 6.25,
"learning_rate": 4.170390229371463e-05,
"loss": 0.3982,
"step": 4660
},
{
"epoch": 6.27,
"learning_rate": 4.140601727733095e-05,
"loss": 0.3858,
"step": 4680
},
{
"epoch": 6.3,
"learning_rate": 4.110813226094727e-05,
"loss": 0.3853,
"step": 4700
},
{
"epoch": 6.33,
"learning_rate": 4.08102472445636e-05,
"loss": 0.3684,
"step": 4720
},
{
"epoch": 6.35,
"learning_rate": 4.051236222817992e-05,
"loss": 0.3623,
"step": 4740
},
{
"epoch": 6.38,
"learning_rate": 4.021447721179625e-05,
"loss": 0.38,
"step": 4760
},
{
"epoch": 6.41,
"learning_rate": 3.991659219541257e-05,
"loss": 0.4107,
"step": 4780
},
{
"epoch": 6.43,
"learning_rate": 3.96187071790289e-05,
"loss": 0.3882,
"step": 4800
},
{
"epoch": 6.46,
"learning_rate": 3.932082216264522e-05,
"loss": 0.3907,
"step": 4820
},
{
"epoch": 6.49,
"learning_rate": 3.9022937146261546e-05,
"loss": 0.3749,
"step": 4840
},
{
"epoch": 6.51,
"learning_rate": 3.872505212987787e-05,
"loss": 0.3849,
"step": 4860
},
{
"epoch": 6.54,
"learning_rate": 3.8427167113494195e-05,
"loss": 0.3726,
"step": 4880
},
{
"epoch": 6.57,
"learning_rate": 3.8129282097110517e-05,
"loss": 0.3939,
"step": 4900
},
{
"epoch": 6.6,
"learning_rate": 3.7831397080726844e-05,
"loss": 0.3641,
"step": 4920
},
{
"epoch": 6.62,
"learning_rate": 3.7533512064343166e-05,
"loss": 0.3786,
"step": 4940
},
{
"epoch": 6.65,
"learning_rate": 3.723562704795949e-05,
"loss": 0.3941,
"step": 4960
},
{
"epoch": 6.68,
"learning_rate": 3.6937742031575815e-05,
"loss": 0.3876,
"step": 4980
},
{
"epoch": 6.7,
"learning_rate": 3.6639857015192136e-05,
"loss": 0.3878,
"step": 5000
},
{
"epoch": 6.73,
"learning_rate": 3.6341971998808464e-05,
"loss": 0.3897,
"step": 5020
},
{
"epoch": 6.76,
"learning_rate": 3.6044086982424785e-05,
"loss": 0.3974,
"step": 5040
},
{
"epoch": 6.78,
"learning_rate": 3.5746201966041106e-05,
"loss": 0.3781,
"step": 5060
},
{
"epoch": 6.81,
"learning_rate": 3.5448316949657434e-05,
"loss": 0.39,
"step": 5080
},
{
"epoch": 6.84,
"learning_rate": 3.5150431933273755e-05,
"loss": 0.4095,
"step": 5100
},
{
"epoch": 6.86,
"learning_rate": 3.485254691689008e-05,
"loss": 0.3831,
"step": 5120
},
{
"epoch": 6.89,
"learning_rate": 3.4554661900506405e-05,
"loss": 0.4011,
"step": 5140
},
{
"epoch": 6.92,
"learning_rate": 3.425677688412273e-05,
"loss": 0.3948,
"step": 5160
},
{
"epoch": 6.94,
"learning_rate": 3.3958891867739054e-05,
"loss": 0.3937,
"step": 5180
},
{
"epoch": 6.97,
"learning_rate": 3.366100685135538e-05,
"loss": 0.3895,
"step": 5200
},
{
"epoch": 7.0,
"learning_rate": 3.33631218349717e-05,
"loss": 0.3834,
"step": 5220
},
{
"epoch": 7.0,
"eval_cer": 0.18220286873791183,
"eval_deletions": 4025,
"eval_hits": 55571,
"eval_insertions": 2507,
"eval_loss": 0.41830888390541077,
"eval_mer": 0.19479823226834747,
"eval_runtime": 915.8341,
"eval_samples_per_second": 6.852,
"eval_steps_per_second": 0.857,
"eval_substitutions": 6912,
"eval_wer": 0.20214109580802309,
"eval_wil": 0.28554315919063866,
"eval_wip": 0.7144568408093613,
"step": 5222
},
{
"epoch": 7.02,
"learning_rate": 3.3065236818588024e-05,
"loss": 0.3722,
"step": 5240
},
{
"epoch": 7.05,
"learning_rate": 3.276735180220435e-05,
"loss": 0.3653,
"step": 5260
},
{
"epoch": 7.08,
"learning_rate": 3.246946678582067e-05,
"loss": 0.3678,
"step": 5280
},
{
"epoch": 7.1,
"learning_rate": 3.2171581769437e-05,
"loss": 0.3713,
"step": 5300
},
{
"epoch": 7.13,
"learning_rate": 3.187369675305332e-05,
"loss": 0.3701,
"step": 5320
},
{
"epoch": 7.16,
"learning_rate": 3.157581173666965e-05,
"loss": 0.3784,
"step": 5340
},
{
"epoch": 7.18,
"learning_rate": 3.127792672028597e-05,
"loss": 0.3671,
"step": 5360
},
{
"epoch": 7.21,
"learning_rate": 3.09800417039023e-05,
"loss": 0.3699,
"step": 5380
},
{
"epoch": 7.24,
"learning_rate": 3.068215668751862e-05,
"loss": 0.3834,
"step": 5400
},
{
"epoch": 7.27,
"learning_rate": 3.0384271671134945e-05,
"loss": 0.3745,
"step": 5420
},
{
"epoch": 7.29,
"learning_rate": 3.008638665475127e-05,
"loss": 0.3831,
"step": 5440
},
{
"epoch": 7.32,
"learning_rate": 2.9788501638367594e-05,
"loss": 0.3732,
"step": 5460
},
{
"epoch": 7.35,
"learning_rate": 2.9490616621983912e-05,
"loss": 0.3685,
"step": 5480
},
{
"epoch": 7.37,
"learning_rate": 2.9192731605600237e-05,
"loss": 0.3707,
"step": 5500
},
{
"epoch": 7.4,
"learning_rate": 2.889484658921656e-05,
"loss": 0.3702,
"step": 5520
},
{
"epoch": 7.43,
"learning_rate": 2.8596961572832886e-05,
"loss": 0.3707,
"step": 5540
},
{
"epoch": 7.45,
"learning_rate": 2.829907655644921e-05,
"loss": 0.3767,
"step": 5560
},
{
"epoch": 7.48,
"learning_rate": 2.8001191540065535e-05,
"loss": 0.345,
"step": 5580
},
{
"epoch": 7.51,
"learning_rate": 2.770330652368186e-05,
"loss": 0.3677,
"step": 5600
},
{
"epoch": 7.53,
"learning_rate": 2.7405421507298184e-05,
"loss": 0.371,
"step": 5620
},
{
"epoch": 7.56,
"learning_rate": 2.710753649091451e-05,
"loss": 0.3642,
"step": 5640
},
{
"epoch": 7.59,
"learning_rate": 2.6809651474530833e-05,
"loss": 0.3785,
"step": 5660
},
{
"epoch": 7.61,
"learning_rate": 2.6511766458147158e-05,
"loss": 0.3696,
"step": 5680
},
{
"epoch": 7.64,
"learning_rate": 2.6213881441763482e-05,
"loss": 0.3801,
"step": 5700
},
{
"epoch": 7.67,
"learning_rate": 2.5915996425379807e-05,
"loss": 0.3793,
"step": 5720
},
{
"epoch": 7.69,
"learning_rate": 2.5618111408996125e-05,
"loss": 0.3621,
"step": 5740
},
{
"epoch": 7.72,
"learning_rate": 2.532022639261245e-05,
"loss": 0.3622,
"step": 5760
},
{
"epoch": 7.75,
"learning_rate": 2.5022341376228774e-05,
"loss": 0.3758,
"step": 5780
},
{
"epoch": 7.77,
"learning_rate": 2.47244563598451e-05,
"loss": 0.3677,
"step": 5800
},
{
"epoch": 7.8,
"learning_rate": 2.4426571343461423e-05,
"loss": 0.3709,
"step": 5820
},
{
"epoch": 7.83,
"learning_rate": 2.4128686327077747e-05,
"loss": 0.3702,
"step": 5840
},
{
"epoch": 7.86,
"learning_rate": 2.3830801310694072e-05,
"loss": 0.3565,
"step": 5860
},
{
"epoch": 7.88,
"learning_rate": 2.3532916294310397e-05,
"loss": 0.3621,
"step": 5880
},
{
"epoch": 7.91,
"learning_rate": 2.323503127792672e-05,
"loss": 0.3875,
"step": 5900
},
{
"epoch": 7.94,
"learning_rate": 2.2937146261543046e-05,
"loss": 0.3771,
"step": 5920
},
{
"epoch": 7.96,
"learning_rate": 2.2639261245159367e-05,
"loss": 0.4039,
"step": 5940
},
{
"epoch": 7.99,
"learning_rate": 2.234137622877569e-05,
"loss": 0.3484,
"step": 5960
},
{
"epoch": 8.0,
"eval_cer": 0.1825718022963319,
"eval_deletions": 4028,
"eval_hits": 55559,
"eval_insertions": 2490,
"eval_loss": 0.41909465193748474,
"eval_mer": 0.19477376155830603,
"eval_runtime": 894.6454,
"eval_samples_per_second": 7.014,
"eval_steps_per_second": 0.877,
"eval_substitutions": 6921,
"eval_wer": 0.2020659168821796,
"eval_wil": 0.2856318460438537,
"eval_wip": 0.7143681539561463,
"step": 5968
},
{
"epoch": 8.02,
"learning_rate": 2.2043491212392016e-05,
"loss": 0.3722,
"step": 5980
},
{
"epoch": 8.04,
"learning_rate": 2.174560619600834e-05,
"loss": 0.3512,
"step": 6000
},
{
"epoch": 8.07,
"learning_rate": 2.1447721179624665e-05,
"loss": 0.377,
"step": 6020
},
{
"epoch": 8.1,
"learning_rate": 2.114983616324099e-05,
"loss": 0.36,
"step": 6040
},
{
"epoch": 8.12,
"learning_rate": 2.0851951146857314e-05,
"loss": 0.383,
"step": 6060
},
{
"epoch": 8.15,
"learning_rate": 2.0554066130473635e-05,
"loss": 0.3542,
"step": 6080
},
{
"epoch": 8.18,
"learning_rate": 2.025618111408996e-05,
"loss": 0.3594,
"step": 6100
},
{
"epoch": 8.2,
"learning_rate": 1.9958296097706285e-05,
"loss": 0.3436,
"step": 6120
},
{
"epoch": 8.23,
"learning_rate": 1.966041108132261e-05,
"loss": 0.3532,
"step": 6140
},
{
"epoch": 8.26,
"learning_rate": 1.9362526064938934e-05,
"loss": 0.3565,
"step": 6160
},
{
"epoch": 8.28,
"learning_rate": 1.9064641048555258e-05,
"loss": 0.3601,
"step": 6180
},
{
"epoch": 8.31,
"learning_rate": 1.8766756032171583e-05,
"loss": 0.3498,
"step": 6200
},
{
"epoch": 8.34,
"learning_rate": 1.8468871015787907e-05,
"loss": 0.3622,
"step": 6220
},
{
"epoch": 8.36,
"learning_rate": 1.8170985999404232e-05,
"loss": 0.3594,
"step": 6240
},
{
"epoch": 8.39,
"learning_rate": 1.7873100983020553e-05,
"loss": 0.3771,
"step": 6260
},
{
"epoch": 8.42,
"learning_rate": 1.7575215966636878e-05,
"loss": 0.3673,
"step": 6280
},
{
"epoch": 8.45,
"learning_rate": 1.7277330950253202e-05,
"loss": 0.36,
"step": 6300
},
{
"epoch": 8.47,
"learning_rate": 1.6979445933869527e-05,
"loss": 0.3746,
"step": 6320
},
{
"epoch": 8.5,
"learning_rate": 1.668156091748585e-05,
"loss": 0.3643,
"step": 6340
},
{
"epoch": 8.53,
"learning_rate": 1.6383675901102176e-05,
"loss": 0.3572,
"step": 6360
},
{
"epoch": 8.55,
"learning_rate": 1.60857908847185e-05,
"loss": 0.3704,
"step": 6380
},
{
"epoch": 8.58,
"learning_rate": 1.5787905868334825e-05,
"loss": 0.3706,
"step": 6400
},
{
"epoch": 8.61,
"learning_rate": 1.549002085195115e-05,
"loss": 0.367,
"step": 6420
},
{
"epoch": 8.63,
"learning_rate": 1.5192135835567473e-05,
"loss": 0.3408,
"step": 6440
},
{
"epoch": 8.66,
"learning_rate": 1.4894250819183797e-05,
"loss": 0.36,
"step": 6460
},
{
"epoch": 8.69,
"learning_rate": 1.4596365802800118e-05,
"loss": 0.3711,
"step": 6480
},
{
"epoch": 8.71,
"learning_rate": 1.4298480786416443e-05,
"loss": 0.3633,
"step": 6500
}
],
"max_steps": 7460,
"num_train_epochs": 10,
"total_flos": 3.15961607503872e+16,
"trial_name": null,
"trial_params": null
}