whisper-small-cantonese / trainer_state.json
alvanli
Added rest of the files for 10.1 CER
97aa366
raw
history blame
69.4 kB
{
"best_metric": 10.058774675781,
"best_model_checkpoint": "./model_out_trpro_more_more/checkpoint-13000",
"epoch": 3.6020224407812718,
"global_step": 13000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 2.5e-06,
"loss": 0.487,
"step": 25
},
{
"epoch": 0.01,
"learning_rate": 5e-06,
"loss": 0.4472,
"step": 50
},
{
"epoch": 0.02,
"learning_rate": 7.5e-06,
"loss": 0.4173,
"step": 75
},
{
"epoch": 0.03,
"learning_rate": 1e-05,
"loss": 0.4623,
"step": 100
},
{
"epoch": 0.03,
"learning_rate": 1.25e-05,
"loss": 0.457,
"step": 125
},
{
"epoch": 0.04,
"learning_rate": 1.5e-05,
"loss": 0.4775,
"step": 150
},
{
"epoch": 0.05,
"learning_rate": 1.75e-05,
"loss": 0.4445,
"step": 175
},
{
"epoch": 0.06,
"learning_rate": 2e-05,
"loss": 0.4751,
"step": 200
},
{
"epoch": 0.06,
"learning_rate": 2.25e-05,
"loss": 0.4633,
"step": 225
},
{
"epoch": 0.07,
"learning_rate": 2.5e-05,
"loss": 0.4682,
"step": 250
},
{
"epoch": 0.08,
"learning_rate": 2.7500000000000004e-05,
"loss": 0.4807,
"step": 275
},
{
"epoch": 0.08,
"learning_rate": 3e-05,
"loss": 0.4882,
"step": 300
},
{
"epoch": 0.09,
"learning_rate": 3.2500000000000004e-05,
"loss": 0.4681,
"step": 325
},
{
"epoch": 0.1,
"learning_rate": 3.5e-05,
"loss": 0.4918,
"step": 350
},
{
"epoch": 0.1,
"learning_rate": 3.7500000000000003e-05,
"loss": 0.4871,
"step": 375
},
{
"epoch": 0.11,
"learning_rate": 4e-05,
"loss": 0.4986,
"step": 400
},
{
"epoch": 0.12,
"learning_rate": 4.25e-05,
"loss": 0.4859,
"step": 425
},
{
"epoch": 0.12,
"learning_rate": 4.5e-05,
"loss": 0.4761,
"step": 450
},
{
"epoch": 0.13,
"learning_rate": 4.75e-05,
"loss": 0.4884,
"step": 475
},
{
"epoch": 0.14,
"learning_rate": 5e-05,
"loss": 0.5453,
"step": 500
},
{
"epoch": 0.14,
"eval_cer": 14.794288634766497,
"eval_loss": 0.3284708261489868,
"eval_runtime": 1225.6472,
"eval_samples_per_second": 4.562,
"eval_steps_per_second": 0.286,
"step": 500
},
{
"epoch": 0.15,
"learning_rate": 4.991379310344828e-05,
"loss": 0.5147,
"step": 525
},
{
"epoch": 0.15,
"learning_rate": 4.982758620689655e-05,
"loss": 0.514,
"step": 550
},
{
"epoch": 0.16,
"learning_rate": 4.9741379310344836e-05,
"loss": 0.5397,
"step": 575
},
{
"epoch": 0.17,
"learning_rate": 4.9655172413793107e-05,
"loss": 0.5269,
"step": 600
},
{
"epoch": 0.17,
"learning_rate": 4.9568965517241384e-05,
"loss": 0.4992,
"step": 625
},
{
"epoch": 0.18,
"learning_rate": 4.9482758620689655e-05,
"loss": 0.538,
"step": 650
},
{
"epoch": 0.19,
"learning_rate": 4.939655172413793e-05,
"loss": 0.5123,
"step": 675
},
{
"epoch": 0.19,
"learning_rate": 4.931034482758621e-05,
"loss": 0.5051,
"step": 700
},
{
"epoch": 0.2,
"learning_rate": 4.922413793103449e-05,
"loss": 0.5342,
"step": 725
},
{
"epoch": 0.21,
"learning_rate": 4.913793103448276e-05,
"loss": 0.5037,
"step": 750
},
{
"epoch": 0.21,
"learning_rate": 4.905172413793104e-05,
"loss": 0.5492,
"step": 775
},
{
"epoch": 0.22,
"learning_rate": 4.896551724137931e-05,
"loss": 0.5256,
"step": 800
},
{
"epoch": 0.23,
"learning_rate": 4.8879310344827586e-05,
"loss": 0.531,
"step": 825
},
{
"epoch": 0.24,
"learning_rate": 4.8793103448275864e-05,
"loss": 0.5245,
"step": 850
},
{
"epoch": 0.24,
"learning_rate": 4.870689655172414e-05,
"loss": 0.4702,
"step": 875
},
{
"epoch": 0.25,
"learning_rate": 4.862068965517241e-05,
"loss": 0.5404,
"step": 900
},
{
"epoch": 0.26,
"learning_rate": 4.853793103448276e-05,
"loss": 0.4968,
"step": 925
},
{
"epoch": 0.26,
"learning_rate": 4.8451724137931036e-05,
"loss": 0.4905,
"step": 950
},
{
"epoch": 0.27,
"learning_rate": 4.836551724137931e-05,
"loss": 0.5281,
"step": 975
},
{
"epoch": 0.28,
"learning_rate": 4.827931034482759e-05,
"loss": 0.5144,
"step": 1000
},
{
"epoch": 0.28,
"eval_cer": 14.356672842266658,
"eval_loss": 0.3405072093009949,
"eval_runtime": 1158.9803,
"eval_samples_per_second": 4.824,
"eval_steps_per_second": 0.302,
"step": 1000
},
{
"epoch": 0.28,
"learning_rate": 4.819310344827587e-05,
"loss": 0.5357,
"step": 1025
},
{
"epoch": 0.29,
"learning_rate": 4.810689655172414e-05,
"loss": 0.5521,
"step": 1050
},
{
"epoch": 0.3,
"learning_rate": 4.802068965517242e-05,
"loss": 0.5413,
"step": 1075
},
{
"epoch": 0.3,
"learning_rate": 4.793448275862069e-05,
"loss": 0.4707,
"step": 1100
},
{
"epoch": 0.31,
"learning_rate": 4.784827586206897e-05,
"loss": 0.4771,
"step": 1125
},
{
"epoch": 0.32,
"learning_rate": 4.7762068965517245e-05,
"loss": 0.5055,
"step": 1150
},
{
"epoch": 0.33,
"learning_rate": 4.767586206896552e-05,
"loss": 0.5464,
"step": 1175
},
{
"epoch": 0.33,
"learning_rate": 4.758965517241379e-05,
"loss": 0.4839,
"step": 1200
},
{
"epoch": 0.34,
"learning_rate": 4.750344827586207e-05,
"loss": 0.5087,
"step": 1225
},
{
"epoch": 0.35,
"learning_rate": 4.741724137931035e-05,
"loss": 0.4728,
"step": 1250
},
{
"epoch": 0.35,
"learning_rate": 4.7331034482758626e-05,
"loss": 0.4457,
"step": 1275
},
{
"epoch": 0.36,
"learning_rate": 4.72448275862069e-05,
"loss": 0.5246,
"step": 1300
},
{
"epoch": 0.37,
"learning_rate": 4.7158620689655175e-05,
"loss": 0.4953,
"step": 1325
},
{
"epoch": 0.37,
"learning_rate": 4.7072413793103446e-05,
"loss": 0.5318,
"step": 1350
},
{
"epoch": 0.38,
"learning_rate": 4.6986206896551724e-05,
"loss": 0.5062,
"step": 1375
},
{
"epoch": 0.39,
"learning_rate": 4.69e-05,
"loss": 0.4979,
"step": 1400
},
{
"epoch": 0.39,
"learning_rate": 4.681379310344828e-05,
"loss": 0.5202,
"step": 1425
},
{
"epoch": 0.4,
"learning_rate": 4.672758620689656e-05,
"loss": 0.5267,
"step": 1450
},
{
"epoch": 0.41,
"learning_rate": 4.664137931034483e-05,
"loss": 0.4971,
"step": 1475
},
{
"epoch": 0.42,
"learning_rate": 4.6555172413793106e-05,
"loss": 0.49,
"step": 1500
},
{
"epoch": 0.42,
"eval_cer": 14.058008049575161,
"eval_loss": 0.3323359787464142,
"eval_runtime": 1178.6215,
"eval_samples_per_second": 4.744,
"eval_steps_per_second": 0.297,
"step": 1500
},
{
"epoch": 0.42,
"learning_rate": 4.6468965517241384e-05,
"loss": 0.4821,
"step": 1525
},
{
"epoch": 0.43,
"learning_rate": 4.638275862068966e-05,
"loss": 0.4603,
"step": 1550
},
{
"epoch": 0.44,
"learning_rate": 4.629655172413793e-05,
"loss": 0.5213,
"step": 1575
},
{
"epoch": 0.44,
"learning_rate": 4.621034482758621e-05,
"loss": 0.4922,
"step": 1600
},
{
"epoch": 0.45,
"learning_rate": 4.612413793103448e-05,
"loss": 0.5202,
"step": 1625
},
{
"epoch": 0.46,
"learning_rate": 4.603793103448276e-05,
"loss": 0.5046,
"step": 1650
},
{
"epoch": 0.46,
"learning_rate": 4.5951724137931036e-05,
"loss": 0.4795,
"step": 1675
},
{
"epoch": 0.47,
"learning_rate": 4.5865517241379314e-05,
"loss": 0.447,
"step": 1700
},
{
"epoch": 0.48,
"learning_rate": 4.5779310344827585e-05,
"loss": 0.4752,
"step": 1725
},
{
"epoch": 0.48,
"learning_rate": 4.569310344827586e-05,
"loss": 0.515,
"step": 1750
},
{
"epoch": 0.49,
"learning_rate": 4.560689655172414e-05,
"loss": 0.4882,
"step": 1775
},
{
"epoch": 0.5,
"learning_rate": 4.552068965517242e-05,
"loss": 0.4721,
"step": 1800
},
{
"epoch": 0.51,
"learning_rate": 4.5434482758620696e-05,
"loss": 0.4542,
"step": 1825
},
{
"epoch": 0.51,
"learning_rate": 4.534827586206897e-05,
"loss": 0.5197,
"step": 1850
},
{
"epoch": 0.52,
"learning_rate": 4.5262068965517245e-05,
"loss": 0.4589,
"step": 1875
},
{
"epoch": 0.53,
"learning_rate": 4.5175862068965516e-05,
"loss": 0.4668,
"step": 1900
},
{
"epoch": 0.53,
"learning_rate": 4.50896551724138e-05,
"loss": 0.4824,
"step": 1925
},
{
"epoch": 0.54,
"learning_rate": 4.500344827586207e-05,
"loss": 0.4967,
"step": 1950
},
{
"epoch": 0.55,
"learning_rate": 4.491724137931035e-05,
"loss": 0.4913,
"step": 1975
},
{
"epoch": 0.55,
"learning_rate": 4.483103448275862e-05,
"loss": 0.461,
"step": 2000
},
{
"epoch": 0.55,
"eval_cer": 13.056602568197789,
"eval_loss": 0.31062883138656616,
"eval_runtime": 1163.7705,
"eval_samples_per_second": 4.804,
"eval_steps_per_second": 0.301,
"step": 2000
},
{
"epoch": 0.56,
"learning_rate": 4.47448275862069e-05,
"loss": 0.4987,
"step": 2025
},
{
"epoch": 0.57,
"learning_rate": 4.4658620689655175e-05,
"loss": 0.5091,
"step": 2050
},
{
"epoch": 0.57,
"learning_rate": 4.457241379310345e-05,
"loss": 0.5096,
"step": 2075
},
{
"epoch": 0.58,
"learning_rate": 4.4486206896551724e-05,
"loss": 0.4716,
"step": 2100
},
{
"epoch": 0.59,
"learning_rate": 4.44e-05,
"loss": 0.4618,
"step": 2125
},
{
"epoch": 0.6,
"learning_rate": 4.431379310344827e-05,
"loss": 0.4588,
"step": 2150
},
{
"epoch": 0.6,
"learning_rate": 4.422758620689656e-05,
"loss": 0.5037,
"step": 2175
},
{
"epoch": 0.61,
"learning_rate": 4.414137931034483e-05,
"loss": 0.4818,
"step": 2200
},
{
"epoch": 0.62,
"learning_rate": 4.4055172413793106e-05,
"loss": 0.4802,
"step": 2225
},
{
"epoch": 0.62,
"learning_rate": 4.3968965517241384e-05,
"loss": 0.4889,
"step": 2250
},
{
"epoch": 0.63,
"learning_rate": 4.3882758620689655e-05,
"loss": 0.4914,
"step": 2275
},
{
"epoch": 0.64,
"learning_rate": 4.379655172413793e-05,
"loss": 0.4867,
"step": 2300
},
{
"epoch": 0.64,
"learning_rate": 4.371034482758621e-05,
"loss": 0.5006,
"step": 2325
},
{
"epoch": 0.65,
"learning_rate": 4.362413793103449e-05,
"loss": 0.4705,
"step": 2350
},
{
"epoch": 0.66,
"learning_rate": 4.353793103448276e-05,
"loss": 0.482,
"step": 2375
},
{
"epoch": 0.66,
"learning_rate": 4.3451724137931037e-05,
"loss": 0.4818,
"step": 2400
},
{
"epoch": 0.67,
"learning_rate": 4.336551724137931e-05,
"loss": 0.4478,
"step": 2425
},
{
"epoch": 0.68,
"learning_rate": 4.327931034482759e-05,
"loss": 0.4759,
"step": 2450
},
{
"epoch": 0.69,
"learning_rate": 4.319310344827586e-05,
"loss": 0.4518,
"step": 2475
},
{
"epoch": 0.69,
"learning_rate": 4.310689655172414e-05,
"loss": 0.4935,
"step": 2500
},
{
"epoch": 0.69,
"eval_cer": 13.829617325752253,
"eval_loss": 0.3001407980918884,
"eval_runtime": 1213.7136,
"eval_samples_per_second": 4.607,
"eval_steps_per_second": 0.288,
"step": 2500
},
{
"epoch": 0.7,
"learning_rate": 4.302068965517241e-05,
"loss": 0.4943,
"step": 2525
},
{
"epoch": 0.71,
"learning_rate": 4.293448275862069e-05,
"loss": 0.4451,
"step": 2550
},
{
"epoch": 0.71,
"learning_rate": 4.284827586206897e-05,
"loss": 0.456,
"step": 2575
},
{
"epoch": 0.72,
"learning_rate": 4.2762068965517245e-05,
"loss": 0.4944,
"step": 2600
},
{
"epoch": 0.73,
"learning_rate": 4.267586206896552e-05,
"loss": 0.4523,
"step": 2625
},
{
"epoch": 0.73,
"learning_rate": 4.2589655172413794e-05,
"loss": 0.4865,
"step": 2650
},
{
"epoch": 0.74,
"learning_rate": 4.250344827586207e-05,
"loss": 0.4429,
"step": 2675
},
{
"epoch": 0.75,
"learning_rate": 4.241724137931035e-05,
"loss": 0.4705,
"step": 2700
},
{
"epoch": 0.75,
"learning_rate": 4.233103448275863e-05,
"loss": 0.437,
"step": 2725
},
{
"epoch": 0.76,
"learning_rate": 4.22448275862069e-05,
"loss": 0.4218,
"step": 2750
},
{
"epoch": 0.77,
"learning_rate": 4.2158620689655176e-05,
"loss": 0.5038,
"step": 2775
},
{
"epoch": 0.78,
"learning_rate": 4.2072413793103447e-05,
"loss": 0.4579,
"step": 2800
},
{
"epoch": 0.78,
"learning_rate": 4.1986206896551724e-05,
"loss": 0.4738,
"step": 2825
},
{
"epoch": 0.79,
"learning_rate": 4.19e-05,
"loss": 0.4642,
"step": 2850
},
{
"epoch": 0.8,
"learning_rate": 4.181379310344828e-05,
"loss": 0.4422,
"step": 2875
},
{
"epoch": 0.8,
"learning_rate": 4.172758620689655e-05,
"loss": 0.4212,
"step": 2900
},
{
"epoch": 0.81,
"learning_rate": 4.164137931034483e-05,
"loss": 0.4803,
"step": 2925
},
{
"epoch": 0.82,
"learning_rate": 4.1555172413793106e-05,
"loss": 0.4333,
"step": 2950
},
{
"epoch": 0.82,
"learning_rate": 4.1468965517241384e-05,
"loss": 0.4663,
"step": 2975
},
{
"epoch": 0.83,
"learning_rate": 4.1382758620689655e-05,
"loss": 0.4836,
"step": 3000
},
{
"epoch": 0.83,
"eval_cer": 13.110905257778061,
"eval_loss": 0.29782944917678833,
"eval_runtime": 1227.0769,
"eval_samples_per_second": 4.556,
"eval_steps_per_second": 0.285,
"step": 3000
},
{
"epoch": 0.84,
"learning_rate": 4.129655172413793e-05,
"loss": 0.482,
"step": 3025
},
{
"epoch": 0.84,
"learning_rate": 4.121034482758621e-05,
"loss": 0.4706,
"step": 3050
},
{
"epoch": 0.85,
"learning_rate": 4.112413793103448e-05,
"loss": 0.4658,
"step": 3075
},
{
"epoch": 0.86,
"learning_rate": 4.1037931034482766e-05,
"loss": 0.4624,
"step": 3100
},
{
"epoch": 0.87,
"learning_rate": 4.095172413793104e-05,
"loss": 0.4754,
"step": 3125
},
{
"epoch": 0.87,
"learning_rate": 4.0865517241379315e-05,
"loss": 0.4866,
"step": 3150
},
{
"epoch": 0.88,
"learning_rate": 4.0779310344827586e-05,
"loss": 0.4723,
"step": 3175
},
{
"epoch": 0.89,
"learning_rate": 4.069310344827586e-05,
"loss": 0.4707,
"step": 3200
},
{
"epoch": 0.89,
"learning_rate": 4.060689655172414e-05,
"loss": 0.3989,
"step": 3225
},
{
"epoch": 0.9,
"learning_rate": 4.052068965517242e-05,
"loss": 0.4714,
"step": 3250
},
{
"epoch": 0.91,
"learning_rate": 4.043448275862069e-05,
"loss": 0.4451,
"step": 3275
},
{
"epoch": 0.91,
"learning_rate": 4.034827586206897e-05,
"loss": 0.4785,
"step": 3300
},
{
"epoch": 0.92,
"learning_rate": 4.026206896551724e-05,
"loss": 0.4509,
"step": 3325
},
{
"epoch": 0.93,
"learning_rate": 4.017586206896552e-05,
"loss": 0.4523,
"step": 3350
},
{
"epoch": 0.94,
"learning_rate": 4.0089655172413794e-05,
"loss": 0.4264,
"step": 3375
},
{
"epoch": 0.94,
"learning_rate": 4.000344827586207e-05,
"loss": 0.4685,
"step": 3400
},
{
"epoch": 0.95,
"learning_rate": 3.991724137931035e-05,
"loss": 0.4632,
"step": 3425
},
{
"epoch": 0.96,
"learning_rate": 3.983103448275862e-05,
"loss": 0.4548,
"step": 3450
},
{
"epoch": 0.96,
"learning_rate": 3.97448275862069e-05,
"loss": 0.4306,
"step": 3475
},
{
"epoch": 0.97,
"learning_rate": 3.9658620689655176e-05,
"loss": 0.4782,
"step": 3500
},
{
"epoch": 0.97,
"eval_cer": 12.711620775570179,
"eval_loss": 0.2923573851585388,
"eval_runtime": 1186.1965,
"eval_samples_per_second": 4.713,
"eval_steps_per_second": 0.295,
"step": 3500
},
{
"epoch": 0.98,
"learning_rate": 3.9572413793103454e-05,
"loss": 0.4219,
"step": 3525
},
{
"epoch": 0.98,
"learning_rate": 3.9486206896551725e-05,
"loss": 0.471,
"step": 3550
},
{
"epoch": 0.99,
"learning_rate": 3.94e-05,
"loss": 0.4687,
"step": 3575
},
{
"epoch": 1.0,
"learning_rate": 3.931379310344827e-05,
"loss": 0.4356,
"step": 3600
},
{
"epoch": 1.0,
"learning_rate": 3.922758620689656e-05,
"loss": 0.4144,
"step": 3625
},
{
"epoch": 1.01,
"learning_rate": 3.914137931034483e-05,
"loss": 0.3605,
"step": 3650
},
{
"epoch": 1.02,
"learning_rate": 3.9055172413793106e-05,
"loss": 0.3652,
"step": 3675
},
{
"epoch": 1.03,
"learning_rate": 3.896896551724138e-05,
"loss": 0.3725,
"step": 3700
},
{
"epoch": 1.03,
"learning_rate": 3.8882758620689655e-05,
"loss": 0.3481,
"step": 3725
},
{
"epoch": 1.04,
"learning_rate": 3.879655172413793e-05,
"loss": 0.3662,
"step": 3750
},
{
"epoch": 1.05,
"learning_rate": 3.871034482758621e-05,
"loss": 0.3835,
"step": 3775
},
{
"epoch": 1.05,
"learning_rate": 3.862413793103448e-05,
"loss": 0.3638,
"step": 3800
},
{
"epoch": 1.06,
"learning_rate": 3.853793103448276e-05,
"loss": 0.3615,
"step": 3825
},
{
"epoch": 1.07,
"learning_rate": 3.845172413793104e-05,
"loss": 0.3327,
"step": 3850
},
{
"epoch": 1.07,
"learning_rate": 3.8365517241379315e-05,
"loss": 0.3679,
"step": 3875
},
{
"epoch": 1.08,
"learning_rate": 3.827931034482759e-05,
"loss": 0.3549,
"step": 3900
},
{
"epoch": 1.09,
"learning_rate": 3.8193103448275863e-05,
"loss": 0.3728,
"step": 3925
},
{
"epoch": 1.09,
"learning_rate": 3.810689655172414e-05,
"loss": 0.3493,
"step": 3950
},
{
"epoch": 1.1,
"learning_rate": 3.802068965517241e-05,
"loss": 0.3509,
"step": 3975
},
{
"epoch": 1.11,
"learning_rate": 3.793448275862069e-05,
"loss": 0.3441,
"step": 4000
},
{
"epoch": 1.11,
"eval_cer": 11.794863604420877,
"eval_loss": 0.2874628007411957,
"eval_runtime": 1164.5329,
"eval_samples_per_second": 4.801,
"eval_steps_per_second": 0.301,
"step": 4000
},
{
"epoch": 1.12,
"learning_rate": 3.784827586206897e-05,
"loss": 0.3513,
"step": 4025
},
{
"epoch": 1.12,
"learning_rate": 3.7762068965517245e-05,
"loss": 0.3442,
"step": 4050
},
{
"epoch": 1.13,
"learning_rate": 3.7675862068965516e-05,
"loss": 0.3183,
"step": 4075
},
{
"epoch": 1.14,
"learning_rate": 3.7589655172413794e-05,
"loss": 0.3674,
"step": 4100
},
{
"epoch": 1.14,
"learning_rate": 3.750344827586207e-05,
"loss": 0.3491,
"step": 4125
},
{
"epoch": 1.15,
"learning_rate": 3.741724137931035e-05,
"loss": 0.337,
"step": 4150
},
{
"epoch": 1.16,
"learning_rate": 3.733103448275862e-05,
"loss": 0.3711,
"step": 4175
},
{
"epoch": 1.16,
"learning_rate": 3.72448275862069e-05,
"loss": 0.3839,
"step": 4200
},
{
"epoch": 1.17,
"learning_rate": 3.7158620689655176e-05,
"loss": 0.3724,
"step": 4225
},
{
"epoch": 1.18,
"learning_rate": 3.707241379310345e-05,
"loss": 0.3444,
"step": 4250
},
{
"epoch": 1.18,
"learning_rate": 3.698620689655173e-05,
"loss": 0.3997,
"step": 4275
},
{
"epoch": 1.19,
"learning_rate": 3.69e-05,
"loss": 0.3742,
"step": 4300
},
{
"epoch": 1.2,
"learning_rate": 3.681379310344828e-05,
"loss": 0.3657,
"step": 4325
},
{
"epoch": 1.21,
"learning_rate": 3.672758620689655e-05,
"loss": 0.3762,
"step": 4350
},
{
"epoch": 1.21,
"learning_rate": 3.664137931034483e-05,
"loss": 0.3815,
"step": 4375
},
{
"epoch": 1.22,
"learning_rate": 3.655517241379311e-05,
"loss": 0.3628,
"step": 4400
},
{
"epoch": 1.23,
"learning_rate": 3.6468965517241384e-05,
"loss": 0.3533,
"step": 4425
},
{
"epoch": 1.23,
"learning_rate": 3.6382758620689655e-05,
"loss": 0.3857,
"step": 4450
},
{
"epoch": 1.24,
"learning_rate": 3.629655172413793e-05,
"loss": 0.3381,
"step": 4475
},
{
"epoch": 1.25,
"learning_rate": 3.6210344827586204e-05,
"loss": 0.3647,
"step": 4500
},
{
"epoch": 1.25,
"eval_cer": 11.617581294320578,
"eval_loss": 0.28388434648513794,
"eval_runtime": 1169.2298,
"eval_samples_per_second": 4.782,
"eval_steps_per_second": 0.299,
"step": 4500
},
{
"epoch": 1.25,
"learning_rate": 3.612413793103449e-05,
"loss": 0.3935,
"step": 4525
},
{
"epoch": 1.26,
"learning_rate": 3.603793103448276e-05,
"loss": 0.3514,
"step": 4550
},
{
"epoch": 1.27,
"learning_rate": 3.595172413793104e-05,
"loss": 0.3925,
"step": 4575
},
{
"epoch": 1.27,
"learning_rate": 3.586551724137931e-05,
"loss": 0.355,
"step": 4600
},
{
"epoch": 1.28,
"learning_rate": 3.5779310344827586e-05,
"loss": 0.3539,
"step": 4625
},
{
"epoch": 1.29,
"learning_rate": 3.5693103448275864e-05,
"loss": 0.3583,
"step": 4650
},
{
"epoch": 1.3,
"learning_rate": 3.560689655172414e-05,
"loss": 0.3312,
"step": 4675
},
{
"epoch": 1.3,
"learning_rate": 3.552068965517242e-05,
"loss": 0.331,
"step": 4700
},
{
"epoch": 1.31,
"learning_rate": 3.543448275862069e-05,
"loss": 0.3703,
"step": 4725
},
{
"epoch": 1.32,
"learning_rate": 3.534827586206897e-05,
"loss": 0.3599,
"step": 4750
},
{
"epoch": 1.32,
"learning_rate": 3.526206896551724e-05,
"loss": 0.3779,
"step": 4775
},
{
"epoch": 1.33,
"learning_rate": 3.517586206896552e-05,
"loss": 0.3338,
"step": 4800
},
{
"epoch": 1.34,
"learning_rate": 3.5089655172413794e-05,
"loss": 0.3249,
"step": 4825
},
{
"epoch": 1.34,
"learning_rate": 3.500344827586207e-05,
"loss": 0.364,
"step": 4850
},
{
"epoch": 1.35,
"learning_rate": 3.491724137931034e-05,
"loss": 0.3755,
"step": 4875
},
{
"epoch": 1.36,
"learning_rate": 3.483103448275862e-05,
"loss": 0.353,
"step": 4900
},
{
"epoch": 1.36,
"learning_rate": 3.47448275862069e-05,
"loss": 0.3483,
"step": 4925
},
{
"epoch": 1.37,
"learning_rate": 3.4658620689655176e-05,
"loss": 0.3675,
"step": 4950
},
{
"epoch": 1.38,
"learning_rate": 3.457241379310345e-05,
"loss": 0.3779,
"step": 4975
},
{
"epoch": 1.39,
"learning_rate": 3.4486206896551725e-05,
"loss": 0.3642,
"step": 5000
},
{
"epoch": 1.39,
"eval_cer": 11.572861432313294,
"eval_loss": 0.28437113761901855,
"eval_runtime": 1173.2699,
"eval_samples_per_second": 4.765,
"eval_steps_per_second": 0.298,
"step": 5000
},
{
"epoch": 1.39,
"learning_rate": 3.4399999999999996e-05,
"loss": 0.3872,
"step": 5025
},
{
"epoch": 1.4,
"learning_rate": 3.431379310344828e-05,
"loss": 0.392,
"step": 5050
},
{
"epoch": 1.41,
"learning_rate": 3.422758620689656e-05,
"loss": 0.3545,
"step": 5075
},
{
"epoch": 1.41,
"learning_rate": 3.41448275862069e-05,
"loss": 0.3627,
"step": 5100
},
{
"epoch": 1.42,
"learning_rate": 3.4058620689655175e-05,
"loss": 0.3362,
"step": 5125
},
{
"epoch": 1.43,
"learning_rate": 3.397241379310345e-05,
"loss": 0.3628,
"step": 5150
},
{
"epoch": 1.43,
"learning_rate": 3.3886206896551724e-05,
"loss": 0.3157,
"step": 5175
},
{
"epoch": 1.44,
"learning_rate": 3.38e-05,
"loss": 0.3649,
"step": 5200
},
{
"epoch": 1.45,
"learning_rate": 3.371379310344828e-05,
"loss": 0.3462,
"step": 5225
},
{
"epoch": 1.45,
"learning_rate": 3.362758620689656e-05,
"loss": 0.3833,
"step": 5250
},
{
"epoch": 1.46,
"learning_rate": 3.354137931034483e-05,
"loss": 0.3643,
"step": 5275
},
{
"epoch": 1.47,
"learning_rate": 3.3455172413793106e-05,
"loss": 0.3788,
"step": 5300
},
{
"epoch": 1.48,
"learning_rate": 3.336896551724138e-05,
"loss": 0.3625,
"step": 5325
},
{
"epoch": 1.48,
"learning_rate": 3.328275862068966e-05,
"loss": 0.363,
"step": 5350
},
{
"epoch": 1.49,
"learning_rate": 3.319655172413793e-05,
"loss": 0.3714,
"step": 5375
},
{
"epoch": 1.5,
"learning_rate": 3.311034482758621e-05,
"loss": 0.3636,
"step": 5400
},
{
"epoch": 1.5,
"learning_rate": 3.302413793103448e-05,
"loss": 0.3349,
"step": 5425
},
{
"epoch": 1.51,
"learning_rate": 3.293793103448276e-05,
"loss": 0.3516,
"step": 5450
},
{
"epoch": 1.52,
"learning_rate": 3.2851724137931036e-05,
"loss": 0.3327,
"step": 5475
},
{
"epoch": 1.52,
"learning_rate": 3.2765517241379314e-05,
"loss": 0.3493,
"step": 5500
},
{
"epoch": 1.52,
"eval_cer": 11.528141570306012,
"eval_loss": 0.2879265248775482,
"eval_runtime": 1178.4372,
"eval_samples_per_second": 4.744,
"eval_steps_per_second": 0.297,
"step": 5500
},
{
"epoch": 1.53,
"learning_rate": 3.2679310344827585e-05,
"loss": 0.3419,
"step": 5525
},
{
"epoch": 1.54,
"learning_rate": 3.259310344827586e-05,
"loss": 0.3876,
"step": 5550
},
{
"epoch": 1.54,
"learning_rate": 3.250689655172414e-05,
"loss": 0.3392,
"step": 5575
},
{
"epoch": 1.55,
"learning_rate": 3.242068965517241e-05,
"loss": 0.3781,
"step": 5600
},
{
"epoch": 1.56,
"learning_rate": 3.2334482758620696e-05,
"loss": 0.3616,
"step": 5625
},
{
"epoch": 1.57,
"learning_rate": 3.224827586206897e-05,
"loss": 0.3596,
"step": 5650
},
{
"epoch": 1.57,
"learning_rate": 3.2162068965517245e-05,
"loss": 0.3681,
"step": 5675
},
{
"epoch": 1.58,
"learning_rate": 3.2075862068965516e-05,
"loss": 0.3373,
"step": 5700
},
{
"epoch": 1.59,
"learning_rate": 3.1989655172413794e-05,
"loss": 0.3311,
"step": 5725
},
{
"epoch": 1.59,
"learning_rate": 3.190344827586207e-05,
"loss": 0.4004,
"step": 5750
},
{
"epoch": 1.6,
"learning_rate": 3.181724137931035e-05,
"loss": 0.3731,
"step": 5775
},
{
"epoch": 1.61,
"learning_rate": 3.173103448275862e-05,
"loss": 0.3443,
"step": 5800
},
{
"epoch": 1.61,
"learning_rate": 3.16448275862069e-05,
"loss": 0.345,
"step": 5825
},
{
"epoch": 1.62,
"learning_rate": 3.155862068965517e-05,
"loss": 0.312,
"step": 5850
},
{
"epoch": 1.63,
"learning_rate": 3.147241379310345e-05,
"loss": 0.3474,
"step": 5875
},
{
"epoch": 1.63,
"learning_rate": 3.1386206896551724e-05,
"loss": 0.3162,
"step": 5900
},
{
"epoch": 1.64,
"learning_rate": 3.13e-05,
"loss": 0.3227,
"step": 5925
},
{
"epoch": 1.65,
"learning_rate": 3.121379310344828e-05,
"loss": 0.3482,
"step": 5950
},
{
"epoch": 1.66,
"learning_rate": 3.112758620689655e-05,
"loss": 0.3398,
"step": 5975
},
{
"epoch": 1.66,
"learning_rate": 3.104137931034483e-05,
"loss": 0.3466,
"step": 6000
},
{
"epoch": 1.66,
"eval_cer": 11.44349326007794,
"eval_loss": 0.28201985359191895,
"eval_runtime": 1169.509,
"eval_samples_per_second": 4.781,
"eval_steps_per_second": 0.299,
"step": 6000
},
{
"epoch": 1.67,
"learning_rate": 3.0955172413793106e-05,
"loss": 0.3256,
"step": 6025
},
{
"epoch": 1.68,
"learning_rate": 3.0868965517241384e-05,
"loss": 0.3208,
"step": 6050
},
{
"epoch": 1.68,
"learning_rate": 3.0782758620689655e-05,
"loss": 0.3267,
"step": 6075
},
{
"epoch": 1.69,
"learning_rate": 3.069655172413793e-05,
"loss": 0.3675,
"step": 6100
},
{
"epoch": 1.7,
"learning_rate": 3.061034482758621e-05,
"loss": 0.3502,
"step": 6125
},
{
"epoch": 1.7,
"learning_rate": 3.052413793103449e-05,
"loss": 0.3586,
"step": 6150
},
{
"epoch": 1.71,
"learning_rate": 3.043793103448276e-05,
"loss": 0.3211,
"step": 6175
},
{
"epoch": 1.72,
"learning_rate": 3.0351724137931037e-05,
"loss": 0.3631,
"step": 6200
},
{
"epoch": 1.72,
"learning_rate": 3.026551724137931e-05,
"loss": 0.3764,
"step": 6225
},
{
"epoch": 1.73,
"learning_rate": 3.017931034482759e-05,
"loss": 0.3636,
"step": 6250
},
{
"epoch": 1.74,
"learning_rate": 3.009310344827586e-05,
"loss": 0.3682,
"step": 6275
},
{
"epoch": 1.75,
"learning_rate": 3.000689655172414e-05,
"loss": 0.3484,
"step": 6300
},
{
"epoch": 1.75,
"learning_rate": 2.9920689655172412e-05,
"loss": 0.336,
"step": 6325
},
{
"epoch": 1.76,
"learning_rate": 2.983448275862069e-05,
"loss": 0.3457,
"step": 6350
},
{
"epoch": 1.77,
"learning_rate": 2.974827586206897e-05,
"loss": 0.3548,
"step": 6375
},
{
"epoch": 1.77,
"learning_rate": 2.966206896551724e-05,
"loss": 0.3455,
"step": 6400
},
{
"epoch": 1.78,
"learning_rate": 2.957586206896552e-05,
"loss": 0.3303,
"step": 6425
},
{
"epoch": 1.79,
"learning_rate": 2.9489655172413794e-05,
"loss": 0.3344,
"step": 6450
},
{
"epoch": 1.79,
"learning_rate": 2.940344827586207e-05,
"loss": 0.3517,
"step": 6475
},
{
"epoch": 1.8,
"learning_rate": 2.9317241379310346e-05,
"loss": 0.3507,
"step": 6500
},
{
"epoch": 1.8,
"eval_cer": 11.012266019293426,
"eval_loss": 0.27717769145965576,
"eval_runtime": 1163.8916,
"eval_samples_per_second": 4.804,
"eval_steps_per_second": 0.301,
"step": 6500
},
{
"epoch": 1.81,
"learning_rate": 2.9231034482758624e-05,
"loss": 0.3283,
"step": 6525
},
{
"epoch": 1.81,
"learning_rate": 2.9144827586206898e-05,
"loss": 0.3808,
"step": 6550
},
{
"epoch": 1.82,
"learning_rate": 2.9058620689655176e-05,
"loss": 0.3473,
"step": 6575
},
{
"epoch": 1.83,
"learning_rate": 2.8972413793103447e-05,
"loss": 0.3502,
"step": 6600
},
{
"epoch": 1.84,
"learning_rate": 2.8886206896551728e-05,
"loss": 0.3507,
"step": 6625
},
{
"epoch": 1.84,
"learning_rate": 2.88e-05,
"loss": 0.3086,
"step": 6650
},
{
"epoch": 1.85,
"learning_rate": 2.8713793103448276e-05,
"loss": 0.3266,
"step": 6675
},
{
"epoch": 1.86,
"learning_rate": 2.862758620689655e-05,
"loss": 0.3236,
"step": 6700
},
{
"epoch": 1.86,
"learning_rate": 2.854137931034483e-05,
"loss": 0.3063,
"step": 6725
},
{
"epoch": 1.87,
"learning_rate": 2.8455172413793106e-05,
"loss": 0.3357,
"step": 6750
},
{
"epoch": 1.88,
"learning_rate": 2.836896551724138e-05,
"loss": 0.3366,
"step": 6775
},
{
"epoch": 1.88,
"learning_rate": 2.828275862068966e-05,
"loss": 0.3459,
"step": 6800
},
{
"epoch": 1.89,
"learning_rate": 2.8196551724137933e-05,
"loss": 0.3396,
"step": 6825
},
{
"epoch": 1.9,
"learning_rate": 2.811034482758621e-05,
"loss": 0.3388,
"step": 6850
},
{
"epoch": 1.9,
"learning_rate": 2.8024137931034485e-05,
"loss": 0.3657,
"step": 6875
},
{
"epoch": 1.91,
"learning_rate": 2.7937931034482763e-05,
"loss": 0.3604,
"step": 6900
},
{
"epoch": 1.92,
"learning_rate": 2.7851724137931033e-05,
"loss": 0.3467,
"step": 6925
},
{
"epoch": 1.93,
"learning_rate": 2.7765517241379315e-05,
"loss": 0.3434,
"step": 6950
},
{
"epoch": 1.93,
"learning_rate": 2.7679310344827586e-05,
"loss": 0.3416,
"step": 6975
},
{
"epoch": 1.94,
"learning_rate": 2.7593103448275863e-05,
"loss": 0.3482,
"step": 7000
},
{
"epoch": 1.94,
"eval_cer": 11.178368363891906,
"eval_loss": 0.2766138017177582,
"eval_runtime": 1182.8471,
"eval_samples_per_second": 4.727,
"eval_steps_per_second": 0.296,
"step": 7000
},
{
"epoch": 1.95,
"learning_rate": 2.7506896551724138e-05,
"loss": 0.3213,
"step": 7025
},
{
"epoch": 1.95,
"learning_rate": 2.7420689655172415e-05,
"loss": 0.359,
"step": 7050
},
{
"epoch": 1.96,
"learning_rate": 2.733448275862069e-05,
"loss": 0.3506,
"step": 7075
},
{
"epoch": 1.97,
"learning_rate": 2.7248275862068968e-05,
"loss": 0.3178,
"step": 7100
},
{
"epoch": 1.97,
"learning_rate": 2.716206896551724e-05,
"loss": 0.3371,
"step": 7125
},
{
"epoch": 1.98,
"learning_rate": 2.707586206896552e-05,
"loss": 0.3236,
"step": 7150
},
{
"epoch": 1.99,
"learning_rate": 2.6989655172413797e-05,
"loss": 0.3231,
"step": 7175
},
{
"epoch": 1.99,
"learning_rate": 2.6903448275862068e-05,
"loss": 0.3007,
"step": 7200
},
{
"epoch": 2.0,
"learning_rate": 2.6820689655172414e-05,
"loss": 0.324,
"step": 7225
},
{
"epoch": 2.01,
"learning_rate": 2.6734482758620692e-05,
"loss": 0.2574,
"step": 7250
},
{
"epoch": 2.02,
"learning_rate": 2.6648275862068966e-05,
"loss": 0.2346,
"step": 7275
},
{
"epoch": 2.02,
"learning_rate": 2.6562068965517244e-05,
"loss": 0.2627,
"step": 7300
},
{
"epoch": 2.03,
"learning_rate": 2.647586206896552e-05,
"loss": 0.2614,
"step": 7325
},
{
"epoch": 2.04,
"learning_rate": 2.6389655172413796e-05,
"loss": 0.2279,
"step": 7350
},
{
"epoch": 2.04,
"learning_rate": 2.630344827586207e-05,
"loss": 0.271,
"step": 7375
},
{
"epoch": 2.05,
"learning_rate": 2.621724137931035e-05,
"loss": 0.2566,
"step": 7400
},
{
"epoch": 2.06,
"learning_rate": 2.613103448275862e-05,
"loss": 0.2471,
"step": 7425
},
{
"epoch": 2.06,
"learning_rate": 2.60448275862069e-05,
"loss": 0.2438,
"step": 7450
},
{
"epoch": 2.07,
"learning_rate": 2.595862068965517e-05,
"loss": 0.2457,
"step": 7475
},
{
"epoch": 2.08,
"learning_rate": 2.587241379310345e-05,
"loss": 0.232,
"step": 7500
},
{
"epoch": 2.08,
"eval_cer": 10.926020571136522,
"eval_loss": 0.2784821093082428,
"eval_runtime": 1168.6896,
"eval_samples_per_second": 4.784,
"eval_steps_per_second": 0.299,
"step": 7500
},
{
"epoch": 2.09,
"learning_rate": 2.5786206896551724e-05,
"loss": 0.256,
"step": 7525
},
{
"epoch": 2.09,
"learning_rate": 2.57e-05,
"loss": 0.2383,
"step": 7550
},
{
"epoch": 2.1,
"learning_rate": 2.5613793103448276e-05,
"loss": 0.2516,
"step": 7575
},
{
"epoch": 2.11,
"learning_rate": 2.5527586206896553e-05,
"loss": 0.229,
"step": 7600
},
{
"epoch": 2.11,
"learning_rate": 2.5441379310344828e-05,
"loss": 0.242,
"step": 7625
},
{
"epoch": 2.12,
"learning_rate": 2.5355172413793105e-05,
"loss": 0.243,
"step": 7650
},
{
"epoch": 2.13,
"learning_rate": 2.5268965517241383e-05,
"loss": 0.2476,
"step": 7675
},
{
"epoch": 2.13,
"learning_rate": 2.5182758620689658e-05,
"loss": 0.2351,
"step": 7700
},
{
"epoch": 2.14,
"learning_rate": 2.5096551724137935e-05,
"loss": 0.2526,
"step": 7725
},
{
"epoch": 2.15,
"learning_rate": 2.5010344827586206e-05,
"loss": 0.2424,
"step": 7750
},
{
"epoch": 2.15,
"learning_rate": 2.4924137931034484e-05,
"loss": 0.264,
"step": 7775
},
{
"epoch": 2.16,
"learning_rate": 2.483793103448276e-05,
"loss": 0.244,
"step": 7800
},
{
"epoch": 2.17,
"learning_rate": 2.4751724137931036e-05,
"loss": 0.2588,
"step": 7825
},
{
"epoch": 2.18,
"learning_rate": 2.4665517241379314e-05,
"loss": 0.2418,
"step": 7850
},
{
"epoch": 2.18,
"learning_rate": 2.4579310344827588e-05,
"loss": 0.2556,
"step": 7875
},
{
"epoch": 2.19,
"learning_rate": 2.4493103448275866e-05,
"loss": 0.2253,
"step": 7900
},
{
"epoch": 2.2,
"learning_rate": 2.440689655172414e-05,
"loss": 0.2629,
"step": 7925
},
{
"epoch": 2.2,
"learning_rate": 2.4320689655172415e-05,
"loss": 0.2512,
"step": 7950
},
{
"epoch": 2.21,
"learning_rate": 2.4234482758620692e-05,
"loss": 0.2186,
"step": 7975
},
{
"epoch": 2.22,
"learning_rate": 2.4148275862068967e-05,
"loss": 0.2539,
"step": 8000
},
{
"epoch": 2.22,
"eval_cer": 10.593815881939564,
"eval_loss": 0.2776859700679779,
"eval_runtime": 1165.9559,
"eval_samples_per_second": 4.795,
"eval_steps_per_second": 0.3,
"step": 8000
},
{
"epoch": 2.22,
"learning_rate": 2.406206896551724e-05,
"loss": 0.2433,
"step": 8025
},
{
"epoch": 2.23,
"learning_rate": 2.397586206896552e-05,
"loss": 0.2383,
"step": 8050
},
{
"epoch": 2.24,
"learning_rate": 2.3889655172413793e-05,
"loss": 0.2454,
"step": 8075
},
{
"epoch": 2.24,
"learning_rate": 2.380344827586207e-05,
"loss": 0.2596,
"step": 8100
},
{
"epoch": 2.25,
"learning_rate": 2.3717241379310345e-05,
"loss": 0.2479,
"step": 8125
},
{
"epoch": 2.26,
"learning_rate": 2.363103448275862e-05,
"loss": 0.2519,
"step": 8150
},
{
"epoch": 2.27,
"learning_rate": 2.3544827586206897e-05,
"loss": 0.2394,
"step": 8175
},
{
"epoch": 2.27,
"learning_rate": 2.345862068965517e-05,
"loss": 0.2336,
"step": 8200
},
{
"epoch": 2.28,
"learning_rate": 2.337241379310345e-05,
"loss": 0.2549,
"step": 8225
},
{
"epoch": 2.29,
"learning_rate": 2.3286206896551727e-05,
"loss": 0.2784,
"step": 8250
},
{
"epoch": 2.29,
"learning_rate": 2.32e-05,
"loss": 0.2365,
"step": 8275
},
{
"epoch": 2.3,
"learning_rate": 2.311379310344828e-05,
"loss": 0.2445,
"step": 8300
},
{
"epoch": 2.31,
"learning_rate": 2.3027586206896554e-05,
"loss": 0.2411,
"step": 8325
},
{
"epoch": 2.31,
"learning_rate": 2.2941379310344828e-05,
"loss": 0.257,
"step": 8350
},
{
"epoch": 2.32,
"learning_rate": 2.2855172413793106e-05,
"loss": 0.2196,
"step": 8375
},
{
"epoch": 2.33,
"learning_rate": 2.276896551724138e-05,
"loss": 0.2583,
"step": 8400
},
{
"epoch": 2.33,
"learning_rate": 2.2682758620689658e-05,
"loss": 0.2522,
"step": 8425
},
{
"epoch": 2.34,
"learning_rate": 2.2596551724137932e-05,
"loss": 0.2712,
"step": 8450
},
{
"epoch": 2.35,
"learning_rate": 2.2510344827586206e-05,
"loss": 0.2624,
"step": 8475
},
{
"epoch": 2.36,
"learning_rate": 2.2424137931034484e-05,
"loss": 0.2377,
"step": 8500
},
{
"epoch": 2.36,
"eval_cer": 10.659298537021657,
"eval_loss": 0.2783927917480469,
"eval_runtime": 1180.6736,
"eval_samples_per_second": 4.735,
"eval_steps_per_second": 0.296,
"step": 8500
},
{
"epoch": 2.36,
"learning_rate": 2.233793103448276e-05,
"loss": 0.2526,
"step": 8525
},
{
"epoch": 2.37,
"learning_rate": 2.2251724137931036e-05,
"loss": 0.2408,
"step": 8550
},
{
"epoch": 2.38,
"learning_rate": 2.216551724137931e-05,
"loss": 0.2778,
"step": 8575
},
{
"epoch": 2.38,
"learning_rate": 2.2079310344827585e-05,
"loss": 0.253,
"step": 8600
},
{
"epoch": 2.39,
"learning_rate": 2.1993103448275863e-05,
"loss": 0.2416,
"step": 8625
},
{
"epoch": 2.4,
"learning_rate": 2.190689655172414e-05,
"loss": 0.2477,
"step": 8650
},
{
"epoch": 2.4,
"learning_rate": 2.1820689655172415e-05,
"loss": 0.2453,
"step": 8675
},
{
"epoch": 2.41,
"learning_rate": 2.1734482758620693e-05,
"loss": 0.2463,
"step": 8700
},
{
"epoch": 2.42,
"learning_rate": 2.1648275862068967e-05,
"loss": 0.2409,
"step": 8725
},
{
"epoch": 2.42,
"learning_rate": 2.1562068965517245e-05,
"loss": 0.2305,
"step": 8750
},
{
"epoch": 2.43,
"learning_rate": 2.147586206896552e-05,
"loss": 0.2342,
"step": 8775
},
{
"epoch": 2.44,
"learning_rate": 2.1389655172413793e-05,
"loss": 0.2551,
"step": 8800
},
{
"epoch": 2.45,
"learning_rate": 2.130344827586207e-05,
"loss": 0.2396,
"step": 8825
},
{
"epoch": 2.45,
"learning_rate": 2.1217241379310345e-05,
"loss": 0.268,
"step": 8850
},
{
"epoch": 2.46,
"learning_rate": 2.1131034482758623e-05,
"loss": 0.2384,
"step": 8875
},
{
"epoch": 2.47,
"learning_rate": 2.1044827586206898e-05,
"loss": 0.2418,
"step": 8900
},
{
"epoch": 2.47,
"learning_rate": 2.0958620689655172e-05,
"loss": 0.2414,
"step": 8925
},
{
"epoch": 2.48,
"learning_rate": 2.087241379310345e-05,
"loss": 0.2232,
"step": 8950
},
{
"epoch": 2.49,
"learning_rate": 2.0786206896551724e-05,
"loss": 0.248,
"step": 8975
},
{
"epoch": 2.49,
"learning_rate": 2.07e-05,
"loss": 0.2384,
"step": 9000
},
{
"epoch": 2.49,
"eval_cer": 10.659298537021657,
"eval_loss": 0.2808603048324585,
"eval_runtime": 1165.5072,
"eval_samples_per_second": 4.797,
"eval_steps_per_second": 0.3,
"step": 9000
},
{
"epoch": 2.5,
"learning_rate": 2.0613793103448276e-05,
"loss": 0.2448,
"step": 9025
},
{
"epoch": 2.51,
"learning_rate": 2.0527586206896554e-05,
"loss": 0.2268,
"step": 9050
},
{
"epoch": 2.51,
"learning_rate": 2.0441379310344828e-05,
"loss": 0.247,
"step": 9075
},
{
"epoch": 2.52,
"learning_rate": 2.0355172413793106e-05,
"loss": 0.2345,
"step": 9100
},
{
"epoch": 2.53,
"learning_rate": 2.026896551724138e-05,
"loss": 0.2117,
"step": 9125
},
{
"epoch": 2.54,
"learning_rate": 2.0182758620689658e-05,
"loss": 0.2265,
"step": 9150
},
{
"epoch": 2.54,
"learning_rate": 2.0096551724137932e-05,
"loss": 0.2464,
"step": 9175
},
{
"epoch": 2.55,
"learning_rate": 2.0010344827586207e-05,
"loss": 0.2741,
"step": 9200
},
{
"epoch": 2.56,
"learning_rate": 1.9924137931034484e-05,
"loss": 0.2397,
"step": 9225
},
{
"epoch": 2.56,
"learning_rate": 1.983793103448276e-05,
"loss": 0.2357,
"step": 9250
},
{
"epoch": 2.57,
"learning_rate": 1.9751724137931037e-05,
"loss": 0.2435,
"step": 9275
},
{
"epoch": 2.58,
"learning_rate": 1.966551724137931e-05,
"loss": 0.2571,
"step": 9300
},
{
"epoch": 2.58,
"learning_rate": 1.9579310344827585e-05,
"loss": 0.2675,
"step": 9325
},
{
"epoch": 2.59,
"learning_rate": 1.9493103448275863e-05,
"loss": 0.2476,
"step": 9350
},
{
"epoch": 2.6,
"learning_rate": 1.9406896551724137e-05,
"loss": 0.2412,
"step": 9375
},
{
"epoch": 2.6,
"learning_rate": 1.9320689655172415e-05,
"loss": 0.2502,
"step": 9400
},
{
"epoch": 2.61,
"learning_rate": 1.923448275862069e-05,
"loss": 0.2327,
"step": 9425
},
{
"epoch": 2.62,
"learning_rate": 1.9148275862068964e-05,
"loss": 0.2198,
"step": 9450
},
{
"epoch": 2.63,
"learning_rate": 1.9062068965517245e-05,
"loss": 0.2614,
"step": 9475
},
{
"epoch": 2.63,
"learning_rate": 1.897586206896552e-05,
"loss": 0.2495,
"step": 9500
},
{
"epoch": 2.63,
"eval_cer": 10.632147192231521,
"eval_loss": 0.281376451253891,
"eval_runtime": 1164.8408,
"eval_samples_per_second": 4.8,
"eval_steps_per_second": 0.3,
"step": 9500
},
{
"epoch": 2.64,
"learning_rate": 1.8889655172413794e-05,
"loss": 0.2442,
"step": 9525
},
{
"epoch": 2.65,
"learning_rate": 1.880344827586207e-05,
"loss": 0.2369,
"step": 9550
},
{
"epoch": 2.65,
"learning_rate": 1.8717241379310346e-05,
"loss": 0.2186,
"step": 9575
},
{
"epoch": 2.66,
"learning_rate": 1.8631034482758623e-05,
"loss": 0.234,
"step": 9600
},
{
"epoch": 2.67,
"learning_rate": 1.8544827586206898e-05,
"loss": 0.2448,
"step": 9625
},
{
"epoch": 2.67,
"learning_rate": 1.8458620689655172e-05,
"loss": 0.2449,
"step": 9650
},
{
"epoch": 2.68,
"learning_rate": 1.837241379310345e-05,
"loss": 0.2322,
"step": 9675
},
{
"epoch": 2.69,
"learning_rate": 1.8286206896551724e-05,
"loss": 0.2489,
"step": 9700
},
{
"epoch": 2.69,
"learning_rate": 1.8200000000000002e-05,
"loss": 0.2569,
"step": 9725
},
{
"epoch": 2.7,
"learning_rate": 1.8113793103448276e-05,
"loss": 0.2331,
"step": 9750
},
{
"epoch": 2.71,
"learning_rate": 1.802758620689655e-05,
"loss": 0.2647,
"step": 9775
},
{
"epoch": 2.72,
"learning_rate": 1.794137931034483e-05,
"loss": 0.2426,
"step": 9800
},
{
"epoch": 2.72,
"learning_rate": 1.7855172413793103e-05,
"loss": 0.2341,
"step": 9825
},
{
"epoch": 2.73,
"learning_rate": 1.776896551724138e-05,
"loss": 0.24,
"step": 9850
},
{
"epoch": 2.74,
"learning_rate": 1.7682758620689658e-05,
"loss": 0.249,
"step": 9875
},
{
"epoch": 2.74,
"learning_rate": 1.7596551724137933e-05,
"loss": 0.2217,
"step": 9900
},
{
"epoch": 2.75,
"learning_rate": 1.751034482758621e-05,
"loss": 0.2467,
"step": 9925
},
{
"epoch": 2.76,
"learning_rate": 1.7424137931034485e-05,
"loss": 0.2529,
"step": 9950
},
{
"epoch": 2.76,
"learning_rate": 1.733793103448276e-05,
"loss": 0.2578,
"step": 9975
},
{
"epoch": 2.77,
"learning_rate": 1.7251724137931037e-05,
"loss": 0.2312,
"step": 10000
},
{
"epoch": 2.77,
"eval_cer": 10.60020443365489,
"eval_loss": 0.28221753239631653,
"eval_runtime": 1168.5035,
"eval_samples_per_second": 4.785,
"eval_steps_per_second": 0.3,
"step": 10000
},
{
"epoch": 2.78,
"learning_rate": 1.716551724137931e-05,
"loss": 0.2284,
"step": 10025
},
{
"epoch": 2.78,
"learning_rate": 1.7079310344827585e-05,
"loss": 0.2348,
"step": 10050
},
{
"epoch": 2.79,
"learning_rate": 1.6993103448275863e-05,
"loss": 0.2244,
"step": 10075
},
{
"epoch": 2.8,
"learning_rate": 1.6906896551724138e-05,
"loss": 0.2522,
"step": 10100
},
{
"epoch": 2.81,
"learning_rate": 1.6820689655172415e-05,
"loss": 0.2476,
"step": 10125
},
{
"epoch": 2.81,
"learning_rate": 1.673448275862069e-05,
"loss": 0.2279,
"step": 10150
},
{
"epoch": 2.82,
"learning_rate": 1.6648275862068964e-05,
"loss": 0.2317,
"step": 10175
},
{
"epoch": 2.83,
"learning_rate": 1.6562068965517242e-05,
"loss": 0.2305,
"step": 10200
},
{
"epoch": 2.83,
"learning_rate": 1.6475862068965516e-05,
"loss": 0.2364,
"step": 10225
},
{
"epoch": 2.84,
"learning_rate": 1.6389655172413794e-05,
"loss": 0.2572,
"step": 10250
},
{
"epoch": 2.85,
"learning_rate": 1.630344827586207e-05,
"loss": 0.2065,
"step": 10275
},
{
"epoch": 2.85,
"learning_rate": 1.6217241379310346e-05,
"loss": 0.2398,
"step": 10300
},
{
"epoch": 2.86,
"learning_rate": 1.6131034482758624e-05,
"loss": 0.2526,
"step": 10325
},
{
"epoch": 2.87,
"learning_rate": 1.6044827586206898e-05,
"loss": 0.2263,
"step": 10350
},
{
"epoch": 2.87,
"learning_rate": 1.5958620689655172e-05,
"loss": 0.2527,
"step": 10375
},
{
"epoch": 2.88,
"learning_rate": 1.587241379310345e-05,
"loss": 0.2401,
"step": 10400
},
{
"epoch": 2.89,
"learning_rate": 1.5786206896551724e-05,
"loss": 0.2221,
"step": 10425
},
{
"epoch": 2.9,
"learning_rate": 1.5700000000000002e-05,
"loss": 0.2251,
"step": 10450
},
{
"epoch": 2.9,
"learning_rate": 1.5613793103448276e-05,
"loss": 0.247,
"step": 10475
},
{
"epoch": 2.91,
"learning_rate": 1.552758620689655e-05,
"loss": 0.2264,
"step": 10500
},
{
"epoch": 2.91,
"eval_cer": 10.73436401967674,
"eval_loss": 0.2812344431877136,
"eval_runtime": 1186.9835,
"eval_samples_per_second": 4.71,
"eval_steps_per_second": 0.295,
"step": 10500
},
{
"epoch": 2.92,
"learning_rate": 1.544137931034483e-05,
"loss": 0.2459,
"step": 10525
},
{
"epoch": 2.92,
"learning_rate": 1.5355172413793103e-05,
"loss": 0.2541,
"step": 10550
},
{
"epoch": 2.93,
"learning_rate": 1.526896551724138e-05,
"loss": 0.2265,
"step": 10575
},
{
"epoch": 2.94,
"learning_rate": 1.5182758620689655e-05,
"loss": 0.2329,
"step": 10600
},
{
"epoch": 2.94,
"learning_rate": 1.5096551724137931e-05,
"loss": 0.2245,
"step": 10625
},
{
"epoch": 2.95,
"learning_rate": 1.5010344827586207e-05,
"loss": 0.265,
"step": 10650
},
{
"epoch": 2.96,
"learning_rate": 1.4924137931034485e-05,
"loss": 0.2423,
"step": 10675
},
{
"epoch": 2.96,
"learning_rate": 1.4837931034482761e-05,
"loss": 0.2319,
"step": 10700
},
{
"epoch": 2.97,
"learning_rate": 1.4751724137931037e-05,
"loss": 0.2286,
"step": 10725
},
{
"epoch": 2.98,
"learning_rate": 1.4665517241379311e-05,
"loss": 0.2352,
"step": 10750
},
{
"epoch": 2.99,
"learning_rate": 1.4579310344827587e-05,
"loss": 0.2635,
"step": 10775
},
{
"epoch": 2.99,
"learning_rate": 1.4493103448275863e-05,
"loss": 0.2269,
"step": 10800
},
{
"epoch": 3.0,
"learning_rate": 1.440689655172414e-05,
"loss": 0.2296,
"step": 10825
},
{
"epoch": 3.01,
"learning_rate": 1.4324137931034484e-05,
"loss": 0.1896,
"step": 10850
},
{
"epoch": 3.01,
"learning_rate": 1.423793103448276e-05,
"loss": 0.1824,
"step": 10875
},
{
"epoch": 3.02,
"learning_rate": 1.4151724137931036e-05,
"loss": 0.1621,
"step": 10900
},
{
"epoch": 3.03,
"learning_rate": 1.4065517241379312e-05,
"loss": 0.1497,
"step": 10925
},
{
"epoch": 3.03,
"learning_rate": 1.3979310344827586e-05,
"loss": 0.1597,
"step": 10950
},
{
"epoch": 3.04,
"learning_rate": 1.3893103448275862e-05,
"loss": 0.1642,
"step": 10975
},
{
"epoch": 3.05,
"learning_rate": 1.3806896551724138e-05,
"loss": 0.1468,
"step": 11000
},
{
"epoch": 3.05,
"eval_cer": 10.796652398901168,
"eval_loss": 0.2832615077495575,
"eval_runtime": 1183.0471,
"eval_samples_per_second": 4.726,
"eval_steps_per_second": 0.296,
"step": 11000
},
{
"epoch": 3.05,
"learning_rate": 1.3720689655172414e-05,
"loss": 0.1427,
"step": 11025
},
{
"epoch": 3.06,
"learning_rate": 1.363448275862069e-05,
"loss": 0.17,
"step": 11050
},
{
"epoch": 3.07,
"learning_rate": 1.3548275862068965e-05,
"loss": 0.1585,
"step": 11075
},
{
"epoch": 3.08,
"learning_rate": 1.3462068965517241e-05,
"loss": 0.1727,
"step": 11100
},
{
"epoch": 3.08,
"learning_rate": 1.3375862068965517e-05,
"loss": 0.1654,
"step": 11125
},
{
"epoch": 3.09,
"learning_rate": 1.3289655172413793e-05,
"loss": 0.1517,
"step": 11150
},
{
"epoch": 3.1,
"learning_rate": 1.320344827586207e-05,
"loss": 0.1536,
"step": 11175
},
{
"epoch": 3.1,
"learning_rate": 1.3117241379310347e-05,
"loss": 0.1692,
"step": 11200
},
{
"epoch": 3.11,
"learning_rate": 1.3031034482758623e-05,
"loss": 0.1673,
"step": 11225
},
{
"epoch": 3.12,
"learning_rate": 1.2944827586206897e-05,
"loss": 0.1554,
"step": 11250
},
{
"epoch": 3.12,
"learning_rate": 1.2858620689655173e-05,
"loss": 0.1616,
"step": 11275
},
{
"epoch": 3.13,
"learning_rate": 1.277241379310345e-05,
"loss": 0.1553,
"step": 11300
},
{
"epoch": 3.14,
"learning_rate": 1.2686206896551725e-05,
"loss": 0.1692,
"step": 11325
},
{
"epoch": 3.14,
"learning_rate": 1.2600000000000001e-05,
"loss": 0.1582,
"step": 11350
},
{
"epoch": 3.15,
"learning_rate": 1.2513793103448276e-05,
"loss": 0.1634,
"step": 11375
},
{
"epoch": 3.16,
"learning_rate": 1.2427586206896552e-05,
"loss": 0.1584,
"step": 11400
},
{
"epoch": 3.17,
"learning_rate": 1.2341379310344828e-05,
"loss": 0.1678,
"step": 11425
},
{
"epoch": 3.17,
"learning_rate": 1.2255172413793104e-05,
"loss": 0.1767,
"step": 11450
},
{
"epoch": 3.18,
"learning_rate": 1.216896551724138e-05,
"loss": 0.1561,
"step": 11475
},
{
"epoch": 3.19,
"learning_rate": 1.2082758620689656e-05,
"loss": 0.1557,
"step": 11500
},
{
"epoch": 3.19,
"eval_cer": 10.533124640643967,
"eval_loss": 0.28545621037483215,
"eval_runtime": 1176.7108,
"eval_samples_per_second": 4.751,
"eval_steps_per_second": 0.297,
"step": 11500
},
{
"epoch": 3.19,
"learning_rate": 1.1996551724137932e-05,
"loss": 0.172,
"step": 11525
},
{
"epoch": 3.2,
"learning_rate": 1.1910344827586208e-05,
"loss": 0.1571,
"step": 11550
},
{
"epoch": 3.21,
"learning_rate": 1.1824137931034484e-05,
"loss": 0.1605,
"step": 11575
},
{
"epoch": 3.21,
"learning_rate": 1.1737931034482758e-05,
"loss": 0.1612,
"step": 11600
},
{
"epoch": 3.22,
"learning_rate": 1.1651724137931034e-05,
"loss": 0.1615,
"step": 11625
},
{
"epoch": 3.23,
"learning_rate": 1.156551724137931e-05,
"loss": 0.1627,
"step": 11650
},
{
"epoch": 3.23,
"learning_rate": 1.1479310344827588e-05,
"loss": 0.1481,
"step": 11675
},
{
"epoch": 3.24,
"learning_rate": 1.1393103448275863e-05,
"loss": 0.1532,
"step": 11700
},
{
"epoch": 3.25,
"learning_rate": 1.1306896551724139e-05,
"loss": 0.1358,
"step": 11725
},
{
"epoch": 3.26,
"learning_rate": 1.1220689655172415e-05,
"loss": 0.1619,
"step": 11750
},
{
"epoch": 3.26,
"learning_rate": 1.113448275862069e-05,
"loss": 0.1721,
"step": 11775
},
{
"epoch": 3.27,
"learning_rate": 1.1048275862068965e-05,
"loss": 0.1662,
"step": 11800
},
{
"epoch": 3.28,
"learning_rate": 1.0962068965517241e-05,
"loss": 0.1511,
"step": 11825
},
{
"epoch": 3.28,
"learning_rate": 1.0875862068965517e-05,
"loss": 0.1554,
"step": 11850
},
{
"epoch": 3.29,
"learning_rate": 1.0789655172413795e-05,
"loss": 0.1767,
"step": 11875
},
{
"epoch": 3.3,
"learning_rate": 1.070344827586207e-05,
"loss": 0.1613,
"step": 11900
},
{
"epoch": 3.3,
"learning_rate": 1.0617241379310345e-05,
"loss": 0.1752,
"step": 11925
},
{
"epoch": 3.31,
"learning_rate": 1.0531034482758621e-05,
"loss": 0.1493,
"step": 11950
},
{
"epoch": 3.32,
"learning_rate": 1.0444827586206897e-05,
"loss": 0.1927,
"step": 11975
},
{
"epoch": 3.32,
"learning_rate": 1.0358620689655173e-05,
"loss": 0.1639,
"step": 12000
},
{
"epoch": 3.32,
"eval_cer": 10.167380054941546,
"eval_loss": 0.28585749864578247,
"eval_runtime": 1168.3755,
"eval_samples_per_second": 4.785,
"eval_steps_per_second": 0.3,
"step": 12000
},
{
"epoch": 3.33,
"learning_rate": 1.0272413793103448e-05,
"loss": 0.1708,
"step": 12025
},
{
"epoch": 3.34,
"learning_rate": 1.0186206896551724e-05,
"loss": 0.1692,
"step": 12050
},
{
"epoch": 3.35,
"learning_rate": 1.0100000000000002e-05,
"loss": 0.1754,
"step": 12075
},
{
"epoch": 3.35,
"learning_rate": 1.0013793103448278e-05,
"loss": 0.1695,
"step": 12100
},
{
"epoch": 3.36,
"learning_rate": 9.927586206896552e-06,
"loss": 0.1695,
"step": 12125
},
{
"epoch": 3.37,
"learning_rate": 9.841379310344828e-06,
"loss": 0.1718,
"step": 12150
},
{
"epoch": 3.37,
"learning_rate": 9.755172413793104e-06,
"loss": 0.151,
"step": 12175
},
{
"epoch": 3.38,
"learning_rate": 9.66896551724138e-06,
"loss": 0.1672,
"step": 12200
},
{
"epoch": 3.39,
"learning_rate": 9.582758620689654e-06,
"loss": 0.1665,
"step": 12225
},
{
"epoch": 3.39,
"learning_rate": 9.49655172413793e-06,
"loss": 0.1567,
"step": 12250
},
{
"epoch": 3.4,
"learning_rate": 9.410344827586208e-06,
"loss": 0.1691,
"step": 12275
},
{
"epoch": 3.41,
"learning_rate": 9.324137931034484e-06,
"loss": 0.1648,
"step": 12300
},
{
"epoch": 3.42,
"learning_rate": 9.237931034482759e-06,
"loss": 0.1576,
"step": 12325
},
{
"epoch": 3.42,
"learning_rate": 9.151724137931035e-06,
"loss": 0.1689,
"step": 12350
},
{
"epoch": 3.43,
"learning_rate": 9.06551724137931e-06,
"loss": 0.159,
"step": 12375
},
{
"epoch": 3.44,
"learning_rate": 8.979310344827587e-06,
"loss": 0.1682,
"step": 12400
},
{
"epoch": 3.44,
"learning_rate": 8.893103448275863e-06,
"loss": 0.1729,
"step": 12425
},
{
"epoch": 3.45,
"learning_rate": 8.806896551724137e-06,
"loss": 0.168,
"step": 12450
},
{
"epoch": 3.46,
"learning_rate": 8.720689655172415e-06,
"loss": 0.1704,
"step": 12475
},
{
"epoch": 3.46,
"learning_rate": 8.634482758620691e-06,
"loss": 0.1603,
"step": 12500
},
{
"epoch": 3.46,
"eval_cer": 10.212099916948828,
"eval_loss": 0.2885717749595642,
"eval_runtime": 1168.0562,
"eval_samples_per_second": 4.787,
"eval_steps_per_second": 0.3,
"step": 12500
},
{
"epoch": 3.47,
"learning_rate": 8.548275862068967e-06,
"loss": 0.1564,
"step": 12525
},
{
"epoch": 3.48,
"learning_rate": 8.462068965517241e-06,
"loss": 0.1511,
"step": 12550
},
{
"epoch": 3.48,
"learning_rate": 8.375862068965517e-06,
"loss": 0.1575,
"step": 12575
},
{
"epoch": 3.49,
"learning_rate": 8.289655172413793e-06,
"loss": 0.1574,
"step": 12600
},
{
"epoch": 3.5,
"learning_rate": 8.20344827586207e-06,
"loss": 0.1673,
"step": 12625
},
{
"epoch": 3.51,
"learning_rate": 8.117241379310346e-06,
"loss": 0.1605,
"step": 12650
},
{
"epoch": 3.51,
"learning_rate": 8.031034482758622e-06,
"loss": 0.1562,
"step": 12675
},
{
"epoch": 3.52,
"learning_rate": 7.944827586206898e-06,
"loss": 0.1629,
"step": 12700
},
{
"epoch": 3.53,
"learning_rate": 7.858620689655174e-06,
"loss": 0.1629,
"step": 12725
},
{
"epoch": 3.53,
"learning_rate": 7.772413793103448e-06,
"loss": 0.1849,
"step": 12750
},
{
"epoch": 3.54,
"learning_rate": 7.686206896551724e-06,
"loss": 0.1501,
"step": 12775
},
{
"epoch": 3.55,
"learning_rate": 7.6e-06,
"loss": 0.1563,
"step": 12800
},
{
"epoch": 3.55,
"learning_rate": 7.513793103448276e-06,
"loss": 0.1596,
"step": 12825
},
{
"epoch": 3.56,
"learning_rate": 7.427586206896551e-06,
"loss": 0.156,
"step": 12850
},
{
"epoch": 3.57,
"learning_rate": 7.341379310344828e-06,
"loss": 0.1655,
"step": 12875
},
{
"epoch": 3.57,
"learning_rate": 7.255172413793104e-06,
"loss": 0.1642,
"step": 12900
},
{
"epoch": 3.58,
"learning_rate": 7.16896551724138e-06,
"loss": 0.1676,
"step": 12925
},
{
"epoch": 3.59,
"learning_rate": 7.0827586206896555e-06,
"loss": 0.149,
"step": 12950
},
{
"epoch": 3.6,
"learning_rate": 6.9965517241379315e-06,
"loss": 0.1734,
"step": 12975
},
{
"epoch": 3.6,
"learning_rate": 6.910344827586207e-06,
"loss": 0.1599,
"step": 13000
},
{
"epoch": 3.6,
"eval_cer": 10.058774675781,
"eval_loss": 0.2872118651866913,
"eval_runtime": 1162.7195,
"eval_samples_per_second": 4.809,
"eval_steps_per_second": 0.301,
"step": 13000
}
],
"max_steps": 15000,
"num_train_epochs": 5,
"total_flos": 3.7516448344080384e+20,
"trial_name": null,
"trial_params": null
}