ISOM5240-whisper-small-zhhk_1 / trainer_state.json
RexChan's picture
Upload trainer_state.json
e5fb948 verified
{
"best_metric": 10.695329087048833,
"best_model_checkpoint": "./model_out/checkpoint-10000",
"epoch": 5.5514433752775725,
"eval_steps": 500,
"global_step": 15000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 2.4000000000000003e-06,
"loss": 1.2092,
"step": 25
},
{
"epoch": 0.02,
"learning_rate": 4.9000000000000005e-06,
"loss": 0.8939,
"step": 50
},
{
"epoch": 0.03,
"learning_rate": 7.4e-06,
"loss": 0.7337,
"step": 75
},
{
"epoch": 0.04,
"learning_rate": 9.900000000000002e-06,
"loss": 0.6496,
"step": 100
},
{
"epoch": 0.05,
"learning_rate": 1.24e-05,
"loss": 0.5998,
"step": 125
},
{
"epoch": 0.06,
"learning_rate": 1.49e-05,
"loss": 0.5262,
"step": 150
},
{
"epoch": 0.06,
"learning_rate": 1.74e-05,
"loss": 0.501,
"step": 175
},
{
"epoch": 0.07,
"learning_rate": 1.9900000000000003e-05,
"loss": 0.4801,
"step": 200
},
{
"epoch": 0.08,
"learning_rate": 2.2400000000000002e-05,
"loss": 0.4495,
"step": 225
},
{
"epoch": 0.09,
"learning_rate": 2.4900000000000002e-05,
"loss": 0.4372,
"step": 250
},
{
"epoch": 0.1,
"learning_rate": 2.7300000000000003e-05,
"loss": 0.425,
"step": 275
},
{
"epoch": 0.11,
"learning_rate": 2.98e-05,
"loss": 0.4097,
"step": 300
},
{
"epoch": 0.12,
"learning_rate": 3.2300000000000006e-05,
"loss": 0.3997,
"step": 325
},
{
"epoch": 0.13,
"learning_rate": 3.48e-05,
"loss": 0.3828,
"step": 350
},
{
"epoch": 0.14,
"learning_rate": 3.73e-05,
"loss": 0.3817,
"step": 375
},
{
"epoch": 0.15,
"learning_rate": 3.9800000000000005e-05,
"loss": 0.3696,
"step": 400
},
{
"epoch": 0.16,
"learning_rate": 4.23e-05,
"loss": 0.3808,
"step": 425
},
{
"epoch": 0.17,
"learning_rate": 4.4800000000000005e-05,
"loss": 0.3764,
"step": 450
},
{
"epoch": 0.18,
"learning_rate": 4.73e-05,
"loss": 0.3712,
"step": 475
},
{
"epoch": 0.19,
"learning_rate": 4.9800000000000004e-05,
"loss": 0.3416,
"step": 500
},
{
"epoch": 0.19,
"eval_cer": 13.553988474370641,
"eval_loss": 0.2625325620174408,
"eval_runtime": 367.6787,
"eval_samples_per_second": 7.02,
"eval_steps_per_second": 0.878,
"step": 500
},
{
"epoch": 0.19,
"learning_rate": 4.992068965517242e-05,
"loss": 0.3682,
"step": 525
},
{
"epoch": 0.2,
"learning_rate": 4.9834482758620695e-05,
"loss": 0.3716,
"step": 550
},
{
"epoch": 0.21,
"learning_rate": 4.9748275862068966e-05,
"loss": 0.3581,
"step": 575
},
{
"epoch": 0.22,
"learning_rate": 4.966206896551724e-05,
"loss": 0.3542,
"step": 600
},
{
"epoch": 0.23,
"learning_rate": 4.9575862068965514e-05,
"loss": 0.347,
"step": 625
},
{
"epoch": 0.24,
"learning_rate": 4.94896551724138e-05,
"loss": 0.3603,
"step": 650
},
{
"epoch": 0.25,
"learning_rate": 4.940344827586207e-05,
"loss": 0.3568,
"step": 675
},
{
"epoch": 0.26,
"learning_rate": 4.931724137931035e-05,
"loss": 0.3408,
"step": 700
},
{
"epoch": 0.27,
"learning_rate": 4.9231034482758625e-05,
"loss": 0.346,
"step": 725
},
{
"epoch": 0.28,
"learning_rate": 4.9144827586206896e-05,
"loss": 0.3382,
"step": 750
},
{
"epoch": 0.29,
"learning_rate": 4.9058620689655174e-05,
"loss": 0.3422,
"step": 775
},
{
"epoch": 0.3,
"learning_rate": 4.897241379310345e-05,
"loss": 0.3423,
"step": 800
},
{
"epoch": 0.31,
"learning_rate": 4.888620689655173e-05,
"loss": 0.3245,
"step": 825
},
{
"epoch": 0.31,
"learning_rate": 4.88e-05,
"loss": 0.3378,
"step": 850
},
{
"epoch": 0.32,
"learning_rate": 4.871379310344828e-05,
"loss": 0.3388,
"step": 875
},
{
"epoch": 0.33,
"learning_rate": 4.8627586206896556e-05,
"loss": 0.3189,
"step": 900
},
{
"epoch": 0.34,
"learning_rate": 4.8541379310344834e-05,
"loss": 0.3264,
"step": 925
},
{
"epoch": 0.35,
"learning_rate": 4.8455172413793105e-05,
"loss": 0.3231,
"step": 950
},
{
"epoch": 0.36,
"learning_rate": 4.836896551724138e-05,
"loss": 0.3046,
"step": 975
},
{
"epoch": 0.37,
"learning_rate": 4.828275862068965e-05,
"loss": 0.3115,
"step": 1000
},
{
"epoch": 0.37,
"eval_cer": 12.473460721868365,
"eval_loss": 0.2618062496185303,
"eval_runtime": 363.8751,
"eval_samples_per_second": 7.093,
"eval_steps_per_second": 0.888,
"step": 1000
},
{
"epoch": 0.38,
"learning_rate": 4.819655172413793e-05,
"loss": 0.3251,
"step": 1025
},
{
"epoch": 0.39,
"learning_rate": 4.811034482758621e-05,
"loss": 0.3158,
"step": 1050
},
{
"epoch": 0.4,
"learning_rate": 4.8024137931034486e-05,
"loss": 0.3093,
"step": 1075
},
{
"epoch": 0.41,
"learning_rate": 4.7937931034482764e-05,
"loss": 0.3238,
"step": 1100
},
{
"epoch": 0.42,
"learning_rate": 4.7851724137931035e-05,
"loss": 0.3128,
"step": 1125
},
{
"epoch": 0.43,
"learning_rate": 4.776551724137931e-05,
"loss": 0.3012,
"step": 1150
},
{
"epoch": 0.43,
"learning_rate": 4.767931034482759e-05,
"loss": 0.3103,
"step": 1175
},
{
"epoch": 0.44,
"learning_rate": 4.759310344827587e-05,
"loss": 0.314,
"step": 1200
},
{
"epoch": 0.45,
"learning_rate": 4.750689655172414e-05,
"loss": 0.325,
"step": 1225
},
{
"epoch": 0.46,
"learning_rate": 4.742068965517242e-05,
"loss": 0.2923,
"step": 1250
},
{
"epoch": 0.47,
"learning_rate": 4.733448275862069e-05,
"loss": 0.3047,
"step": 1275
},
{
"epoch": 0.48,
"learning_rate": 4.7251724137931034e-05,
"loss": 0.3007,
"step": 1300
},
{
"epoch": 0.49,
"learning_rate": 4.716551724137931e-05,
"loss": 0.3233,
"step": 1325
},
{
"epoch": 0.5,
"learning_rate": 4.707931034482759e-05,
"loss": 0.2983,
"step": 1350
},
{
"epoch": 0.51,
"learning_rate": 4.699310344827587e-05,
"loss": 0.2961,
"step": 1375
},
{
"epoch": 0.52,
"learning_rate": 4.690689655172414e-05,
"loss": 0.2929,
"step": 1400
},
{
"epoch": 0.53,
"learning_rate": 4.6820689655172416e-05,
"loss": 0.2839,
"step": 1425
},
{
"epoch": 0.54,
"learning_rate": 4.673448275862069e-05,
"loss": 0.2973,
"step": 1450
},
{
"epoch": 0.55,
"learning_rate": 4.664827586206897e-05,
"loss": 0.2896,
"step": 1475
},
{
"epoch": 0.56,
"learning_rate": 4.656206896551724e-05,
"loss": 0.289,
"step": 1500
},
{
"epoch": 0.56,
"eval_cer": 11.809978768577494,
"eval_loss": 0.2488066703081131,
"eval_runtime": 362.9261,
"eval_samples_per_second": 7.112,
"eval_steps_per_second": 0.89,
"step": 1500
},
{
"epoch": 0.56,
"learning_rate": 4.647586206896552e-05,
"loss": 0.2872,
"step": 1525
},
{
"epoch": 0.57,
"learning_rate": 4.638965517241379e-05,
"loss": 0.2965,
"step": 1550
},
{
"epoch": 0.58,
"learning_rate": 4.630344827586207e-05,
"loss": 0.3036,
"step": 1575
},
{
"epoch": 0.59,
"learning_rate": 4.621724137931035e-05,
"loss": 0.2891,
"step": 1600
},
{
"epoch": 0.6,
"learning_rate": 4.6131034482758624e-05,
"loss": 0.2896,
"step": 1625
},
{
"epoch": 0.61,
"learning_rate": 4.60448275862069e-05,
"loss": 0.2992,
"step": 1650
},
{
"epoch": 0.62,
"learning_rate": 4.595862068965517e-05,
"loss": 0.2865,
"step": 1675
},
{
"epoch": 0.63,
"learning_rate": 4.587241379310345e-05,
"loss": 0.3007,
"step": 1700
},
{
"epoch": 0.64,
"learning_rate": 4.578620689655173e-05,
"loss": 0.3108,
"step": 1725
},
{
"epoch": 0.65,
"learning_rate": 4.5700000000000006e-05,
"loss": 0.2861,
"step": 1750
},
{
"epoch": 0.66,
"learning_rate": 4.561379310344828e-05,
"loss": 0.2996,
"step": 1775
},
{
"epoch": 0.67,
"learning_rate": 4.5527586206896555e-05,
"loss": 0.295,
"step": 1800
},
{
"epoch": 0.68,
"learning_rate": 4.5441379310344826e-05,
"loss": 0.2848,
"step": 1825
},
{
"epoch": 0.68,
"learning_rate": 4.5355172413793104e-05,
"loss": 0.2997,
"step": 1850
},
{
"epoch": 0.69,
"learning_rate": 4.526896551724138e-05,
"loss": 0.2828,
"step": 1875
},
{
"epoch": 0.7,
"learning_rate": 4.518620689655173e-05,
"loss": 0.2889,
"step": 1900
},
{
"epoch": 0.71,
"learning_rate": 4.5100000000000005e-05,
"loss": 0.2893,
"step": 1925
},
{
"epoch": 0.72,
"learning_rate": 4.5013793103448276e-05,
"loss": 0.276,
"step": 1950
},
{
"epoch": 0.73,
"learning_rate": 4.4927586206896554e-05,
"loss": 0.2784,
"step": 1975
},
{
"epoch": 0.74,
"learning_rate": 4.4841379310344825e-05,
"loss": 0.2826,
"step": 2000
},
{
"epoch": 0.74,
"eval_cer": 13.197603882317258,
"eval_loss": 0.2393285036087036,
"eval_runtime": 368.8043,
"eval_samples_per_second": 6.998,
"eval_steps_per_second": 0.876,
"step": 2000
},
{
"epoch": 0.75,
"learning_rate": 4.475517241379311e-05,
"loss": 0.2707,
"step": 2025
},
{
"epoch": 0.76,
"learning_rate": 4.466896551724138e-05,
"loss": 0.2824,
"step": 2050
},
{
"epoch": 0.77,
"learning_rate": 4.458275862068966e-05,
"loss": 0.296,
"step": 2075
},
{
"epoch": 0.78,
"learning_rate": 4.4496551724137936e-05,
"loss": 0.2786,
"step": 2100
},
{
"epoch": 0.79,
"learning_rate": 4.441034482758621e-05,
"loss": 0.2802,
"step": 2125
},
{
"epoch": 0.8,
"learning_rate": 4.4324137931034485e-05,
"loss": 0.2791,
"step": 2150
},
{
"epoch": 0.8,
"learning_rate": 4.423793103448276e-05,
"loss": 0.2953,
"step": 2175
},
{
"epoch": 0.81,
"learning_rate": 4.415172413793104e-05,
"loss": 0.2807,
"step": 2200
},
{
"epoch": 0.82,
"learning_rate": 4.406551724137931e-05,
"loss": 0.2776,
"step": 2225
},
{
"epoch": 0.83,
"learning_rate": 4.397931034482759e-05,
"loss": 0.2798,
"step": 2250
},
{
"epoch": 0.84,
"learning_rate": 4.389310344827586e-05,
"loss": 0.2804,
"step": 2275
},
{
"epoch": 0.85,
"learning_rate": 4.3806896551724144e-05,
"loss": 0.2892,
"step": 2300
},
{
"epoch": 0.86,
"learning_rate": 4.3720689655172415e-05,
"loss": 0.2773,
"step": 2325
},
{
"epoch": 0.87,
"learning_rate": 4.363448275862069e-05,
"loss": 0.2733,
"step": 2350
},
{
"epoch": 0.88,
"learning_rate": 4.3548275862068964e-05,
"loss": 0.2805,
"step": 2375
},
{
"epoch": 0.89,
"learning_rate": 4.346206896551724e-05,
"loss": 0.274,
"step": 2400
},
{
"epoch": 0.9,
"learning_rate": 4.337586206896552e-05,
"loss": 0.2711,
"step": 2425
},
{
"epoch": 0.91,
"learning_rate": 4.32896551724138e-05,
"loss": 0.2684,
"step": 2450
},
{
"epoch": 0.92,
"learning_rate": 4.320344827586207e-05,
"loss": 0.2731,
"step": 2475
},
{
"epoch": 0.93,
"learning_rate": 4.3117241379310346e-05,
"loss": 0.2689,
"step": 2500
},
{
"epoch": 0.93,
"eval_cer": 13.10282074613285,
"eval_loss": 0.24490514397621155,
"eval_runtime": 366.5991,
"eval_samples_per_second": 7.04,
"eval_steps_per_second": 0.881,
"step": 2500
},
{
"epoch": 0.93,
"learning_rate": 4.3031034482758624e-05,
"loss": 0.2702,
"step": 2525
},
{
"epoch": 0.94,
"learning_rate": 4.29448275862069e-05,
"loss": 0.2623,
"step": 2550
},
{
"epoch": 0.95,
"learning_rate": 4.285862068965518e-05,
"loss": 0.2749,
"step": 2575
},
{
"epoch": 0.96,
"learning_rate": 4.277241379310345e-05,
"loss": 0.2658,
"step": 2600
},
{
"epoch": 0.97,
"learning_rate": 4.268620689655173e-05,
"loss": 0.2637,
"step": 2625
},
{
"epoch": 0.98,
"learning_rate": 4.26e-05,
"loss": 0.2823,
"step": 2650
},
{
"epoch": 0.99,
"learning_rate": 4.2513793103448276e-05,
"loss": 0.2641,
"step": 2675
},
{
"epoch": 1.0,
"learning_rate": 4.2427586206896554e-05,
"loss": 0.2721,
"step": 2700
},
{
"epoch": 1.01,
"learning_rate": 4.234137931034483e-05,
"loss": 0.1882,
"step": 2725
},
{
"epoch": 1.02,
"learning_rate": 4.22551724137931e-05,
"loss": 0.1776,
"step": 2750
},
{
"epoch": 1.03,
"learning_rate": 4.216896551724138e-05,
"loss": 0.1768,
"step": 2775
},
{
"epoch": 1.04,
"learning_rate": 4.208275862068966e-05,
"loss": 0.1714,
"step": 2800
},
{
"epoch": 1.05,
"learning_rate": 4.1996551724137936e-05,
"loss": 0.1771,
"step": 2825
},
{
"epoch": 1.05,
"learning_rate": 4.191034482758621e-05,
"loss": 0.193,
"step": 2850
},
{
"epoch": 1.06,
"learning_rate": 4.1824137931034485e-05,
"loss": 0.1788,
"step": 2875
},
{
"epoch": 1.07,
"learning_rate": 4.1737931034482756e-05,
"loss": 0.1784,
"step": 2900
},
{
"epoch": 1.08,
"learning_rate": 4.1651724137931034e-05,
"loss": 0.1879,
"step": 2925
},
{
"epoch": 1.09,
"learning_rate": 4.156551724137932e-05,
"loss": 0.1825,
"step": 2950
},
{
"epoch": 1.1,
"learning_rate": 4.147931034482759e-05,
"loss": 0.1854,
"step": 2975
},
{
"epoch": 1.11,
"learning_rate": 4.139310344827587e-05,
"loss": 0.1882,
"step": 3000
},
{
"epoch": 1.11,
"eval_cer": 11.67728237791932,
"eval_loss": 0.25736209750175476,
"eval_runtime": 361.0573,
"eval_samples_per_second": 7.148,
"eval_steps_per_second": 0.895,
"step": 3000
},
{
"epoch": 1.12,
"learning_rate": 4.130689655172414e-05,
"loss": 0.1827,
"step": 3025
},
{
"epoch": 1.13,
"learning_rate": 4.1220689655172415e-05,
"loss": 0.1889,
"step": 3050
},
{
"epoch": 1.14,
"learning_rate": 4.113448275862069e-05,
"loss": 0.1938,
"step": 3075
},
{
"epoch": 1.15,
"learning_rate": 4.104827586206897e-05,
"loss": 0.182,
"step": 3100
},
{
"epoch": 1.16,
"learning_rate": 4.096206896551724e-05,
"loss": 0.1884,
"step": 3125
},
{
"epoch": 1.17,
"learning_rate": 4.087586206896552e-05,
"loss": 0.182,
"step": 3150
},
{
"epoch": 1.18,
"learning_rate": 4.078965517241379e-05,
"loss": 0.1795,
"step": 3175
},
{
"epoch": 1.18,
"learning_rate": 4.070344827586207e-05,
"loss": 0.176,
"step": 3200
},
{
"epoch": 1.19,
"learning_rate": 4.0617241379310346e-05,
"loss": 0.1834,
"step": 3225
},
{
"epoch": 1.2,
"learning_rate": 4.0531034482758624e-05,
"loss": 0.1886,
"step": 3250
},
{
"epoch": 1.21,
"learning_rate": 4.0444827586206895e-05,
"loss": 0.1908,
"step": 3275
},
{
"epoch": 1.22,
"learning_rate": 4.035862068965517e-05,
"loss": 0.1813,
"step": 3300
},
{
"epoch": 1.23,
"learning_rate": 4.027241379310345e-05,
"loss": 0.1875,
"step": 3325
},
{
"epoch": 1.24,
"learning_rate": 4.018620689655173e-05,
"loss": 0.1925,
"step": 3350
},
{
"epoch": 1.25,
"learning_rate": 4.0100000000000006e-05,
"loss": 0.1855,
"step": 3375
},
{
"epoch": 1.26,
"learning_rate": 4.001379310344828e-05,
"loss": 0.194,
"step": 3400
},
{
"epoch": 1.27,
"learning_rate": 3.9927586206896554e-05,
"loss": 0.1827,
"step": 3425
},
{
"epoch": 1.28,
"learning_rate": 3.9841379310344825e-05,
"loss": 0.1806,
"step": 3450
},
{
"epoch": 1.29,
"learning_rate": 3.975517241379311e-05,
"loss": 0.1783,
"step": 3475
},
{
"epoch": 1.3,
"learning_rate": 3.966896551724138e-05,
"loss": 0.1889,
"step": 3500
},
{
"epoch": 1.3,
"eval_cer": 11.434637549287231,
"eval_loss": 0.24476666748523712,
"eval_runtime": 361.1188,
"eval_samples_per_second": 7.147,
"eval_steps_per_second": 0.894,
"step": 3500
},
{
"epoch": 1.3,
"learning_rate": 3.958275862068966e-05,
"loss": 0.185,
"step": 3525
},
{
"epoch": 1.31,
"learning_rate": 3.949655172413793e-05,
"loss": 0.1833,
"step": 3550
},
{
"epoch": 1.32,
"learning_rate": 3.941034482758621e-05,
"loss": 0.1749,
"step": 3575
},
{
"epoch": 1.33,
"learning_rate": 3.9324137931034485e-05,
"loss": 0.1906,
"step": 3600
},
{
"epoch": 1.34,
"learning_rate": 3.923793103448276e-05,
"loss": 0.1934,
"step": 3625
},
{
"epoch": 1.35,
"learning_rate": 3.9151724137931034e-05,
"loss": 0.1782,
"step": 3650
},
{
"epoch": 1.36,
"learning_rate": 3.906551724137931e-05,
"loss": 0.19,
"step": 3675
},
{
"epoch": 1.37,
"learning_rate": 3.897931034482758e-05,
"loss": 0.1943,
"step": 3700
},
{
"epoch": 1.38,
"learning_rate": 3.889310344827587e-05,
"loss": 0.1838,
"step": 3725
},
{
"epoch": 1.39,
"learning_rate": 3.8806896551724145e-05,
"loss": 0.2248,
"step": 3750
},
{
"epoch": 1.4,
"learning_rate": 3.8720689655172416e-05,
"loss": 0.1853,
"step": 3775
},
{
"epoch": 1.41,
"learning_rate": 3.863448275862069e-05,
"loss": 0.1867,
"step": 3800
},
{
"epoch": 1.42,
"learning_rate": 3.8548275862068964e-05,
"loss": 0.1801,
"step": 3825
},
{
"epoch": 1.42,
"learning_rate": 3.846206896551724e-05,
"loss": 0.182,
"step": 3850
},
{
"epoch": 1.43,
"learning_rate": 3.837586206896552e-05,
"loss": 0.1843,
"step": 3875
},
{
"epoch": 1.44,
"learning_rate": 3.82896551724138e-05,
"loss": 0.184,
"step": 3900
},
{
"epoch": 1.45,
"learning_rate": 3.820344827586207e-05,
"loss": 0.1808,
"step": 3925
},
{
"epoch": 1.46,
"learning_rate": 3.8117241379310346e-05,
"loss": 0.188,
"step": 3950
},
{
"epoch": 1.47,
"learning_rate": 3.8031034482758624e-05,
"loss": 0.1691,
"step": 3975
},
{
"epoch": 1.48,
"learning_rate": 3.79448275862069e-05,
"loss": 0.1923,
"step": 4000
},
{
"epoch": 1.48,
"eval_cer": 11.8895966029724,
"eval_loss": 0.24924282729625702,
"eval_runtime": 360.2098,
"eval_samples_per_second": 7.165,
"eval_steps_per_second": 0.897,
"step": 4000
},
{
"epoch": 1.49,
"learning_rate": 3.785862068965517e-05,
"loss": 0.175,
"step": 4025
},
{
"epoch": 1.5,
"learning_rate": 3.777241379310345e-05,
"loss": 0.1814,
"step": 4050
},
{
"epoch": 1.51,
"learning_rate": 3.768620689655172e-05,
"loss": 0.1808,
"step": 4075
},
{
"epoch": 1.52,
"learning_rate": 3.76e-05,
"loss": 0.1875,
"step": 4100
},
{
"epoch": 1.53,
"learning_rate": 3.7513793103448284e-05,
"loss": 0.1831,
"step": 4125
},
{
"epoch": 1.54,
"learning_rate": 3.7427586206896555e-05,
"loss": 0.184,
"step": 4150
},
{
"epoch": 1.55,
"learning_rate": 3.734137931034483e-05,
"loss": 0.1774,
"step": 4175
},
{
"epoch": 1.55,
"learning_rate": 3.72551724137931e-05,
"loss": 0.1716,
"step": 4200
},
{
"epoch": 1.56,
"learning_rate": 3.716896551724138e-05,
"loss": 0.1845,
"step": 4225
},
{
"epoch": 1.57,
"learning_rate": 3.708275862068966e-05,
"loss": 0.1854,
"step": 4250
},
{
"epoch": 1.58,
"learning_rate": 3.6996551724137937e-05,
"loss": 0.1824,
"step": 4275
},
{
"epoch": 1.59,
"learning_rate": 3.691034482758621e-05,
"loss": 0.1796,
"step": 4300
},
{
"epoch": 1.6,
"learning_rate": 3.6824137931034485e-05,
"loss": 0.1882,
"step": 4325
},
{
"epoch": 1.61,
"learning_rate": 3.6737931034482756e-05,
"loss": 0.1866,
"step": 4350
},
{
"epoch": 1.62,
"learning_rate": 3.6651724137931034e-05,
"loss": 0.1881,
"step": 4375
},
{
"epoch": 1.63,
"learning_rate": 3.656551724137931e-05,
"loss": 0.1822,
"step": 4400
},
{
"epoch": 1.64,
"learning_rate": 3.647931034482759e-05,
"loss": 0.1803,
"step": 4425
},
{
"epoch": 1.65,
"learning_rate": 3.639310344827586e-05,
"loss": 0.1851,
"step": 4450
},
{
"epoch": 1.66,
"learning_rate": 3.630689655172414e-05,
"loss": 0.1825,
"step": 4475
},
{
"epoch": 1.67,
"learning_rate": 3.6220689655172416e-05,
"loss": 0.1886,
"step": 4500
},
{
"epoch": 1.67,
"eval_cer": 11.75310888686685,
"eval_loss": 0.23957248032093048,
"eval_runtime": 361.2575,
"eval_samples_per_second": 7.144,
"eval_steps_per_second": 0.894,
"step": 4500
},
{
"epoch": 1.67,
"learning_rate": 3.6134482758620694e-05,
"loss": 0.1944,
"step": 4525
},
{
"epoch": 1.68,
"learning_rate": 3.604827586206897e-05,
"loss": 0.1896,
"step": 4550
},
{
"epoch": 1.69,
"learning_rate": 3.596206896551724e-05,
"loss": 0.1819,
"step": 4575
},
{
"epoch": 1.7,
"learning_rate": 3.587586206896552e-05,
"loss": 0.1892,
"step": 4600
},
{
"epoch": 1.71,
"learning_rate": 3.578965517241379e-05,
"loss": 0.1914,
"step": 4625
},
{
"epoch": 1.72,
"learning_rate": 3.5703448275862076e-05,
"loss": 0.181,
"step": 4650
},
{
"epoch": 1.73,
"learning_rate": 3.5617241379310346e-05,
"loss": 0.1799,
"step": 4675
},
{
"epoch": 1.74,
"learning_rate": 3.5531034482758624e-05,
"loss": 0.1784,
"step": 4700
},
{
"epoch": 1.75,
"learning_rate": 3.5444827586206895e-05,
"loss": 0.1812,
"step": 4725
},
{
"epoch": 1.76,
"learning_rate": 3.535862068965517e-05,
"loss": 0.1782,
"step": 4750
},
{
"epoch": 1.77,
"learning_rate": 3.527241379310345e-05,
"loss": 0.1783,
"step": 4775
},
{
"epoch": 1.78,
"learning_rate": 3.518620689655173e-05,
"loss": 0.2002,
"step": 4800
},
{
"epoch": 1.79,
"learning_rate": 3.51e-05,
"loss": 0.1948,
"step": 4825
},
{
"epoch": 1.79,
"learning_rate": 3.501379310344828e-05,
"loss": 0.1791,
"step": 4850
},
{
"epoch": 1.8,
"learning_rate": 3.492758620689655e-05,
"loss": 0.2018,
"step": 4875
},
{
"epoch": 1.81,
"learning_rate": 3.484137931034483e-05,
"loss": 0.1828,
"step": 4900
},
{
"epoch": 1.82,
"learning_rate": 3.475517241379311e-05,
"loss": 0.1834,
"step": 4925
},
{
"epoch": 1.83,
"learning_rate": 3.466896551724138e-05,
"loss": 0.1821,
"step": 4950
},
{
"epoch": 1.84,
"learning_rate": 3.458275862068966e-05,
"loss": 0.1864,
"step": 4975
},
{
"epoch": 1.85,
"learning_rate": 3.449655172413793e-05,
"loss": 0.1899,
"step": 5000
},
{
"epoch": 1.85,
"eval_cer": 12.048832271762208,
"eval_loss": 0.23497651517391205,
"eval_runtime": 363.6465,
"eval_samples_per_second": 7.098,
"eval_steps_per_second": 0.888,
"step": 5000
},
{
"epoch": 1.86,
"learning_rate": 3.441034482758621e-05,
"loss": 0.1805,
"step": 5025
},
{
"epoch": 1.87,
"learning_rate": 3.4324137931034485e-05,
"loss": 0.1826,
"step": 5050
},
{
"epoch": 1.88,
"learning_rate": 3.423793103448276e-05,
"loss": 0.1863,
"step": 5075
},
{
"epoch": 1.89,
"learning_rate": 3.4151724137931034e-05,
"loss": 0.1754,
"step": 5100
},
{
"epoch": 1.9,
"learning_rate": 3.406551724137931e-05,
"loss": 0.1801,
"step": 5125
},
{
"epoch": 1.91,
"learning_rate": 3.397931034482758e-05,
"loss": 0.1846,
"step": 5150
},
{
"epoch": 1.92,
"learning_rate": 3.389310344827587e-05,
"loss": 0.1717,
"step": 5175
},
{
"epoch": 1.92,
"learning_rate": 3.380689655172414e-05,
"loss": 0.1754,
"step": 5200
},
{
"epoch": 1.93,
"learning_rate": 3.3720689655172416e-05,
"loss": 0.1819,
"step": 5225
},
{
"epoch": 1.94,
"learning_rate": 3.363448275862069e-05,
"loss": 0.1792,
"step": 5250
},
{
"epoch": 1.95,
"learning_rate": 3.3548275862068965e-05,
"loss": 0.1837,
"step": 5275
},
{
"epoch": 1.96,
"learning_rate": 3.346206896551724e-05,
"loss": 0.188,
"step": 5300
},
{
"epoch": 1.97,
"learning_rate": 3.337586206896552e-05,
"loss": 0.1875,
"step": 5325
},
{
"epoch": 1.98,
"learning_rate": 3.32896551724138e-05,
"loss": 0.1909,
"step": 5350
},
{
"epoch": 1.99,
"learning_rate": 3.320344827586207e-05,
"loss": 0.1714,
"step": 5375
},
{
"epoch": 2.0,
"learning_rate": 3.311724137931035e-05,
"loss": 0.1921,
"step": 5400
},
{
"epoch": 2.01,
"learning_rate": 3.3031034482758624e-05,
"loss": 0.1109,
"step": 5425
},
{
"epoch": 2.02,
"learning_rate": 3.29448275862069e-05,
"loss": 0.0946,
"step": 5450
},
{
"epoch": 2.03,
"learning_rate": 3.285862068965517e-05,
"loss": 0.0967,
"step": 5475
},
{
"epoch": 2.04,
"learning_rate": 3.277241379310345e-05,
"loss": 0.1009,
"step": 5500
},
{
"epoch": 2.04,
"eval_cer": 11.404306945708221,
"eval_loss": 0.24928008019924164,
"eval_runtime": 359.254,
"eval_samples_per_second": 7.184,
"eval_steps_per_second": 0.899,
"step": 5500
},
{
"epoch": 2.04,
"learning_rate": 3.268620689655172e-05,
"loss": 0.0954,
"step": 5525
},
{
"epoch": 2.05,
"learning_rate": 3.26e-05,
"loss": 0.0999,
"step": 5550
},
{
"epoch": 2.06,
"learning_rate": 3.251379310344828e-05,
"loss": 0.0946,
"step": 5575
},
{
"epoch": 2.07,
"learning_rate": 3.2427586206896555e-05,
"loss": 0.0949,
"step": 5600
},
{
"epoch": 2.08,
"learning_rate": 3.2341379310344826e-05,
"loss": 0.1032,
"step": 5625
},
{
"epoch": 2.09,
"learning_rate": 3.2255172413793104e-05,
"loss": 0.1002,
"step": 5650
},
{
"epoch": 2.1,
"learning_rate": 3.216896551724138e-05,
"loss": 0.0973,
"step": 5675
},
{
"epoch": 2.11,
"learning_rate": 3.208275862068966e-05,
"loss": 0.0984,
"step": 5700
},
{
"epoch": 2.12,
"learning_rate": 3.199655172413794e-05,
"loss": 0.1031,
"step": 5725
},
{
"epoch": 2.13,
"learning_rate": 3.191034482758621e-05,
"loss": 0.0968,
"step": 5750
},
{
"epoch": 2.14,
"learning_rate": 3.1824137931034486e-05,
"loss": 0.0983,
"step": 5775
},
{
"epoch": 2.15,
"learning_rate": 3.173793103448276e-05,
"loss": 0.1088,
"step": 5800
},
{
"epoch": 2.16,
"learning_rate": 3.165172413793104e-05,
"loss": 0.0998,
"step": 5825
},
{
"epoch": 2.17,
"learning_rate": 3.156551724137931e-05,
"loss": 0.0991,
"step": 5850
},
{
"epoch": 2.17,
"learning_rate": 3.147931034482759e-05,
"loss": 0.103,
"step": 5875
},
{
"epoch": 2.18,
"learning_rate": 3.139310344827586e-05,
"loss": 0.1046,
"step": 5900
},
{
"epoch": 2.19,
"learning_rate": 3.130689655172414e-05,
"loss": 0.099,
"step": 5925
},
{
"epoch": 2.2,
"learning_rate": 3.1220689655172416e-05,
"loss": 0.1002,
"step": 5950
},
{
"epoch": 2.21,
"learning_rate": 3.1134482758620694e-05,
"loss": 0.1028,
"step": 5975
},
{
"epoch": 2.22,
"learning_rate": 3.1048275862068965e-05,
"loss": 0.1003,
"step": 6000
},
{
"epoch": 2.22,
"eval_cer": 11.373976342129207,
"eval_loss": 0.2572503387928009,
"eval_runtime": 359.7119,
"eval_samples_per_second": 7.175,
"eval_steps_per_second": 0.898,
"step": 6000
},
{
"epoch": 2.23,
"learning_rate": 3.096206896551724e-05,
"loss": 0.0954,
"step": 6025
},
{
"epoch": 2.24,
"learning_rate": 3.0875862068965514e-05,
"loss": 0.0981,
"step": 6050
},
{
"epoch": 2.25,
"learning_rate": 3.07896551724138e-05,
"loss": 0.1006,
"step": 6075
},
{
"epoch": 2.26,
"learning_rate": 3.070344827586207e-05,
"loss": 0.1005,
"step": 6100
},
{
"epoch": 2.27,
"learning_rate": 3.061724137931035e-05,
"loss": 0.1007,
"step": 6125
},
{
"epoch": 2.28,
"learning_rate": 3.0531034482758625e-05,
"loss": 0.1005,
"step": 6150
},
{
"epoch": 2.29,
"learning_rate": 3.04448275862069e-05,
"loss": 0.0968,
"step": 6175
},
{
"epoch": 2.29,
"learning_rate": 3.0358620689655177e-05,
"loss": 0.1034,
"step": 6200
},
{
"epoch": 2.3,
"learning_rate": 3.0272413793103448e-05,
"loss": 0.0982,
"step": 6225
},
{
"epoch": 2.31,
"learning_rate": 3.018620689655173e-05,
"loss": 0.1033,
"step": 6250
},
{
"epoch": 2.32,
"learning_rate": 3.01e-05,
"loss": 0.1031,
"step": 6275
},
{
"epoch": 2.33,
"learning_rate": 3.0013793103448278e-05,
"loss": 0.0985,
"step": 6300
},
{
"epoch": 2.34,
"learning_rate": 2.9931034482758624e-05,
"loss": 0.1405,
"step": 6325
},
{
"epoch": 2.35,
"learning_rate": 2.9844827586206898e-05,
"loss": 0.1059,
"step": 6350
},
{
"epoch": 2.36,
"learning_rate": 2.9758620689655176e-05,
"loss": 0.098,
"step": 6375
},
{
"epoch": 2.37,
"learning_rate": 2.967241379310345e-05,
"loss": 0.1029,
"step": 6400
},
{
"epoch": 2.38,
"learning_rate": 2.9586206896551728e-05,
"loss": 0.1082,
"step": 6425
},
{
"epoch": 2.39,
"learning_rate": 2.95e-05,
"loss": 0.1049,
"step": 6450
},
{
"epoch": 2.4,
"learning_rate": 2.941379310344828e-05,
"loss": 0.0996,
"step": 6475
},
{
"epoch": 2.41,
"learning_rate": 2.932758620689655e-05,
"loss": 0.1031,
"step": 6500
},
{
"epoch": 2.41,
"eval_cer": 12.954959053685169,
"eval_loss": 0.26603972911834717,
"eval_runtime": 369.9518,
"eval_samples_per_second": 6.977,
"eval_steps_per_second": 0.873,
"step": 6500
},
{
"epoch": 2.41,
"learning_rate": 2.924137931034483e-05,
"loss": 0.1062,
"step": 6525
},
{
"epoch": 2.42,
"learning_rate": 2.9155172413793103e-05,
"loss": 0.1057,
"step": 6550
},
{
"epoch": 2.43,
"learning_rate": 2.906896551724138e-05,
"loss": 0.1063,
"step": 6575
},
{
"epoch": 2.44,
"learning_rate": 2.8982758620689655e-05,
"loss": 0.1064,
"step": 6600
},
{
"epoch": 2.45,
"learning_rate": 2.8896551724137933e-05,
"loss": 0.1026,
"step": 6625
},
{
"epoch": 2.46,
"learning_rate": 2.881034482758621e-05,
"loss": 0.1068,
"step": 6650
},
{
"epoch": 2.47,
"learning_rate": 2.8724137931034485e-05,
"loss": 0.1008,
"step": 6675
},
{
"epoch": 2.48,
"learning_rate": 2.8637931034482763e-05,
"loss": 0.1118,
"step": 6700
},
{
"epoch": 2.49,
"learning_rate": 2.8551724137931034e-05,
"loss": 0.1017,
"step": 6725
},
{
"epoch": 2.5,
"learning_rate": 2.8465517241379315e-05,
"loss": 0.1059,
"step": 6750
},
{
"epoch": 2.51,
"learning_rate": 2.8379310344827586e-05,
"loss": 0.1071,
"step": 6775
},
{
"epoch": 2.52,
"learning_rate": 2.8293103448275863e-05,
"loss": 0.1057,
"step": 6800
},
{
"epoch": 2.53,
"learning_rate": 2.8206896551724138e-05,
"loss": 0.1067,
"step": 6825
},
{
"epoch": 2.54,
"learning_rate": 2.8120689655172415e-05,
"loss": 0.1068,
"step": 6850
},
{
"epoch": 2.54,
"learning_rate": 2.803448275862069e-05,
"loss": 0.1175,
"step": 6875
},
{
"epoch": 2.55,
"learning_rate": 2.7948275862068968e-05,
"loss": 0.0992,
"step": 6900
},
{
"epoch": 2.56,
"learning_rate": 2.7862068965517242e-05,
"loss": 0.1085,
"step": 6925
},
{
"epoch": 2.57,
"learning_rate": 2.777586206896552e-05,
"loss": 0.1066,
"step": 6950
},
{
"epoch": 2.58,
"learning_rate": 2.768965517241379e-05,
"loss": 0.1013,
"step": 6975
},
{
"epoch": 2.59,
"learning_rate": 2.7603448275862072e-05,
"loss": 0.1048,
"step": 7000
},
{
"epoch": 2.59,
"eval_cer": 12.026084319077949,
"eval_loss": 0.25693368911743164,
"eval_runtime": 363.146,
"eval_samples_per_second": 7.107,
"eval_steps_per_second": 0.889,
"step": 7000
},
{
"epoch": 2.6,
"learning_rate": 2.7517241379310343e-05,
"loss": 0.1017,
"step": 7025
},
{
"epoch": 2.61,
"learning_rate": 2.743103448275862e-05,
"loss": 0.1024,
"step": 7050
},
{
"epoch": 2.62,
"learning_rate": 2.73448275862069e-05,
"loss": 0.108,
"step": 7075
},
{
"epoch": 2.63,
"learning_rate": 2.7258620689655173e-05,
"loss": 0.0979,
"step": 7100
},
{
"epoch": 2.64,
"learning_rate": 2.717241379310345e-05,
"loss": 0.1076,
"step": 7125
},
{
"epoch": 2.65,
"learning_rate": 2.7086206896551725e-05,
"loss": 0.1024,
"step": 7150
},
{
"epoch": 2.66,
"learning_rate": 2.7000000000000002e-05,
"loss": 0.1037,
"step": 7175
},
{
"epoch": 2.66,
"learning_rate": 2.6913793103448277e-05,
"loss": 0.1063,
"step": 7200
},
{
"epoch": 2.67,
"learning_rate": 2.6827586206896554e-05,
"loss": 0.1006,
"step": 7225
},
{
"epoch": 2.68,
"learning_rate": 2.674137931034483e-05,
"loss": 0.1093,
"step": 7250
},
{
"epoch": 2.69,
"learning_rate": 2.6655172413793107e-05,
"loss": 0.1057,
"step": 7275
},
{
"epoch": 2.7,
"learning_rate": 2.6568965517241378e-05,
"loss": 0.1015,
"step": 7300
},
{
"epoch": 2.71,
"learning_rate": 2.648275862068966e-05,
"loss": 0.1057,
"step": 7325
},
{
"epoch": 2.72,
"learning_rate": 2.639655172413793e-05,
"loss": 0.1067,
"step": 7350
},
{
"epoch": 2.73,
"learning_rate": 2.6310344827586207e-05,
"loss": 0.104,
"step": 7375
},
{
"epoch": 2.74,
"learning_rate": 2.6224137931034482e-05,
"loss": 0.1035,
"step": 7400
},
{
"epoch": 2.75,
"learning_rate": 2.613793103448276e-05,
"loss": 0.1054,
"step": 7425
},
{
"epoch": 2.76,
"learning_rate": 2.6051724137931037e-05,
"loss": 0.1083,
"step": 7450
},
{
"epoch": 2.77,
"learning_rate": 2.596551724137931e-05,
"loss": 0.1081,
"step": 7475
},
{
"epoch": 2.78,
"learning_rate": 2.587931034482759e-05,
"loss": 0.1056,
"step": 7500
},
{
"epoch": 2.78,
"eval_cer": 11.631786472550804,
"eval_loss": 0.26435503363609314,
"eval_runtime": 358.991,
"eval_samples_per_second": 7.19,
"eval_steps_per_second": 0.9,
"step": 7500
},
{
"epoch": 2.78,
"learning_rate": 2.5793103448275864e-05,
"loss": 0.103,
"step": 7525
},
{
"epoch": 2.79,
"learning_rate": 2.570689655172414e-05,
"loss": 0.1068,
"step": 7550
},
{
"epoch": 2.8,
"learning_rate": 2.5620689655172416e-05,
"loss": 0.1,
"step": 7575
},
{
"epoch": 2.81,
"learning_rate": 2.5534482758620693e-05,
"loss": 0.101,
"step": 7600
},
{
"epoch": 2.82,
"learning_rate": 2.5448275862068964e-05,
"loss": 0.1086,
"step": 7625
},
{
"epoch": 2.83,
"learning_rate": 2.5362068965517246e-05,
"loss": 0.1039,
"step": 7650
},
{
"epoch": 2.84,
"learning_rate": 2.5275862068965516e-05,
"loss": 0.1007,
"step": 7675
},
{
"epoch": 2.85,
"learning_rate": 2.5189655172413794e-05,
"loss": 0.1029,
"step": 7700
},
{
"epoch": 2.86,
"learning_rate": 2.510344827586207e-05,
"loss": 0.0989,
"step": 7725
},
{
"epoch": 2.87,
"learning_rate": 2.5017241379310346e-05,
"loss": 0.1042,
"step": 7750
},
{
"epoch": 2.88,
"learning_rate": 2.493103448275862e-05,
"loss": 0.1018,
"step": 7775
},
{
"epoch": 2.89,
"learning_rate": 2.48448275862069e-05,
"loss": 0.1017,
"step": 7800
},
{
"epoch": 2.9,
"learning_rate": 2.4758620689655173e-05,
"loss": 0.1028,
"step": 7825
},
{
"epoch": 2.91,
"learning_rate": 2.467241379310345e-05,
"loss": 0.0994,
"step": 7850
},
{
"epoch": 2.91,
"learning_rate": 2.4586206896551725e-05,
"loss": 0.1049,
"step": 7875
},
{
"epoch": 2.92,
"learning_rate": 2.45e-05,
"loss": 0.1005,
"step": 7900
},
{
"epoch": 2.93,
"learning_rate": 2.4413793103448277e-05,
"loss": 0.1027,
"step": 7925
},
{
"epoch": 2.94,
"learning_rate": 2.432758620689655e-05,
"loss": 0.1096,
"step": 7950
},
{
"epoch": 2.95,
"learning_rate": 2.424137931034483e-05,
"loss": 0.0999,
"step": 7975
},
{
"epoch": 2.96,
"learning_rate": 2.4155172413793103e-05,
"loss": 0.1021,
"step": 8000
},
{
"epoch": 2.96,
"eval_cer": 11.506672732787383,
"eval_loss": 0.2621135413646698,
"eval_runtime": 358.5537,
"eval_samples_per_second": 7.198,
"eval_steps_per_second": 0.901,
"step": 8000
},
{
"epoch": 2.97,
"learning_rate": 2.4068965517241378e-05,
"loss": 0.0984,
"step": 8025
},
{
"epoch": 2.98,
"learning_rate": 2.398275862068966e-05,
"loss": 0.0976,
"step": 8050
},
{
"epoch": 2.99,
"learning_rate": 2.3896551724137933e-05,
"loss": 0.1034,
"step": 8075
},
{
"epoch": 3.0,
"learning_rate": 2.3810344827586208e-05,
"loss": 0.1011,
"step": 8100
},
{
"epoch": 3.01,
"learning_rate": 2.3724137931034485e-05,
"loss": 0.0651,
"step": 8125
},
{
"epoch": 3.02,
"learning_rate": 2.363793103448276e-05,
"loss": 0.0473,
"step": 8150
},
{
"epoch": 3.03,
"learning_rate": 2.3551724137931037e-05,
"loss": 0.0448,
"step": 8175
},
{
"epoch": 3.03,
"learning_rate": 2.3465517241379312e-05,
"loss": 0.0473,
"step": 8200
},
{
"epoch": 3.04,
"learning_rate": 2.3379310344827586e-05,
"loss": 0.0457,
"step": 8225
},
{
"epoch": 3.05,
"learning_rate": 2.3293103448275864e-05,
"loss": 0.0451,
"step": 8250
},
{
"epoch": 3.06,
"learning_rate": 2.3206896551724138e-05,
"loss": 0.0453,
"step": 8275
},
{
"epoch": 3.07,
"learning_rate": 2.3120689655172416e-05,
"loss": 0.0502,
"step": 8300
},
{
"epoch": 3.08,
"learning_rate": 2.303448275862069e-05,
"loss": 0.0471,
"step": 8325
},
{
"epoch": 3.09,
"learning_rate": 2.2948275862068965e-05,
"loss": 0.0476,
"step": 8350
},
{
"epoch": 3.1,
"learning_rate": 2.2862068965517242e-05,
"loss": 0.0473,
"step": 8375
},
{
"epoch": 3.11,
"learning_rate": 2.2775862068965517e-05,
"loss": 0.0476,
"step": 8400
},
{
"epoch": 3.12,
"learning_rate": 2.2689655172413794e-05,
"loss": 0.0466,
"step": 8425
},
{
"epoch": 3.13,
"learning_rate": 2.2603448275862072e-05,
"loss": 0.0478,
"step": 8450
},
{
"epoch": 3.14,
"learning_rate": 2.2517241379310347e-05,
"loss": 0.0464,
"step": 8475
},
{
"epoch": 3.15,
"learning_rate": 2.2431034482758624e-05,
"loss": 0.0468,
"step": 8500
},
{
"epoch": 3.15,
"eval_cer": 11.332271762208068,
"eval_loss": 0.2711898386478424,
"eval_runtime": 359.6452,
"eval_samples_per_second": 7.177,
"eval_steps_per_second": 0.898,
"step": 8500
},
{
"epoch": 3.16,
"learning_rate": 2.23448275862069e-05,
"loss": 0.0468,
"step": 8525
},
{
"epoch": 3.16,
"learning_rate": 2.2258620689655173e-05,
"loss": 0.0526,
"step": 8550
},
{
"epoch": 3.17,
"learning_rate": 2.217241379310345e-05,
"loss": 0.0465,
"step": 8575
},
{
"epoch": 3.18,
"learning_rate": 2.2086206896551725e-05,
"loss": 0.0499,
"step": 8600
},
{
"epoch": 3.19,
"learning_rate": 2.2000000000000003e-05,
"loss": 0.0466,
"step": 8625
},
{
"epoch": 3.2,
"learning_rate": 2.1913793103448277e-05,
"loss": 0.0451,
"step": 8650
},
{
"epoch": 3.21,
"learning_rate": 2.182758620689655e-05,
"loss": 0.047,
"step": 8675
},
{
"epoch": 3.22,
"learning_rate": 2.174137931034483e-05,
"loss": 0.0482,
"step": 8700
},
{
"epoch": 3.23,
"learning_rate": 2.1655172413793104e-05,
"loss": 0.0513,
"step": 8725
},
{
"epoch": 3.24,
"learning_rate": 2.1568965517241378e-05,
"loss": 0.0483,
"step": 8750
},
{
"epoch": 3.25,
"learning_rate": 2.1482758620689656e-05,
"loss": 0.0491,
"step": 8775
},
{
"epoch": 3.26,
"learning_rate": 2.139655172413793e-05,
"loss": 0.0483,
"step": 8800
},
{
"epoch": 3.27,
"learning_rate": 2.1310344827586208e-05,
"loss": 0.0504,
"step": 8825
},
{
"epoch": 3.28,
"learning_rate": 2.1224137931034486e-05,
"loss": 0.0514,
"step": 8850
},
{
"epoch": 3.28,
"learning_rate": 2.113793103448276e-05,
"loss": 0.0484,
"step": 8875
},
{
"epoch": 3.29,
"learning_rate": 2.1051724137931038e-05,
"loss": 0.0474,
"step": 8900
},
{
"epoch": 3.3,
"learning_rate": 2.0965517241379312e-05,
"loss": 0.0462,
"step": 8925
},
{
"epoch": 3.31,
"learning_rate": 2.0879310344827586e-05,
"loss": 0.0501,
"step": 8950
},
{
"epoch": 3.32,
"learning_rate": 2.0793103448275864e-05,
"loss": 0.0487,
"step": 8975
},
{
"epoch": 3.33,
"learning_rate": 2.070689655172414e-05,
"loss": 0.0499,
"step": 9000
},
{
"epoch": 3.33,
"eval_cer": 11.555959963603275,
"eval_loss": 0.28167569637298584,
"eval_runtime": 359.0627,
"eval_samples_per_second": 7.188,
"eval_steps_per_second": 0.9,
"step": 9000
},
{
"epoch": 3.34,
"learning_rate": 2.0620689655172416e-05,
"loss": 0.0496,
"step": 9025
},
{
"epoch": 3.35,
"learning_rate": 2.053448275862069e-05,
"loss": 0.0471,
"step": 9050
},
{
"epoch": 3.36,
"learning_rate": 2.0448275862068965e-05,
"loss": 0.0469,
"step": 9075
},
{
"epoch": 3.37,
"learning_rate": 2.0362068965517243e-05,
"loss": 0.0511,
"step": 9100
},
{
"epoch": 3.38,
"learning_rate": 2.0275862068965517e-05,
"loss": 0.0476,
"step": 9125
},
{
"epoch": 3.39,
"learning_rate": 2.0189655172413795e-05,
"loss": 0.0483,
"step": 9150
},
{
"epoch": 3.4,
"learning_rate": 2.010344827586207e-05,
"loss": 0.0484,
"step": 9175
},
{
"epoch": 3.4,
"learning_rate": 2.0017241379310343e-05,
"loss": 0.0524,
"step": 9200
},
{
"epoch": 3.41,
"learning_rate": 1.993103448275862e-05,
"loss": 0.0469,
"step": 9225
},
{
"epoch": 3.42,
"learning_rate": 1.98448275862069e-05,
"loss": 0.0454,
"step": 9250
},
{
"epoch": 3.43,
"learning_rate": 1.9758620689655173e-05,
"loss": 0.0489,
"step": 9275
},
{
"epoch": 3.44,
"learning_rate": 1.967241379310345e-05,
"loss": 0.0482,
"step": 9300
},
{
"epoch": 3.45,
"learning_rate": 1.9586206896551725e-05,
"loss": 0.0487,
"step": 9325
},
{
"epoch": 3.46,
"learning_rate": 1.9500000000000003e-05,
"loss": 0.048,
"step": 9350
},
{
"epoch": 3.47,
"learning_rate": 1.9413793103448277e-05,
"loss": 0.0477,
"step": 9375
},
{
"epoch": 3.48,
"learning_rate": 1.9327586206896552e-05,
"loss": 0.0511,
"step": 9400
},
{
"epoch": 3.49,
"learning_rate": 1.924137931034483e-05,
"loss": 0.0468,
"step": 9425
},
{
"epoch": 3.5,
"learning_rate": 1.9155172413793104e-05,
"loss": 0.0502,
"step": 9450
},
{
"epoch": 3.51,
"learning_rate": 1.906896551724138e-05,
"loss": 0.0476,
"step": 9475
},
{
"epoch": 3.52,
"learning_rate": 1.8982758620689656e-05,
"loss": 0.0489,
"step": 9500
},
{
"epoch": 3.52,
"eval_cer": 11.548377312708523,
"eval_loss": 0.28109079599380493,
"eval_runtime": 359.545,
"eval_samples_per_second": 7.179,
"eval_steps_per_second": 0.898,
"step": 9500
},
{
"epoch": 3.53,
"learning_rate": 1.889655172413793e-05,
"loss": 0.0463,
"step": 9525
},
{
"epoch": 3.53,
"learning_rate": 1.8810344827586208e-05,
"loss": 0.0457,
"step": 9550
},
{
"epoch": 3.54,
"learning_rate": 1.8724137931034482e-05,
"loss": 0.0463,
"step": 9575
},
{
"epoch": 3.55,
"learning_rate": 1.863793103448276e-05,
"loss": 0.0526,
"step": 9600
},
{
"epoch": 3.56,
"learning_rate": 1.8551724137931034e-05,
"loss": 0.0499,
"step": 9625
},
{
"epoch": 3.57,
"learning_rate": 1.8465517241379312e-05,
"loss": 0.0488,
"step": 9650
},
{
"epoch": 3.58,
"learning_rate": 1.837931034482759e-05,
"loss": 0.0491,
"step": 9675
},
{
"epoch": 3.59,
"learning_rate": 1.8293103448275864e-05,
"loss": 0.046,
"step": 9700
},
{
"epoch": 3.6,
"learning_rate": 1.820689655172414e-05,
"loss": 0.0475,
"step": 9725
},
{
"epoch": 3.61,
"learning_rate": 1.8120689655172416e-05,
"loss": 0.0512,
"step": 9750
},
{
"epoch": 3.62,
"learning_rate": 1.803448275862069e-05,
"loss": 0.0478,
"step": 9775
},
{
"epoch": 3.63,
"learning_rate": 1.7948275862068965e-05,
"loss": 0.0602,
"step": 9800
},
{
"epoch": 3.64,
"learning_rate": 1.7862068965517243e-05,
"loss": 0.0482,
"step": 9825
},
{
"epoch": 3.65,
"learning_rate": 1.7775862068965517e-05,
"loss": 0.0463,
"step": 9850
},
{
"epoch": 3.65,
"learning_rate": 1.7689655172413795e-05,
"loss": 0.0513,
"step": 9875
},
{
"epoch": 3.66,
"learning_rate": 1.760344827586207e-05,
"loss": 0.0495,
"step": 9900
},
{
"epoch": 3.67,
"learning_rate": 1.7517241379310344e-05,
"loss": 0.0547,
"step": 9925
},
{
"epoch": 3.68,
"learning_rate": 1.743103448275862e-05,
"loss": 0.0485,
"step": 9950
},
{
"epoch": 3.69,
"learning_rate": 1.7344827586206896e-05,
"loss": 0.0498,
"step": 9975
},
{
"epoch": 3.7,
"learning_rate": 1.7258620689655173e-05,
"loss": 0.0479,
"step": 10000
},
{
"epoch": 3.7,
"eval_cer": 10.695329087048833,
"eval_loss": 0.27692216634750366,
"eval_runtime": 359.4994,
"eval_samples_per_second": 7.179,
"eval_steps_per_second": 0.898,
"step": 10000
},
{
"epoch": 3.71,
"learning_rate": 1.7172413793103448e-05,
"loss": 0.0456,
"step": 10025
},
{
"epoch": 3.72,
"learning_rate": 1.7086206896551722e-05,
"loss": 0.0512,
"step": 10050
},
{
"epoch": 3.73,
"learning_rate": 1.7000000000000003e-05,
"loss": 0.0487,
"step": 10075
},
{
"epoch": 3.74,
"learning_rate": 1.6913793103448278e-05,
"loss": 0.0463,
"step": 10100
},
{
"epoch": 3.75,
"learning_rate": 1.6827586206896552e-05,
"loss": 0.0625,
"step": 10125
},
{
"epoch": 3.76,
"learning_rate": 1.674137931034483e-05,
"loss": 0.0474,
"step": 10150
},
{
"epoch": 3.77,
"learning_rate": 1.6655172413793104e-05,
"loss": 0.0476,
"step": 10175
},
{
"epoch": 3.77,
"learning_rate": 1.6568965517241382e-05,
"loss": 0.0502,
"step": 10200
},
{
"epoch": 3.78,
"learning_rate": 1.6482758620689656e-05,
"loss": 0.0522,
"step": 10225
},
{
"epoch": 3.79,
"learning_rate": 1.639655172413793e-05,
"loss": 0.0458,
"step": 10250
},
{
"epoch": 3.8,
"learning_rate": 1.6310344827586208e-05,
"loss": 0.0482,
"step": 10275
},
{
"epoch": 3.81,
"learning_rate": 1.6224137931034483e-05,
"loss": 0.0466,
"step": 10300
},
{
"epoch": 3.82,
"learning_rate": 1.613793103448276e-05,
"loss": 0.0466,
"step": 10325
},
{
"epoch": 3.83,
"learning_rate": 1.6051724137931035e-05,
"loss": 0.0451,
"step": 10350
},
{
"epoch": 3.84,
"learning_rate": 1.596551724137931e-05,
"loss": 0.0492,
"step": 10375
},
{
"epoch": 3.85,
"learning_rate": 1.5879310344827587e-05,
"loss": 0.045,
"step": 10400
},
{
"epoch": 3.86,
"learning_rate": 1.579310344827586e-05,
"loss": 0.0457,
"step": 10425
},
{
"epoch": 3.87,
"learning_rate": 1.570689655172414e-05,
"loss": 0.0471,
"step": 10450
},
{
"epoch": 3.88,
"learning_rate": 1.5620689655172417e-05,
"loss": 0.0473,
"step": 10475
},
{
"epoch": 3.89,
"learning_rate": 1.553448275862069e-05,
"loss": 0.049,
"step": 10500
},
{
"epoch": 3.89,
"eval_cer": 11.51046405823476,
"eval_loss": 0.28438711166381836,
"eval_runtime": 359.2212,
"eval_samples_per_second": 7.185,
"eval_steps_per_second": 0.899,
"step": 10500
},
{
"epoch": 3.9,
"learning_rate": 1.544827586206897e-05,
"loss": 0.0458,
"step": 10525
},
{
"epoch": 3.9,
"learning_rate": 1.5362068965517243e-05,
"loss": 0.0456,
"step": 10550
},
{
"epoch": 3.91,
"learning_rate": 1.5275862068965517e-05,
"loss": 0.0474,
"step": 10575
},
{
"epoch": 3.92,
"learning_rate": 1.5189655172413793e-05,
"loss": 0.0447,
"step": 10600
},
{
"epoch": 3.93,
"learning_rate": 1.510344827586207e-05,
"loss": 0.046,
"step": 10625
},
{
"epoch": 3.94,
"learning_rate": 1.5017241379310346e-05,
"loss": 0.0442,
"step": 10650
},
{
"epoch": 3.95,
"learning_rate": 1.4931034482758622e-05,
"loss": 0.0457,
"step": 10675
},
{
"epoch": 3.96,
"learning_rate": 1.4844827586206898e-05,
"loss": 0.0482,
"step": 10700
},
{
"epoch": 3.97,
"learning_rate": 1.4758620689655172e-05,
"loss": 0.0476,
"step": 10725
},
{
"epoch": 3.98,
"learning_rate": 1.4672413793103448e-05,
"loss": 0.0449,
"step": 10750
},
{
"epoch": 3.99,
"learning_rate": 1.4586206896551724e-05,
"loss": 0.0478,
"step": 10775
},
{
"epoch": 4.0,
"learning_rate": 1.45e-05,
"loss": 0.0517,
"step": 10800
},
{
"epoch": 4.01,
"learning_rate": 1.4413793103448276e-05,
"loss": 0.0285,
"step": 10825
},
{
"epoch": 4.02,
"learning_rate": 1.432758620689655e-05,
"loss": 0.0197,
"step": 10850
},
{
"epoch": 4.02,
"learning_rate": 1.424137931034483e-05,
"loss": 0.0183,
"step": 10875
},
{
"epoch": 4.03,
"learning_rate": 1.4155172413793106e-05,
"loss": 0.0186,
"step": 10900
},
{
"epoch": 4.04,
"learning_rate": 1.406896551724138e-05,
"loss": 0.0203,
"step": 10925
},
{
"epoch": 4.05,
"learning_rate": 1.3982758620689656e-05,
"loss": 0.0185,
"step": 10950
},
{
"epoch": 4.06,
"learning_rate": 1.3896551724137932e-05,
"loss": 0.0173,
"step": 10975
},
{
"epoch": 4.07,
"learning_rate": 1.3810344827586208e-05,
"loss": 0.0193,
"step": 11000
},
{
"epoch": 4.07,
"eval_cer": 11.116166211707613,
"eval_loss": 0.29933854937553406,
"eval_runtime": 359.5628,
"eval_samples_per_second": 7.178,
"eval_steps_per_second": 0.898,
"step": 11000
},
{
"epoch": 4.08,
"learning_rate": 1.3724137931034484e-05,
"loss": 0.0176,
"step": 11025
},
{
"epoch": 4.09,
"learning_rate": 1.3637931034482759e-05,
"loss": 0.0171,
"step": 11050
},
{
"epoch": 4.1,
"learning_rate": 1.3551724137931035e-05,
"loss": 0.0173,
"step": 11075
},
{
"epoch": 4.11,
"learning_rate": 1.3465517241379311e-05,
"loss": 0.0185,
"step": 11100
},
{
"epoch": 4.12,
"learning_rate": 1.3379310344827587e-05,
"loss": 0.0161,
"step": 11125
},
{
"epoch": 4.13,
"learning_rate": 1.3293103448275861e-05,
"loss": 0.0186,
"step": 11150
},
{
"epoch": 4.14,
"learning_rate": 1.3206896551724137e-05,
"loss": 0.0191,
"step": 11175
},
{
"epoch": 4.15,
"learning_rate": 1.3120689655172413e-05,
"loss": 0.0177,
"step": 11200
},
{
"epoch": 4.15,
"learning_rate": 1.303448275862069e-05,
"loss": 0.0184,
"step": 11225
},
{
"epoch": 4.16,
"learning_rate": 1.2948275862068966e-05,
"loss": 0.0195,
"step": 11250
},
{
"epoch": 4.17,
"learning_rate": 1.2862068965517243e-05,
"loss": 0.019,
"step": 11275
},
{
"epoch": 4.18,
"learning_rate": 1.277586206896552e-05,
"loss": 0.0183,
"step": 11300
},
{
"epoch": 4.19,
"learning_rate": 1.2689655172413795e-05,
"loss": 0.0188,
"step": 11325
},
{
"epoch": 4.2,
"learning_rate": 1.260344827586207e-05,
"loss": 0.0209,
"step": 11350
},
{
"epoch": 4.21,
"learning_rate": 1.2517241379310346e-05,
"loss": 0.0171,
"step": 11375
},
{
"epoch": 4.22,
"learning_rate": 1.2431034482758622e-05,
"loss": 0.0163,
"step": 11400
},
{
"epoch": 4.23,
"learning_rate": 1.2344827586206898e-05,
"loss": 0.0181,
"step": 11425
},
{
"epoch": 4.24,
"learning_rate": 1.2258620689655174e-05,
"loss": 0.0169,
"step": 11450
},
{
"epoch": 4.25,
"learning_rate": 1.2172413793103448e-05,
"loss": 0.0197,
"step": 11475
},
{
"epoch": 4.26,
"learning_rate": 1.2086206896551724e-05,
"loss": 0.0163,
"step": 11500
},
{
"epoch": 4.26,
"eval_cer": 11.010009099181074,
"eval_loss": 0.3046729266643524,
"eval_runtime": 358.1502,
"eval_samples_per_second": 7.206,
"eval_steps_per_second": 0.902,
"step": 11500
},
{
"epoch": 4.27,
"learning_rate": 1.2e-05,
"loss": 0.0174,
"step": 11525
},
{
"epoch": 4.27,
"learning_rate": 1.1913793103448276e-05,
"loss": 0.0179,
"step": 11550
},
{
"epoch": 4.28,
"learning_rate": 1.1827586206896552e-05,
"loss": 0.0189,
"step": 11575
},
{
"epoch": 4.29,
"learning_rate": 1.1741379310344828e-05,
"loss": 0.0202,
"step": 11600
},
{
"epoch": 4.3,
"learning_rate": 1.1655172413793104e-05,
"loss": 0.017,
"step": 11625
},
{
"epoch": 4.31,
"learning_rate": 1.156896551724138e-05,
"loss": 0.0186,
"step": 11650
},
{
"epoch": 4.32,
"learning_rate": 1.1482758620689655e-05,
"loss": 0.0193,
"step": 11675
},
{
"epoch": 4.33,
"learning_rate": 1.1396551724137931e-05,
"loss": 0.0173,
"step": 11700
},
{
"epoch": 4.34,
"learning_rate": 1.1310344827586207e-05,
"loss": 0.018,
"step": 11725
},
{
"epoch": 4.35,
"learning_rate": 1.1224137931034483e-05,
"loss": 0.0166,
"step": 11750
},
{
"epoch": 4.36,
"learning_rate": 1.1137931034482759e-05,
"loss": 0.0176,
"step": 11775
},
{
"epoch": 4.37,
"learning_rate": 1.1051724137931035e-05,
"loss": 0.0195,
"step": 11800
},
{
"epoch": 4.38,
"learning_rate": 1.0965517241379311e-05,
"loss": 0.0179,
"step": 11825
},
{
"epoch": 4.39,
"learning_rate": 1.0879310344827587e-05,
"loss": 0.0174,
"step": 11850
},
{
"epoch": 4.39,
"learning_rate": 1.0793103448275863e-05,
"loss": 0.0183,
"step": 11875
},
{
"epoch": 4.4,
"learning_rate": 1.0706896551724138e-05,
"loss": 0.0178,
"step": 11900
},
{
"epoch": 4.41,
"learning_rate": 1.0620689655172414e-05,
"loss": 0.0177,
"step": 11925
},
{
"epoch": 4.42,
"learning_rate": 1.053448275862069e-05,
"loss": 0.0186,
"step": 11950
},
{
"epoch": 4.43,
"learning_rate": 1.0448275862068967e-05,
"loss": 0.0174,
"step": 11975
},
{
"epoch": 4.44,
"learning_rate": 1.0362068965517242e-05,
"loss": 0.0175,
"step": 12000
},
{
"epoch": 4.44,
"eval_cer": 10.911434637549288,
"eval_loss": 0.30570268630981445,
"eval_runtime": 358.4295,
"eval_samples_per_second": 7.201,
"eval_steps_per_second": 0.901,
"step": 12000
},
{
"epoch": 4.45,
"learning_rate": 1.0275862068965518e-05,
"loss": 0.019,
"step": 12025
},
{
"epoch": 4.46,
"learning_rate": 1.0189655172413794e-05,
"loss": 0.0179,
"step": 12050
},
{
"epoch": 4.47,
"learning_rate": 1.010344827586207e-05,
"loss": 0.0192,
"step": 12075
},
{
"epoch": 4.48,
"learning_rate": 1.0017241379310344e-05,
"loss": 0.0167,
"step": 12100
},
{
"epoch": 4.49,
"learning_rate": 9.93103448275862e-06,
"loss": 0.0178,
"step": 12125
},
{
"epoch": 4.5,
"learning_rate": 9.844827586206896e-06,
"loss": 0.0209,
"step": 12150
},
{
"epoch": 4.51,
"learning_rate": 9.758620689655174e-06,
"loss": 0.019,
"step": 12175
},
{
"epoch": 4.52,
"learning_rate": 9.672413793103448e-06,
"loss": 0.0169,
"step": 12200
},
{
"epoch": 4.52,
"learning_rate": 9.586206896551724e-06,
"loss": 0.0175,
"step": 12225
},
{
"epoch": 4.53,
"learning_rate": 9.5e-06,
"loss": 0.0166,
"step": 12250
},
{
"epoch": 4.54,
"learning_rate": 9.413793103448277e-06,
"loss": 0.0184,
"step": 12275
},
{
"epoch": 4.55,
"learning_rate": 9.327586206896553e-06,
"loss": 0.0256,
"step": 12300
},
{
"epoch": 4.56,
"learning_rate": 9.241379310344827e-06,
"loss": 0.0177,
"step": 12325
},
{
"epoch": 4.57,
"learning_rate": 9.155172413793103e-06,
"loss": 0.018,
"step": 12350
},
{
"epoch": 4.58,
"learning_rate": 9.06896551724138e-06,
"loss": 0.0187,
"step": 12375
},
{
"epoch": 4.59,
"learning_rate": 8.982758620689657e-06,
"loss": 0.0175,
"step": 12400
},
{
"epoch": 4.6,
"learning_rate": 8.896551724137931e-06,
"loss": 0.0204,
"step": 12425
},
{
"epoch": 4.61,
"learning_rate": 8.810344827586207e-06,
"loss": 0.0177,
"step": 12450
},
{
"epoch": 4.62,
"learning_rate": 8.724137931034483e-06,
"loss": 0.0174,
"step": 12475
},
{
"epoch": 4.63,
"learning_rate": 8.63793103448276e-06,
"loss": 0.0186,
"step": 12500
},
{
"epoch": 4.63,
"eval_cer": 11.127540188049743,
"eval_loss": 0.30721473693847656,
"eval_runtime": 359.2661,
"eval_samples_per_second": 7.184,
"eval_steps_per_second": 0.899,
"step": 12500
},
{
"epoch": 4.64,
"learning_rate": 8.555172413793104e-06,
"loss": 0.0166,
"step": 12525
},
{
"epoch": 4.64,
"learning_rate": 8.46896551724138e-06,
"loss": 0.0196,
"step": 12550
},
{
"epoch": 4.65,
"learning_rate": 8.382758620689656e-06,
"loss": 0.0187,
"step": 12575
},
{
"epoch": 4.66,
"learning_rate": 8.296551724137932e-06,
"loss": 0.0179,
"step": 12600
},
{
"epoch": 4.67,
"learning_rate": 8.210344827586206e-06,
"loss": 0.016,
"step": 12625
},
{
"epoch": 4.68,
"learning_rate": 8.124137931034482e-06,
"loss": 0.022,
"step": 12650
},
{
"epoch": 4.69,
"learning_rate": 8.03793103448276e-06,
"loss": 0.0156,
"step": 12675
},
{
"epoch": 4.7,
"learning_rate": 7.951724137931036e-06,
"loss": 0.0172,
"step": 12700
},
{
"epoch": 4.71,
"learning_rate": 7.86551724137931e-06,
"loss": 0.0188,
"step": 12725
},
{
"epoch": 4.72,
"learning_rate": 7.779310344827586e-06,
"loss": 0.0208,
"step": 12750
},
{
"epoch": 4.73,
"learning_rate": 7.693103448275862e-06,
"loss": 0.0176,
"step": 12775
},
{
"epoch": 4.74,
"learning_rate": 7.6068965517241385e-06,
"loss": 0.018,
"step": 12800
},
{
"epoch": 4.75,
"learning_rate": 7.520689655172414e-06,
"loss": 0.0168,
"step": 12825
},
{
"epoch": 4.76,
"learning_rate": 7.43448275862069e-06,
"loss": 0.0159,
"step": 12850
},
{
"epoch": 4.76,
"learning_rate": 7.348275862068967e-06,
"loss": 0.0166,
"step": 12875
},
{
"epoch": 4.77,
"learning_rate": 7.262068965517242e-06,
"loss": 0.0182,
"step": 12900
},
{
"epoch": 4.78,
"learning_rate": 7.175862068965518e-06,
"loss": 0.0166,
"step": 12925
},
{
"epoch": 4.79,
"learning_rate": 7.089655172413794e-06,
"loss": 0.0168,
"step": 12950
},
{
"epoch": 4.8,
"learning_rate": 7.003448275862069e-06,
"loss": 0.0174,
"step": 12975
},
{
"epoch": 4.81,
"learning_rate": 6.917241379310345e-06,
"loss": 0.0176,
"step": 13000
},
{
"epoch": 4.81,
"eval_cer": 11.229905975128904,
"eval_loss": 0.30905863642692566,
"eval_runtime": 359.315,
"eval_samples_per_second": 7.183,
"eval_steps_per_second": 0.899,
"step": 13000
},
{
"epoch": 4.82,
"learning_rate": 6.83103448275862e-06,
"loss": 0.0175,
"step": 13025
},
{
"epoch": 4.83,
"learning_rate": 6.744827586206896e-06,
"loss": 0.017,
"step": 13050
},
{
"epoch": 4.84,
"learning_rate": 6.658620689655173e-06,
"loss": 0.0158,
"step": 13075
},
{
"epoch": 4.85,
"learning_rate": 6.572413793103449e-06,
"loss": 0.0182,
"step": 13100
},
{
"epoch": 4.86,
"learning_rate": 6.4862068965517245e-06,
"loss": 0.0175,
"step": 13125
},
{
"epoch": 4.87,
"learning_rate": 6.4000000000000006e-06,
"loss": 0.0191,
"step": 13150
},
{
"epoch": 4.88,
"learning_rate": 6.313793103448276e-06,
"loss": 0.0153,
"step": 13175
},
{
"epoch": 4.89,
"learning_rate": 6.227586206896552e-06,
"loss": 0.0197,
"step": 13200
},
{
"epoch": 4.89,
"learning_rate": 6.141379310344828e-06,
"loss": 0.0165,
"step": 13225
},
{
"epoch": 4.9,
"learning_rate": 6.055172413793104e-06,
"loss": 0.0157,
"step": 13250
},
{
"epoch": 4.91,
"learning_rate": 5.96896551724138e-06,
"loss": 0.0156,
"step": 13275
},
{
"epoch": 4.92,
"learning_rate": 5.882758620689655e-06,
"loss": 0.0175,
"step": 13300
},
{
"epoch": 4.93,
"learning_rate": 5.796551724137932e-06,
"loss": 0.0166,
"step": 13325
},
{
"epoch": 4.94,
"learning_rate": 5.710344827586207e-06,
"loss": 0.016,
"step": 13350
},
{
"epoch": 4.95,
"learning_rate": 5.624137931034483e-06,
"loss": 0.0146,
"step": 13375
},
{
"epoch": 4.96,
"learning_rate": 5.5379310344827585e-06,
"loss": 0.0181,
"step": 13400
},
{
"epoch": 4.97,
"learning_rate": 5.451724137931035e-06,
"loss": 0.016,
"step": 13425
},
{
"epoch": 4.98,
"learning_rate": 5.3655172413793106e-06,
"loss": 0.0186,
"step": 13450
},
{
"epoch": 4.99,
"learning_rate": 5.279310344827587e-06,
"loss": 0.0157,
"step": 13475
},
{
"epoch": 5.0,
"learning_rate": 5.193103448275862e-06,
"loss": 0.0173,
"step": 13500
},
{
"epoch": 5.0,
"eval_cer": 11.082044282681224,
"eval_loss": 0.31240326166152954,
"eval_runtime": 358.0319,
"eval_samples_per_second": 7.209,
"eval_steps_per_second": 0.902,
"step": 13500
},
{
"epoch": 5.01,
"learning_rate": 5.106896551724139e-06,
"loss": 0.0115,
"step": 13525
},
{
"epoch": 5.01,
"learning_rate": 5.020689655172414e-06,
"loss": 0.0083,
"step": 13550
},
{
"epoch": 5.02,
"learning_rate": 4.93448275862069e-06,
"loss": 0.0078,
"step": 13575
},
{
"epoch": 5.03,
"learning_rate": 4.848275862068966e-06,
"loss": 0.0081,
"step": 13600
},
{
"epoch": 5.04,
"learning_rate": 4.762068965517242e-06,
"loss": 0.007,
"step": 13625
},
{
"epoch": 5.05,
"learning_rate": 4.675862068965517e-06,
"loss": 0.008,
"step": 13650
},
{
"epoch": 5.06,
"learning_rate": 4.589655172413793e-06,
"loss": 0.0076,
"step": 13675
},
{
"epoch": 5.07,
"learning_rate": 4.503448275862069e-06,
"loss": 0.0093,
"step": 13700
},
{
"epoch": 5.08,
"learning_rate": 4.417241379310345e-06,
"loss": 0.0073,
"step": 13725
},
{
"epoch": 5.09,
"learning_rate": 4.331034482758621e-06,
"loss": 0.0072,
"step": 13750
},
{
"epoch": 5.1,
"learning_rate": 4.244827586206897e-06,
"loss": 0.0079,
"step": 13775
},
{
"epoch": 5.11,
"learning_rate": 4.158620689655173e-06,
"loss": 0.0076,
"step": 13800
},
{
"epoch": 5.12,
"learning_rate": 4.072413793103449e-06,
"loss": 0.0071,
"step": 13825
},
{
"epoch": 5.13,
"learning_rate": 3.986206896551725e-06,
"loss": 0.0071,
"step": 13850
},
{
"epoch": 5.14,
"learning_rate": 3.9e-06,
"loss": 0.0067,
"step": 13875
},
{
"epoch": 5.14,
"learning_rate": 3.813793103448276e-06,
"loss": 0.0077,
"step": 13900
},
{
"epoch": 5.15,
"learning_rate": 3.7275862068965524e-06,
"loss": 0.0089,
"step": 13925
},
{
"epoch": 5.16,
"learning_rate": 3.641379310344828e-06,
"loss": 0.0077,
"step": 13950
},
{
"epoch": 5.17,
"learning_rate": 3.5551724137931037e-06,
"loss": 0.008,
"step": 13975
},
{
"epoch": 5.18,
"learning_rate": 3.4689655172413793e-06,
"loss": 0.007,
"step": 14000
},
{
"epoch": 5.18,
"eval_cer": 10.862147406733394,
"eval_loss": 0.3205486238002777,
"eval_runtime": 361.4728,
"eval_samples_per_second": 7.14,
"eval_steps_per_second": 0.894,
"step": 14000
},
{
"epoch": 5.19,
"learning_rate": 3.382758620689655e-06,
"loss": 0.0088,
"step": 14025
},
{
"epoch": 5.2,
"learning_rate": 3.2965517241379314e-06,
"loss": 0.0072,
"step": 14050
},
{
"epoch": 5.21,
"learning_rate": 3.210344827586207e-06,
"loss": 0.0085,
"step": 14075
},
{
"epoch": 5.22,
"learning_rate": 3.1241379310344826e-06,
"loss": 0.0068,
"step": 14100
},
{
"epoch": 5.23,
"learning_rate": 3.0379310344827587e-06,
"loss": 0.0081,
"step": 14125
},
{
"epoch": 5.24,
"learning_rate": 2.9517241379310343e-06,
"loss": 0.0074,
"step": 14150
},
{
"epoch": 5.25,
"learning_rate": 2.8655172413793104e-06,
"loss": 0.0077,
"step": 14175
},
{
"epoch": 5.26,
"learning_rate": 2.779310344827586e-06,
"loss": 0.0083,
"step": 14200
},
{
"epoch": 5.26,
"learning_rate": 2.693103448275862e-06,
"loss": 0.008,
"step": 14225
},
{
"epoch": 5.27,
"learning_rate": 2.606896551724138e-06,
"loss": 0.0064,
"step": 14250
},
{
"epoch": 5.28,
"learning_rate": 2.520689655172414e-06,
"loss": 0.0076,
"step": 14275
},
{
"epoch": 5.29,
"learning_rate": 2.4344827586206897e-06,
"loss": 0.0066,
"step": 14300
},
{
"epoch": 5.3,
"learning_rate": 2.3482758620689658e-06,
"loss": 0.008,
"step": 14325
},
{
"epoch": 5.31,
"learning_rate": 2.2620689655172414e-06,
"loss": 0.0072,
"step": 14350
},
{
"epoch": 5.32,
"learning_rate": 2.1758620689655174e-06,
"loss": 0.0078,
"step": 14375
},
{
"epoch": 5.33,
"learning_rate": 2.089655172413793e-06,
"loss": 0.0091,
"step": 14400
},
{
"epoch": 5.34,
"learning_rate": 2.003448275862069e-06,
"loss": 0.007,
"step": 14425
},
{
"epoch": 5.35,
"learning_rate": 1.9172413793103447e-06,
"loss": 0.0078,
"step": 14450
},
{
"epoch": 5.36,
"learning_rate": 1.8310344827586208e-06,
"loss": 0.0074,
"step": 14475
},
{
"epoch": 5.37,
"learning_rate": 1.7448275862068964e-06,
"loss": 0.0073,
"step": 14500
},
{
"epoch": 5.37,
"eval_cer": 10.790112223233242,
"eval_loss": 0.320094496011734,
"eval_runtime": 359.2372,
"eval_samples_per_second": 7.185,
"eval_steps_per_second": 0.899,
"step": 14500
},
{
"epoch": 5.38,
"learning_rate": 1.6586206896551724e-06,
"loss": 0.0094,
"step": 14525
},
{
"epoch": 5.38,
"learning_rate": 1.5724137931034483e-06,
"loss": 0.0142,
"step": 14550
},
{
"epoch": 5.39,
"learning_rate": 1.4862068965517243e-06,
"loss": 0.0085,
"step": 14575
},
{
"epoch": 5.4,
"learning_rate": 1.4034482758620691e-06,
"loss": 0.0113,
"step": 14600
},
{
"epoch": 5.41,
"learning_rate": 1.317241379310345e-06,
"loss": 0.0085,
"step": 14625
},
{
"epoch": 5.42,
"learning_rate": 1.2310344827586208e-06,
"loss": 0.0076,
"step": 14650
},
{
"epoch": 5.43,
"learning_rate": 1.1448275862068966e-06,
"loss": 0.0074,
"step": 14675
},
{
"epoch": 5.44,
"learning_rate": 1.0586206896551725e-06,
"loss": 0.0079,
"step": 14700
},
{
"epoch": 5.45,
"learning_rate": 9.724137931034483e-07,
"loss": 0.0067,
"step": 14725
},
{
"epoch": 5.46,
"learning_rate": 8.862068965517242e-07,
"loss": 0.0075,
"step": 14750
},
{
"epoch": 5.47,
"learning_rate": 8.000000000000001e-07,
"loss": 0.0082,
"step": 14775
},
{
"epoch": 5.48,
"learning_rate": 7.137931034482759e-07,
"loss": 0.008,
"step": 14800
},
{
"epoch": 5.49,
"learning_rate": 6.275862068965517e-07,
"loss": 0.0066,
"step": 14825
},
{
"epoch": 5.5,
"learning_rate": 5.413793103448277e-07,
"loss": 0.0077,
"step": 14850
},
{
"epoch": 5.51,
"learning_rate": 4.551724137931035e-07,
"loss": 0.0075,
"step": 14875
},
{
"epoch": 5.51,
"learning_rate": 3.689655172413793e-07,
"loss": 0.0086,
"step": 14900
},
{
"epoch": 5.52,
"learning_rate": 2.827586206896552e-07,
"loss": 0.0067,
"step": 14925
},
{
"epoch": 5.53,
"learning_rate": 1.9655172413793103e-07,
"loss": 0.0075,
"step": 14950
},
{
"epoch": 5.54,
"learning_rate": 1.1034482758620691e-07,
"loss": 0.0067,
"step": 14975
},
{
"epoch": 5.55,
"learning_rate": 2.413793103448276e-08,
"loss": 0.0072,
"step": 15000
},
{
"epoch": 5.55,
"eval_cer": 10.763572945101608,
"eval_loss": 0.3218766450881958,
"eval_runtime": 359.0558,
"eval_samples_per_second": 7.188,
"eval_steps_per_second": 0.9,
"step": 15000
}
],
"logging_steps": 25,
"max_steps": 15000,
"num_input_tokens_seen": 0,
"num_train_epochs": 6,
"save_steps": 500,
"total_flos": 4.328506848669696e+20,
"train_batch_size": 25,
"trial_name": null,
"trial_params": null
}