{ "best_metric": 14.237288135593221, "best_model_checkpoint": "whisper-small-pt-cv11-v4_2/checkpoint-4500", "epoch": 11.049723756906078, "global_step": 6000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.05, "learning_rate": 1.0500000000000001e-07, "loss": 2.896, "step": 25 }, { "epoch": 0.09, "learning_rate": 2.3000000000000002e-07, "loss": 2.8972, "step": 50 }, { "epoch": 0.14, "learning_rate": 3.55e-07, "loss": 2.8065, "step": 75 }, { "epoch": 0.18, "learning_rate": 4.800000000000001e-07, "loss": 2.5992, "step": 100 }, { "epoch": 0.23, "learning_rate": 6.05e-07, "loss": 2.3887, "step": 125 }, { "epoch": 0.28, "learning_rate": 7.3e-07, "loss": 2.2139, "step": 150 }, { "epoch": 0.32, "learning_rate": 8.550000000000002e-07, "loss": 2.0286, "step": 175 }, { "epoch": 0.37, "learning_rate": 9.800000000000001e-07, "loss": 1.8338, "step": 200 }, { "epoch": 0.41, "learning_rate": 1.105e-06, "loss": 1.5814, "step": 225 }, { "epoch": 0.46, "learning_rate": 1.23e-06, "loss": 1.3991, "step": 250 }, { "epoch": 0.51, "learning_rate": 1.3550000000000002e-06, "loss": 1.4435, "step": 275 }, { "epoch": 0.55, "learning_rate": 1.48e-06, "loss": 1.3595, "step": 300 }, { "epoch": 0.6, "learning_rate": 1.605e-06, "loss": 1.3124, "step": 325 }, { "epoch": 0.64, "learning_rate": 1.73e-06, "loss": 1.2619, "step": 350 }, { "epoch": 0.69, "learning_rate": 1.8550000000000002e-06, "loss": 1.2197, "step": 375 }, { "epoch": 0.74, "learning_rate": 1.98e-06, "loss": 1.1799, "step": 400 }, { "epoch": 0.78, "learning_rate": 2.105e-06, "loss": 1.149, "step": 425 }, { "epoch": 0.83, "learning_rate": 2.2300000000000002e-06, "loss": 1.2276, "step": 450 }, { "epoch": 0.87, "learning_rate": 2.355e-06, "loss": 1.1621, "step": 475 }, { "epoch": 0.92, "learning_rate": 2.4800000000000004e-06, "loss": 1.1113, "step": 500 }, { "epoch": 0.92, "eval_cer": 6.791931472782537, "eval_loss": 0.3896943926811218, "eval_runtime": 192.8939, "eval_samples_per_second": 5.184, "eval_steps_per_second": 0.327, "eval_wer": 16.87211093990755, "step": 500 }, { "epoch": 0.97, "learning_rate": 2.6050000000000005e-06, "loss": 1.0856, "step": 525 }, { "epoch": 1.01, "learning_rate": 2.7300000000000005e-06, "loss": 1.0536, "step": 550 }, { "epoch": 1.06, "learning_rate": 2.855e-06, "loss": 1.0266, "step": 575 }, { "epoch": 1.1, "learning_rate": 2.9800000000000003e-06, "loss": 1.0755, "step": 600 }, { "epoch": 1.15, "learning_rate": 3.1050000000000003e-06, "loss": 0.974, "step": 625 }, { "epoch": 1.2, "learning_rate": 3.2300000000000004e-06, "loss": 1.0042, "step": 650 }, { "epoch": 1.24, "learning_rate": 3.3550000000000005e-06, "loss": 1.0075, "step": 675 }, { "epoch": 1.29, "learning_rate": 3.48e-06, "loss": 1.0212, "step": 700 }, { "epoch": 1.34, "learning_rate": 3.6050000000000002e-06, "loss": 0.9171, "step": 725 }, { "epoch": 1.38, "learning_rate": 3.7300000000000003e-06, "loss": 0.9601, "step": 750 }, { "epoch": 1.43, "learning_rate": 3.855e-06, "loss": 0.9757, "step": 775 }, { "epoch": 1.47, "learning_rate": 3.980000000000001e-06, "loss": 0.9636, "step": 800 }, { "epoch": 1.52, "learning_rate": 4.1050000000000005e-06, "loss": 0.9497, "step": 825 }, { "epoch": 1.57, "learning_rate": 4.23e-06, "loss": 0.9699, "step": 850 }, { "epoch": 1.61, "learning_rate": 4.355000000000001e-06, "loss": 0.9353, "step": 875 }, { "epoch": 1.66, "learning_rate": 4.48e-06, "loss": 0.9322, "step": 900 }, { "epoch": 1.7, "learning_rate": 4.605000000000001e-06, "loss": 0.9297, "step": 925 }, { "epoch": 1.75, "learning_rate": 4.7300000000000005e-06, "loss": 0.9201, "step": 950 }, { "epoch": 1.8, "learning_rate": 4.855e-06, "loss": 0.8761, "step": 975 }, { "epoch": 1.84, "learning_rate": 4.980000000000001e-06, "loss": 0.9009, "step": 1000 }, { "epoch": 1.84, "eval_cer": 6.231003039513678, "eval_loss": 0.331758975982666, "eval_runtime": 194.6762, "eval_samples_per_second": 5.137, "eval_steps_per_second": 0.324, "eval_wer": 15.932203389830507, "step": 1000 }, { "epoch": 1.89, "learning_rate": 4.9883333333333336e-06, "loss": 0.8478, "step": 1025 }, { "epoch": 1.93, "learning_rate": 4.974444444444445e-06, "loss": 0.8491, "step": 1050 }, { "epoch": 1.98, "learning_rate": 4.9605555555555564e-06, "loss": 0.8931, "step": 1075 }, { "epoch": 2.03, "learning_rate": 4.946666666666667e-06, "loss": 0.8313, "step": 1100 }, { "epoch": 2.07, "learning_rate": 4.932777777777778e-06, "loss": 0.7996, "step": 1125 }, { "epoch": 2.12, "learning_rate": 4.918888888888889e-06, "loss": 0.8184, "step": 1150 }, { "epoch": 2.16, "learning_rate": 4.9050000000000005e-06, "loss": 0.8026, "step": 1175 }, { "epoch": 2.21, "learning_rate": 4.891111111111111e-06, "loss": 0.7925, "step": 1200 }, { "epoch": 2.26, "learning_rate": 4.8772222222222225e-06, "loss": 0.8085, "step": 1225 }, { "epoch": 2.3, "learning_rate": 4.863333333333334e-06, "loss": 0.7819, "step": 1250 }, { "epoch": 2.35, "learning_rate": 4.849444444444445e-06, "loss": 0.7734, "step": 1275 }, { "epoch": 2.39, "learning_rate": 4.835555555555556e-06, "loss": 0.7978, "step": 1300 }, { "epoch": 2.44, "learning_rate": 4.821666666666667e-06, "loss": 0.7987, "step": 1325 }, { "epoch": 2.49, "learning_rate": 4.807777777777779e-06, "loss": 0.7843, "step": 1350 }, { "epoch": 2.53, "learning_rate": 4.793888888888889e-06, "loss": 0.7916, "step": 1375 }, { "epoch": 2.58, "learning_rate": 4.78e-06, "loss": 0.7978, "step": 1400 }, { "epoch": 2.62, "learning_rate": 4.766111111111111e-06, "loss": 0.7801, "step": 1425 }, { "epoch": 2.67, "learning_rate": 4.752222222222223e-06, "loss": 0.7604, "step": 1450 }, { "epoch": 2.72, "learning_rate": 4.7383333333333334e-06, "loss": 0.7827, "step": 1475 }, { "epoch": 2.76, "learning_rate": 4.724444444444445e-06, "loss": 0.7631, "step": 1500 }, { "epoch": 2.76, "eval_cer": 5.893893340701852, "eval_loss": 0.317706823348999, "eval_runtime": 195.4814, "eval_samples_per_second": 5.116, "eval_steps_per_second": 0.322, "eval_wer": 15.485362095531588, "step": 1500 }, { "epoch": 2.81, "learning_rate": 4.710555555555556e-06, "loss": 0.7634, "step": 1525 }, { "epoch": 2.85, "learning_rate": 4.696666666666667e-06, "loss": 0.768, "step": 1550 }, { "epoch": 2.9, "learning_rate": 4.682777777777778e-06, "loss": 0.7462, "step": 1575 }, { "epoch": 2.95, "learning_rate": 4.66888888888889e-06, "loss": 0.7295, "step": 1600 }, { "epoch": 2.99, "learning_rate": 4.655e-06, "loss": 0.7625, "step": 1625 }, { "epoch": 3.04, "learning_rate": 4.641111111111111e-06, "loss": 0.697, "step": 1650 }, { "epoch": 3.08, "learning_rate": 4.627222222222222e-06, "loss": 0.7002, "step": 1675 }, { "epoch": 3.13, "learning_rate": 4.613333333333334e-06, "loss": 0.675, "step": 1700 }, { "epoch": 3.18, "learning_rate": 4.599444444444444e-06, "loss": 0.6891, "step": 1725 }, { "epoch": 3.22, "learning_rate": 4.585555555555556e-06, "loss": 0.6812, "step": 1750 }, { "epoch": 3.27, "learning_rate": 4.571666666666667e-06, "loss": 0.7044, "step": 1775 }, { "epoch": 3.31, "learning_rate": 4.557777777777778e-06, "loss": 0.6219, "step": 1800 }, { "epoch": 3.36, "learning_rate": 4.543888888888889e-06, "loss": 0.6708, "step": 1825 }, { "epoch": 3.41, "learning_rate": 4.530000000000001e-06, "loss": 0.7114, "step": 1850 }, { "epoch": 3.45, "learning_rate": 4.516111111111111e-06, "loss": 0.6869, "step": 1875 }, { "epoch": 3.5, "learning_rate": 4.502222222222223e-06, "loss": 0.6751, "step": 1900 }, { "epoch": 3.55, "learning_rate": 4.488333333333333e-06, "loss": 0.7053, "step": 1925 }, { "epoch": 3.59, "learning_rate": 4.474444444444445e-06, "loss": 0.7169, "step": 1950 }, { "epoch": 3.64, "learning_rate": 4.460555555555556e-06, "loss": 0.679, "step": 1975 }, { "epoch": 3.68, "learning_rate": 4.446666666666667e-06, "loss": 0.7163, "step": 2000 }, { "epoch": 3.68, "eval_cer": 5.797181541862393, "eval_loss": 0.3130023181438446, "eval_runtime": 195.4882, "eval_samples_per_second": 5.115, "eval_steps_per_second": 0.322, "eval_wer": 14.899845916795067, "step": 2000 }, { "epoch": 3.73, "learning_rate": 4.432777777777778e-06, "loss": 0.6868, "step": 2025 }, { "epoch": 3.78, "learning_rate": 4.41888888888889e-06, "loss": 0.6776, "step": 2050 }, { "epoch": 3.82, "learning_rate": 4.405e-06, "loss": 0.6681, "step": 2075 }, { "epoch": 3.87, "learning_rate": 4.391111111111112e-06, "loss": 0.7182, "step": 2100 }, { "epoch": 3.91, "learning_rate": 4.377222222222223e-06, "loss": 0.7055, "step": 2125 }, { "epoch": 3.96, "learning_rate": 4.363333333333334e-06, "loss": 0.6449, "step": 2150 }, { "epoch": 4.01, "learning_rate": 4.349444444444445e-06, "loss": 0.6913, "step": 2175 }, { "epoch": 4.05, "learning_rate": 4.3355555555555565e-06, "loss": 0.6075, "step": 2200 }, { "epoch": 4.1, "learning_rate": 4.321666666666667e-06, "loss": 0.6183, "step": 2225 }, { "epoch": 4.14, "learning_rate": 4.307777777777778e-06, "loss": 0.6472, "step": 2250 }, { "epoch": 4.19, "learning_rate": 4.293888888888889e-06, "loss": 0.6082, "step": 2275 }, { "epoch": 4.24, "learning_rate": 4.2800000000000005e-06, "loss": 0.617, "step": 2300 }, { "epoch": 4.28, "learning_rate": 4.266111111111111e-06, "loss": 0.6224, "step": 2325 }, { "epoch": 4.33, "learning_rate": 4.2522222222222225e-06, "loss": 0.621, "step": 2350 }, { "epoch": 4.37, "learning_rate": 4.238333333333334e-06, "loss": 0.6287, "step": 2375 }, { "epoch": 4.42, "learning_rate": 4.2244444444444446e-06, "loss": 0.5887, "step": 2400 }, { "epoch": 4.47, "learning_rate": 4.210555555555556e-06, "loss": 0.6475, "step": 2425 }, { "epoch": 4.51, "learning_rate": 4.1966666666666674e-06, "loss": 0.6084, "step": 2450 }, { "epoch": 4.56, "learning_rate": 4.182777777777778e-06, "loss": 0.6165, "step": 2475 }, { "epoch": 4.6, "learning_rate": 4.168888888888889e-06, "loss": 0.6334, "step": 2500 }, { "epoch": 4.6, "eval_cer": 5.686653771760155, "eval_loss": 0.3033701181411743, "eval_runtime": 194.8371, "eval_samples_per_second": 5.132, "eval_steps_per_second": 0.323, "eval_wer": 14.791987673343607, "step": 2500 }, { "epoch": 4.65, "learning_rate": 4.155e-06, "loss": 0.6216, "step": 2525 }, { "epoch": 4.7, "learning_rate": 4.1411111111111115e-06, "loss": 0.6087, "step": 2550 }, { "epoch": 4.74, "learning_rate": 4.127222222222222e-06, "loss": 0.5892, "step": 2575 }, { "epoch": 4.79, "learning_rate": 4.1133333333333335e-06, "loss": 0.6139, "step": 2600 }, { "epoch": 4.83, "learning_rate": 4.099444444444445e-06, "loss": 0.639, "step": 2625 }, { "epoch": 4.88, "learning_rate": 4.085555555555556e-06, "loss": 0.6201, "step": 2650 }, { "epoch": 4.93, "learning_rate": 4.071666666666667e-06, "loss": 0.6029, "step": 2675 }, { "epoch": 4.97, "learning_rate": 4.057777777777778e-06, "loss": 0.6426, "step": 2700 }, { "epoch": 5.02, "learning_rate": 4.04388888888889e-06, "loss": 0.6085, "step": 2725 }, { "epoch": 5.06, "learning_rate": 4.03e-06, "loss": 0.5573, "step": 2750 }, { "epoch": 5.11, "learning_rate": 4.016111111111111e-06, "loss": 0.554, "step": 2775 }, { "epoch": 5.16, "learning_rate": 4.002222222222222e-06, "loss": 0.5427, "step": 2800 }, { "epoch": 5.2, "learning_rate": 3.988333333333334e-06, "loss": 0.5982, "step": 2825 }, { "epoch": 5.25, "learning_rate": 3.974444444444444e-06, "loss": 0.5579, "step": 2850 }, { "epoch": 5.29, "learning_rate": 3.960555555555556e-06, "loss": 0.5616, "step": 2875 }, { "epoch": 5.34, "learning_rate": 3.946666666666667e-06, "loss": 0.5577, "step": 2900 }, { "epoch": 5.39, "learning_rate": 3.932777777777778e-06, "loss": 0.5384, "step": 2925 }, { "epoch": 5.43, "learning_rate": 3.918888888888889e-06, "loss": 0.5388, "step": 2950 }, { "epoch": 5.48, "learning_rate": 3.905000000000001e-06, "loss": 0.5828, "step": 2975 }, { "epoch": 5.52, "learning_rate": 3.891111111111111e-06, "loss": 0.5746, "step": 3000 }, { "epoch": 5.52, "eval_cer": 5.639679469466703, "eval_loss": 0.3028756082057953, "eval_runtime": 194.4573, "eval_samples_per_second": 5.143, "eval_steps_per_second": 0.324, "eval_wer": 14.622496147919877, "step": 3000 }, { "epoch": 5.57, "learning_rate": 3.877222222222223e-06, "loss": 0.5889, "step": 3025 }, { "epoch": 5.62, "learning_rate": 3.863333333333333e-06, "loss": 0.5638, "step": 3050 }, { "epoch": 5.66, "learning_rate": 3.849444444444445e-06, "loss": 0.5671, "step": 3075 }, { "epoch": 5.71, "learning_rate": 3.835555555555555e-06, "loss": 0.5558, "step": 3100 }, { "epoch": 5.76, "learning_rate": 3.821666666666667e-06, "loss": 0.6015, "step": 3125 }, { "epoch": 5.8, "learning_rate": 3.8077777777777782e-06, "loss": 0.5976, "step": 3150 }, { "epoch": 5.85, "learning_rate": 3.7938888888888892e-06, "loss": 0.5681, "step": 3175 }, { "epoch": 5.89, "learning_rate": 3.7800000000000002e-06, "loss": 0.5663, "step": 3200 }, { "epoch": 5.94, "learning_rate": 3.7661111111111117e-06, "loss": 0.6237, "step": 3225 }, { "epoch": 5.99, "learning_rate": 3.7522222222222227e-06, "loss": 0.5551, "step": 3250 }, { "epoch": 6.03, "learning_rate": 3.7383333333333337e-06, "loss": 0.5262, "step": 3275 }, { "epoch": 6.08, "learning_rate": 3.724444444444445e-06, "loss": 0.5251, "step": 3300 }, { "epoch": 6.12, "learning_rate": 3.710555555555556e-06, "loss": 0.5494, "step": 3325 }, { "epoch": 6.17, "learning_rate": 3.6966666666666667e-06, "loss": 0.5059, "step": 3350 }, { "epoch": 6.22, "learning_rate": 3.6827777777777777e-06, "loss": 0.5413, "step": 3375 }, { "epoch": 6.26, "learning_rate": 3.668888888888889e-06, "loss": 0.5153, "step": 3400 }, { "epoch": 6.31, "learning_rate": 3.655e-06, "loss": 0.5056, "step": 3425 }, { "epoch": 6.35, "learning_rate": 3.641111111111111e-06, "loss": 0.5541, "step": 3450 }, { "epoch": 6.4, "learning_rate": 3.6272222222222226e-06, "loss": 0.5282, "step": 3475 }, { "epoch": 6.45, "learning_rate": 3.6133333333333336e-06, "loss": 0.5359, "step": 3500 }, { "epoch": 6.45, "eval_cer": 5.578889195910473, "eval_loss": 0.30176836252212524, "eval_runtime": 195.0766, "eval_samples_per_second": 5.126, "eval_steps_per_second": 0.323, "eval_wer": 14.48382126348228, "step": 3500 }, { "epoch": 6.49, "learning_rate": 3.599444444444445e-06, "loss": 0.5223, "step": 3525 }, { "epoch": 6.54, "learning_rate": 3.585555555555556e-06, "loss": 0.5334, "step": 3550 }, { "epoch": 6.58, "learning_rate": 3.571666666666667e-06, "loss": 0.5108, "step": 3575 }, { "epoch": 6.63, "learning_rate": 3.5577777777777785e-06, "loss": 0.5279, "step": 3600 }, { "epoch": 6.68, "learning_rate": 3.543888888888889e-06, "loss": 0.5001, "step": 3625 }, { "epoch": 6.72, "learning_rate": 3.53e-06, "loss": 0.5107, "step": 3650 }, { "epoch": 6.77, "learning_rate": 3.516111111111111e-06, "loss": 0.5466, "step": 3675 }, { "epoch": 6.81, "learning_rate": 3.5022222222222225e-06, "loss": 0.5232, "step": 3700 }, { "epoch": 6.86, "learning_rate": 3.4883333333333335e-06, "loss": 0.5042, "step": 3725 }, { "epoch": 6.91, "learning_rate": 3.4744444444444445e-06, "loss": 0.5325, "step": 3750 }, { "epoch": 6.95, "learning_rate": 3.460555555555556e-06, "loss": 0.5466, "step": 3775 }, { "epoch": 7.0, "learning_rate": 3.446666666666667e-06, "loss": 0.5299, "step": 3800 }, { "epoch": 7.04, "learning_rate": 3.4327777777777784e-06, "loss": 0.4977, "step": 3825 }, { "epoch": 7.09, "learning_rate": 3.4188888888888894e-06, "loss": 0.4844, "step": 3850 }, { "epoch": 7.14, "learning_rate": 3.4050000000000004e-06, "loss": 0.4878, "step": 3875 }, { "epoch": 7.18, "learning_rate": 3.391111111111111e-06, "loss": 0.4973, "step": 3900 }, { "epoch": 7.23, "learning_rate": 3.3772222222222225e-06, "loss": 0.5158, "step": 3925 }, { "epoch": 7.27, "learning_rate": 3.3633333333333335e-06, "loss": 0.5247, "step": 3950 }, { "epoch": 7.32, "learning_rate": 3.3494444444444445e-06, "loss": 0.4648, "step": 3975 }, { "epoch": 7.37, "learning_rate": 3.335555555555556e-06, "loss": 0.5058, "step": 4000 }, { "epoch": 7.37, "eval_cer": 5.683890577507598, "eval_loss": 0.30099615454673767, "eval_runtime": 195.0615, "eval_samples_per_second": 5.127, "eval_steps_per_second": 0.323, "eval_wer": 14.591679506933744, "step": 4000 }, { "epoch": 7.41, "learning_rate": 3.321666666666667e-06, "loss": 0.4619, "step": 4025 }, { "epoch": 7.46, "learning_rate": 3.307777777777778e-06, "loss": 0.4968, "step": 4050 }, { "epoch": 7.5, "learning_rate": 3.2938888888888894e-06, "loss": 0.4982, "step": 4075 }, { "epoch": 7.55, "learning_rate": 3.2800000000000004e-06, "loss": 0.5049, "step": 4100 }, { "epoch": 7.6, "learning_rate": 3.2661111111111114e-06, "loss": 0.4799, "step": 4125 }, { "epoch": 7.64, "learning_rate": 3.252222222222223e-06, "loss": 0.4756, "step": 4150 }, { "epoch": 7.69, "learning_rate": 3.2383333333333334e-06, "loss": 0.4924, "step": 4175 }, { "epoch": 7.73, "learning_rate": 3.2244444444444444e-06, "loss": 0.5195, "step": 4200 }, { "epoch": 7.78, "learning_rate": 3.210555555555556e-06, "loss": 0.4897, "step": 4225 }, { "epoch": 7.83, "learning_rate": 3.196666666666667e-06, "loss": 0.486, "step": 4250 }, { "epoch": 7.87, "learning_rate": 3.182777777777778e-06, "loss": 0.4907, "step": 4275 }, { "epoch": 7.92, "learning_rate": 3.1688888888888893e-06, "loss": 0.4874, "step": 4300 }, { "epoch": 7.97, "learning_rate": 3.1550000000000003e-06, "loss": 0.4948, "step": 4325 }, { "epoch": 8.01, "learning_rate": 3.1411111111111113e-06, "loss": 0.5243, "step": 4350 }, { "epoch": 8.06, "learning_rate": 3.1272222222222227e-06, "loss": 0.4634, "step": 4375 }, { "epoch": 8.1, "learning_rate": 3.1133333333333337e-06, "loss": 0.4867, "step": 4400 }, { "epoch": 8.15, "learning_rate": 3.0994444444444447e-06, "loss": 0.461, "step": 4425 }, { "epoch": 8.2, "learning_rate": 3.085555555555556e-06, "loss": 0.4692, "step": 4450 }, { "epoch": 8.24, "learning_rate": 3.0716666666666668e-06, "loss": 0.4392, "step": 4475 }, { "epoch": 8.29, "learning_rate": 3.0577777777777778e-06, "loss": 0.4833, "step": 4500 }, { "epoch": 8.29, "eval_cer": 5.523625310859353, "eval_loss": 0.302284836769104, "eval_runtime": 194.3049, "eval_samples_per_second": 5.147, "eval_steps_per_second": 0.324, "eval_wer": 14.237288135593221, "step": 4500 }, { "epoch": 8.33, "learning_rate": 3.043888888888889e-06, "loss": 0.4612, "step": 4525 }, { "epoch": 8.38, "learning_rate": 3.0300000000000002e-06, "loss": 0.4538, "step": 4550 }, { "epoch": 8.43, "learning_rate": 3.0161111111111112e-06, "loss": 0.4591, "step": 4575 }, { "epoch": 8.47, "learning_rate": 3.0022222222222227e-06, "loss": 0.4643, "step": 4600 }, { "epoch": 8.52, "learning_rate": 2.9883333333333337e-06, "loss": 0.47, "step": 4625 }, { "epoch": 8.56, "learning_rate": 2.9744444444444447e-06, "loss": 0.4553, "step": 4650 }, { "epoch": 8.61, "learning_rate": 2.960555555555556e-06, "loss": 0.4579, "step": 4675 }, { "epoch": 8.66, "learning_rate": 2.946666666666667e-06, "loss": 0.4745, "step": 4700 }, { "epoch": 8.7, "learning_rate": 2.932777777777778e-06, "loss": 0.4729, "step": 4725 }, { "epoch": 8.75, "learning_rate": 2.9188888888888887e-06, "loss": 0.4767, "step": 4750 }, { "epoch": 8.79, "learning_rate": 2.905e-06, "loss": 0.46, "step": 4775 }, { "epoch": 8.84, "learning_rate": 2.8916666666666672e-06, "loss": 0.4952, "step": 4800 }, { "epoch": 8.89, "learning_rate": 2.8777777777777782e-06, "loss": 0.4839, "step": 4825 }, { "epoch": 8.93, "learning_rate": 2.8638888888888892e-06, "loss": 0.4707, "step": 4850 }, { "epoch": 8.98, "learning_rate": 2.85e-06, "loss": 0.4619, "step": 4875 }, { "epoch": 9.02, "learning_rate": 2.8361111111111113e-06, "loss": 0.4389, "step": 4900 }, { "epoch": 9.07, "learning_rate": 2.8222222222222223e-06, "loss": 0.4506, "step": 4925 }, { "epoch": 9.12, "learning_rate": 2.8083333333333333e-06, "loss": 0.4349, "step": 4950 }, { "epoch": 9.16, "learning_rate": 2.7944444444444447e-06, "loss": 0.4476, "step": 4975 }, { "epoch": 9.21, "learning_rate": 2.7805555555555557e-06, "loss": 0.4398, "step": 5000 }, { "epoch": 9.21, "eval_cer": 5.584415584415584, "eval_loss": 0.30045086145401, "eval_runtime": 194.9477, "eval_samples_per_second": 5.13, "eval_steps_per_second": 0.323, "eval_wer": 14.422187981510016, "step": 5000 }, { "epoch": 9.25, "learning_rate": 2.766666666666667e-06, "loss": 0.4335, "step": 5025 }, { "epoch": 9.3, "learning_rate": 2.752777777777778e-06, "loss": 0.4271, "step": 5050 }, { "epoch": 9.35, "learning_rate": 2.738888888888889e-06, "loss": 0.436, "step": 5075 }, { "epoch": 9.39, "learning_rate": 2.7250000000000006e-06, "loss": 0.4453, "step": 5100 }, { "epoch": 9.44, "learning_rate": 2.7111111111111116e-06, "loss": 0.4268, "step": 5125 }, { "epoch": 9.48, "learning_rate": 2.697222222222222e-06, "loss": 0.4433, "step": 5150 }, { "epoch": 9.53, "learning_rate": 2.683333333333333e-06, "loss": 0.4383, "step": 5175 }, { "epoch": 9.58, "learning_rate": 2.6694444444444446e-06, "loss": 0.4435, "step": 5200 }, { "epoch": 9.62, "learning_rate": 2.6555555555555556e-06, "loss": 0.4233, "step": 5225 }, { "epoch": 9.67, "learning_rate": 2.6416666666666666e-06, "loss": 0.4409, "step": 5250 }, { "epoch": 9.71, "learning_rate": 2.627777777777778e-06, "loss": 0.4442, "step": 5275 }, { "epoch": 9.76, "learning_rate": 2.613888888888889e-06, "loss": 0.422, "step": 5300 }, { "epoch": 9.81, "learning_rate": 2.6e-06, "loss": 0.4539, "step": 5325 }, { "epoch": 9.85, "learning_rate": 2.5861111111111115e-06, "loss": 0.4397, "step": 5350 }, { "epoch": 9.9, "learning_rate": 2.5722222222222225e-06, "loss": 0.4493, "step": 5375 }, { "epoch": 9.94, "learning_rate": 2.558333333333334e-06, "loss": 0.4273, "step": 5400 }, { "epoch": 9.99, "learning_rate": 2.5444444444444446e-06, "loss": 0.4391, "step": 5425 }, { "epoch": 10.04, "learning_rate": 2.5305555555555556e-06, "loss": 0.4174, "step": 5450 }, { "epoch": 10.08, "learning_rate": 2.5166666666666666e-06, "loss": 0.4112, "step": 5475 }, { "epoch": 10.13, "learning_rate": 2.502777777777778e-06, "loss": 0.4359, "step": 5500 }, { "epoch": 10.13, "eval_cer": 5.625863498203923, "eval_loss": 0.29985830187797546, "eval_runtime": 195.24, "eval_samples_per_second": 5.122, "eval_steps_per_second": 0.323, "eval_wer": 14.48382126348228, "step": 5500 }, { "epoch": 10.17, "learning_rate": 2.488888888888889e-06, "loss": 0.4273, "step": 5525 }, { "epoch": 10.22, "learning_rate": 2.475e-06, "loss": 0.4231, "step": 5550 }, { "epoch": 10.27, "learning_rate": 2.4611111111111115e-06, "loss": 0.4263, "step": 5575 }, { "epoch": 10.31, "learning_rate": 2.4472222222222225e-06, "loss": 0.4026, "step": 5600 }, { "epoch": 10.36, "learning_rate": 2.4333333333333335e-06, "loss": 0.4191, "step": 5625 }, { "epoch": 10.41, "learning_rate": 2.4194444444444445e-06, "loss": 0.4029, "step": 5650 }, { "epoch": 10.45, "learning_rate": 2.4055555555555555e-06, "loss": 0.3939, "step": 5675 }, { "epoch": 10.5, "learning_rate": 2.391666666666667e-06, "loss": 0.4283, "step": 5700 }, { "epoch": 10.54, "learning_rate": 2.377777777777778e-06, "loss": 0.4169, "step": 5725 }, { "epoch": 10.59, "learning_rate": 2.3638888888888894e-06, "loss": 0.4069, "step": 5750 }, { "epoch": 10.64, "learning_rate": 2.35e-06, "loss": 0.4227, "step": 5775 }, { "epoch": 10.68, "learning_rate": 2.3361111111111114e-06, "loss": 0.4175, "step": 5800 }, { "epoch": 10.73, "learning_rate": 2.3222222222222224e-06, "loss": 0.4264, "step": 5825 }, { "epoch": 10.77, "learning_rate": 2.3083333333333334e-06, "loss": 0.4331, "step": 5850 }, { "epoch": 10.82, "learning_rate": 2.294444444444445e-06, "loss": 0.4301, "step": 5875 }, { "epoch": 10.87, "learning_rate": 2.280555555555556e-06, "loss": 0.4209, "step": 5900 }, { "epoch": 10.91, "learning_rate": 2.266666666666667e-06, "loss": 0.4345, "step": 5925 }, { "epoch": 10.96, "learning_rate": 2.252777777777778e-06, "loss": 0.4092, "step": 5950 }, { "epoch": 11.0, "learning_rate": 2.238888888888889e-06, "loss": 0.4078, "step": 5975 }, { "epoch": 11.05, "learning_rate": 2.2250000000000003e-06, "loss": 0.4036, "step": 6000 }, { "epoch": 11.05, "eval_cer": 5.562310030395136, "eval_loss": 0.299458771944046, "eval_runtime": 195.4099, "eval_samples_per_second": 5.117, "eval_steps_per_second": 0.322, "eval_wer": 14.28351309707242, "step": 6000 }, { "epoch": 11.05, "step": 6000, "total_flos": 5.540839686144e+19, "train_loss": 0.6941633464495341, "train_runtime": 11106.8839, "train_samples_per_second": 28.811, "train_steps_per_second": 0.9 } ], "max_steps": 10000, "num_train_epochs": 19, "total_flos": 5.540839686144e+19, "trial_name": null, "trial_params": null }