{ "best_metric": 0.06361353581250748, "best_model_checkpoint": "./wav2vec2-xls-r-300m-ar/checkpoint-10000", "epoch": 88.49557522123894, "global_step": 10000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.22, "learning_rate": 1.4999999999999999e-05, "loss": 9.1248, "step": 25 }, { "epoch": 0.44, "learning_rate": 2.9999999999999997e-05, "loss": 6.6739, "step": 50 }, { "epoch": 0.66, "learning_rate": 4.4999999999999996e-05, "loss": 4.632, "step": 75 }, { "epoch": 0.88, "learning_rate": 5.9999999999999995e-05, "loss": 3.917, "step": 100 }, { "epoch": 1.11, "learning_rate": 7.5e-05, "loss": 3.3449, "step": 125 }, { "epoch": 1.33, "learning_rate": 8.999999999999999e-05, "loss": 2.8486, "step": 150 }, { "epoch": 1.55, "learning_rate": 0.00010499999999999999, "loss": 2.5717, "step": 175 }, { "epoch": 1.77, "learning_rate": 0.00011999999999999999, "loss": 2.3746, "step": 200 }, { "epoch": 1.99, "learning_rate": 0.000135, "loss": 2.2744, "step": 225 }, { "epoch": 2.21, "learning_rate": 0.00015, "loss": 2.0759, "step": 250 }, { "epoch": 2.43, "learning_rate": 0.000165, "loss": 1.7813, "step": 275 }, { "epoch": 2.65, "learning_rate": 0.00017999999999999998, "loss": 1.2002, "step": 300 }, { "epoch": 2.88, "learning_rate": 0.000195, "loss": 0.7684, "step": 325 }, { "epoch": 3.1, "learning_rate": 0.00020999999999999998, "loss": 0.527, "step": 350 }, { "epoch": 3.32, "learning_rate": 0.000225, "loss": 0.3734, "step": 375 }, { "epoch": 3.54, "learning_rate": 0.00023999999999999998, "loss": 0.3184, "step": 400 }, { "epoch": 3.76, "learning_rate": 0.00025499999999999996, "loss": 0.2822, "step": 425 }, { "epoch": 3.98, "learning_rate": 0.00027, "loss": 0.2074, "step": 450 }, { "epoch": 4.2, "learning_rate": 0.000285, "loss": 0.1985, "step": 475 }, { "epoch": 4.42, "learning_rate": 0.0003, "loss": 0.1917, "step": 500 }, { "epoch": 4.65, "learning_rate": 0.00029921052631578946, "loss": 0.1853, "step": 525 }, { "epoch": 4.87, "learning_rate": 0.00029842105263157894, "loss": 0.1828, "step": 550 }, { "epoch": 5.09, "learning_rate": 0.00029763157894736837, "loss": 0.1715, "step": 575 }, { "epoch": 5.31, "learning_rate": 0.00029684210526315785, "loss": 0.1457, "step": 600 }, { "epoch": 5.53, "learning_rate": 0.00029605263157894733, "loss": 0.1247, "step": 625 }, { "epoch": 5.75, "learning_rate": 0.0002952631578947368, "loss": 0.1266, "step": 650 }, { "epoch": 5.97, "learning_rate": 0.0002944736842105263, "loss": 0.117, "step": 675 }, { "epoch": 6.19, "learning_rate": 0.0002936842105263158, "loss": 0.117, "step": 700 }, { "epoch": 6.42, "learning_rate": 0.00029289473684210527, "loss": 0.125, "step": 725 }, { "epoch": 6.64, "learning_rate": 0.0002921052631578947, "loss": 0.1084, "step": 750 }, { "epoch": 6.86, "learning_rate": 0.0002913157894736842, "loss": 0.1178, "step": 775 }, { "epoch": 7.08, "learning_rate": 0.00029052631578947366, "loss": 0.1193, "step": 800 }, { "epoch": 7.3, "learning_rate": 0.00028973684210526314, "loss": 0.0925, "step": 825 }, { "epoch": 7.52, "learning_rate": 0.00028894736842105263, "loss": 0.085, "step": 850 }, { "epoch": 7.74, "learning_rate": 0.00028815789473684206, "loss": 0.0865, "step": 875 }, { "epoch": 7.96, "learning_rate": 0.00028736842105263154, "loss": 0.0811, "step": 900 }, { "epoch": 8.19, "learning_rate": 0.000286578947368421, "loss": 0.0835, "step": 925 }, { "epoch": 8.41, "learning_rate": 0.0002857894736842105, "loss": 0.0886, "step": 950 }, { "epoch": 8.63, "learning_rate": 0.000285, "loss": 0.0991, "step": 975 }, { "epoch": 8.85, "learning_rate": 0.0002842105263157894, "loss": 0.0793, "step": 1000 }, { "epoch": 8.85, "eval_loss": 0.16262491047382355, "eval_runtime": 11.6481, "eval_samples_per_second": 8.585, "eval_steps_per_second": 1.116, "eval_wer": 0.07856032524213799, "step": 1000 }, { "epoch": 9.07, "learning_rate": 0.00028342105263157895, "loss": 0.0863, "step": 1025 }, { "epoch": 9.29, "learning_rate": 0.0002826315789473684, "loss": 0.0638, "step": 1050 }, { "epoch": 9.51, "learning_rate": 0.00028184210526315787, "loss": 0.0604, "step": 1075 }, { "epoch": 9.73, "learning_rate": 0.00028105263157894735, "loss": 0.0621, "step": 1100 }, { "epoch": 9.96, "learning_rate": 0.00028026315789473683, "loss": 0.0707, "step": 1125 }, { "epoch": 10.18, "learning_rate": 0.0002794736842105263, "loss": 0.0799, "step": 1150 }, { "epoch": 10.4, "learning_rate": 0.00027868421052631574, "loss": 0.0694, "step": 1175 }, { "epoch": 10.62, "learning_rate": 0.00027789473684210523, "loss": 0.0655, "step": 1200 }, { "epoch": 10.84, "learning_rate": 0.0002771052631578947, "loss": 0.0776, "step": 1225 }, { "epoch": 11.06, "learning_rate": 0.0002763157894736842, "loss": 0.0582, "step": 1250 }, { "epoch": 11.28, "learning_rate": 0.0002755263157894737, "loss": 0.0547, "step": 1275 }, { "epoch": 11.5, "learning_rate": 0.0002747368421052631, "loss": 0.0597, "step": 1300 }, { "epoch": 11.73, "learning_rate": 0.0002739473684210526, "loss": 0.0492, "step": 1325 }, { "epoch": 11.95, "learning_rate": 0.00027315789473684207, "loss": 0.0699, "step": 1350 }, { "epoch": 12.17, "learning_rate": 0.00027236842105263155, "loss": 0.0594, "step": 1375 }, { "epoch": 12.39, "learning_rate": 0.00027157894736842104, "loss": 0.0541, "step": 1400 }, { "epoch": 12.61, "learning_rate": 0.0002707894736842105, "loss": 0.0522, "step": 1425 }, { "epoch": 12.83, "learning_rate": 0.00027, "loss": 0.0645, "step": 1450 }, { "epoch": 13.05, "learning_rate": 0.00026921052631578943, "loss": 0.0583, "step": 1475 }, { "epoch": 13.27, "learning_rate": 0.0002684210526315789, "loss": 0.0443, "step": 1500 }, { "epoch": 13.5, "learning_rate": 0.0002676315789473684, "loss": 0.0548, "step": 1525 }, { "epoch": 13.72, "learning_rate": 0.0002668421052631579, "loss": 0.0484, "step": 1550 }, { "epoch": 13.94, "learning_rate": 0.00026605263157894736, "loss": 0.0775, "step": 1575 }, { "epoch": 14.16, "learning_rate": 0.0002652631578947368, "loss": 0.0554, "step": 1600 }, { "epoch": 14.38, "learning_rate": 0.0002644736842105263, "loss": 0.0456, "step": 1625 }, { "epoch": 14.6, "learning_rate": 0.00026368421052631576, "loss": 0.0563, "step": 1650 }, { "epoch": 14.82, "learning_rate": 0.00026289473684210524, "loss": 0.0493, "step": 1675 }, { "epoch": 15.04, "learning_rate": 0.0002621052631578947, "loss": 0.0452, "step": 1700 }, { "epoch": 15.27, "learning_rate": 0.0002613157894736842, "loss": 0.0389, "step": 1725 }, { "epoch": 15.49, "learning_rate": 0.0002605263157894737, "loss": 0.0436, "step": 1750 }, { "epoch": 15.71, "learning_rate": 0.0002597368421052631, "loss": 0.0407, "step": 1775 }, { "epoch": 15.93, "learning_rate": 0.0002589473684210526, "loss": 0.0448, "step": 1800 }, { "epoch": 16.15, "learning_rate": 0.0002581578947368421, "loss": 0.0457, "step": 1825 }, { "epoch": 16.37, "learning_rate": 0.00025736842105263157, "loss": 0.0461, "step": 1850 }, { "epoch": 16.59, "learning_rate": 0.00025657894736842105, "loss": 0.0452, "step": 1875 }, { "epoch": 16.81, "learning_rate": 0.0002557894736842105, "loss": 0.0484, "step": 1900 }, { "epoch": 17.04, "learning_rate": 0.00025499999999999996, "loss": 0.0376, "step": 1925 }, { "epoch": 17.26, "learning_rate": 0.00025421052631578945, "loss": 0.0409, "step": 1950 }, { "epoch": 17.48, "learning_rate": 0.00025342105263157893, "loss": 0.0315, "step": 1975 }, { "epoch": 17.7, "learning_rate": 0.00025263157894736836, "loss": 0.0396, "step": 2000 }, { "epoch": 17.7, "eval_loss": 0.21985749900341034, "eval_runtime": 11.0577, "eval_samples_per_second": 9.043, "eval_steps_per_second": 1.176, "eval_wer": 0.08071266292000479, "step": 2000 }, { "epoch": 17.92, "learning_rate": 0.00025184210526315784, "loss": 0.0406, "step": 2025 }, { "epoch": 18.14, "learning_rate": 0.0002510526315789474, "loss": 0.0322, "step": 2050 }, { "epoch": 18.36, "learning_rate": 0.0002502631578947368, "loss": 0.0393, "step": 2075 }, { "epoch": 18.58, "learning_rate": 0.0002494736842105263, "loss": 0.0438, "step": 2100 }, { "epoch": 18.81, "learning_rate": 0.0002486842105263158, "loss": 0.0372, "step": 2125 }, { "epoch": 19.03, "learning_rate": 0.00024789473684210526, "loss": 0.0335, "step": 2150 }, { "epoch": 19.25, "learning_rate": 0.00024710526315789474, "loss": 0.0327, "step": 2175 }, { "epoch": 19.47, "learning_rate": 0.00024631578947368417, "loss": 0.0307, "step": 2200 }, { "epoch": 19.69, "learning_rate": 0.00024552631578947365, "loss": 0.0374, "step": 2225 }, { "epoch": 19.91, "learning_rate": 0.00024473684210526314, "loss": 0.035, "step": 2250 }, { "epoch": 20.13, "learning_rate": 0.0002439473684210526, "loss": 0.0336, "step": 2275 }, { "epoch": 20.35, "learning_rate": 0.00024315789473684207, "loss": 0.0426, "step": 2300 }, { "epoch": 20.58, "learning_rate": 0.00024236842105263156, "loss": 0.0375, "step": 2325 }, { "epoch": 20.8, "learning_rate": 0.000241578947368421, "loss": 0.0352, "step": 2350 }, { "epoch": 21.02, "learning_rate": 0.00024078947368421052, "loss": 0.0319, "step": 2375 }, { "epoch": 21.24, "learning_rate": 0.00023999999999999998, "loss": 0.0339, "step": 2400 }, { "epoch": 21.46, "learning_rate": 0.00023921052631578946, "loss": 0.0335, "step": 2425 }, { "epoch": 21.68, "learning_rate": 0.00023842105263157895, "loss": 0.0329, "step": 2450 }, { "epoch": 21.9, "learning_rate": 0.0002376315789473684, "loss": 0.0373, "step": 2475 }, { "epoch": 22.12, "learning_rate": 0.00023684210526315788, "loss": 0.0371, "step": 2500 }, { "epoch": 22.35, "learning_rate": 0.00023605263157894734, "loss": 0.0273, "step": 2525 }, { "epoch": 22.57, "learning_rate": 0.00023526315789473682, "loss": 0.0361, "step": 2550 }, { "epoch": 22.79, "learning_rate": 0.00023447368421052628, "loss": 0.0367, "step": 2575 }, { "epoch": 23.01, "learning_rate": 0.00023368421052631576, "loss": 0.0374, "step": 2600 }, { "epoch": 23.23, "learning_rate": 0.00023289473684210524, "loss": 0.0264, "step": 2625 }, { "epoch": 23.45, "learning_rate": 0.0002321052631578947, "loss": 0.0331, "step": 2650 }, { "epoch": 23.67, "learning_rate": 0.00023131578947368418, "loss": 0.0407, "step": 2675 }, { "epoch": 23.89, "learning_rate": 0.00023052631578947364, "loss": 0.031, "step": 2700 }, { "epoch": 24.12, "learning_rate": 0.00022973684210526315, "loss": 0.0294, "step": 2725 }, { "epoch": 24.34, "learning_rate": 0.00022894736842105263, "loss": 0.0345, "step": 2750 }, { "epoch": 24.56, "learning_rate": 0.0002281578947368421, "loss": 0.027, "step": 2775 }, { "epoch": 24.78, "learning_rate": 0.00022736842105263157, "loss": 0.0241, "step": 2800 }, { "epoch": 25.0, "learning_rate": 0.00022657894736842103, "loss": 0.0304, "step": 2825 }, { "epoch": 25.22, "learning_rate": 0.0002257894736842105, "loss": 0.026, "step": 2850 }, { "epoch": 25.44, "learning_rate": 0.000225, "loss": 0.0325, "step": 2875 }, { "epoch": 25.66, "learning_rate": 0.00022421052631578945, "loss": 0.0374, "step": 2900 }, { "epoch": 25.88, "learning_rate": 0.00022342105263157893, "loss": 0.036, "step": 2925 }, { "epoch": 26.11, "learning_rate": 0.0002226315789473684, "loss": 0.0339, "step": 2950 }, { "epoch": 26.33, "learning_rate": 0.00022184210526315787, "loss": 0.0312, "step": 2975 }, { "epoch": 26.55, "learning_rate": 0.00022105263157894733, "loss": 0.0285, "step": 3000 }, { "epoch": 26.55, "eval_loss": 0.22891539335250854, "eval_runtime": 10.8646, "eval_samples_per_second": 9.204, "eval_steps_per_second": 1.197, "eval_wer": 0.0693531029534856, "step": 3000 }, { "epoch": 26.77, "learning_rate": 0.0002202631578947368, "loss": 0.0238, "step": 3025 }, { "epoch": 26.99, "learning_rate": 0.00021947368421052632, "loss": 0.0226, "step": 3050 }, { "epoch": 27.21, "learning_rate": 0.00021868421052631578, "loss": 0.0255, "step": 3075 }, { "epoch": 27.43, "learning_rate": 0.00021789473684210526, "loss": 0.0326, "step": 3100 }, { "epoch": 27.65, "learning_rate": 0.00021710526315789472, "loss": 0.0629, "step": 3125 }, { "epoch": 27.88, "learning_rate": 0.0002163157894736842, "loss": 0.0348, "step": 3150 }, { "epoch": 28.1, "learning_rate": 0.00021552631578947368, "loss": 0.0311, "step": 3175 }, { "epoch": 28.32, "learning_rate": 0.00021473684210526314, "loss": 0.023, "step": 3200 }, { "epoch": 28.54, "learning_rate": 0.00021394736842105262, "loss": 0.0214, "step": 3225 }, { "epoch": 28.76, "learning_rate": 0.00021315789473684208, "loss": 0.0201, "step": 3250 }, { "epoch": 28.98, "learning_rate": 0.00021236842105263156, "loss": 0.0202, "step": 3275 }, { "epoch": 29.2, "learning_rate": 0.00021157894736842102, "loss": 0.0231, "step": 3300 }, { "epoch": 29.42, "learning_rate": 0.0002107894736842105, "loss": 0.0291, "step": 3325 }, { "epoch": 29.65, "learning_rate": 0.00020999999999999998, "loss": 0.0241, "step": 3350 }, { "epoch": 29.87, "learning_rate": 0.00020921052631578944, "loss": 0.0261, "step": 3375 }, { "epoch": 30.09, "learning_rate": 0.00020842105263157895, "loss": 0.0281, "step": 3400 }, { "epoch": 30.31, "learning_rate": 0.0002076315789473684, "loss": 0.0237, "step": 3425 }, { "epoch": 30.53, "learning_rate": 0.0002068421052631579, "loss": 0.0199, "step": 3450 }, { "epoch": 30.75, "learning_rate": 0.00020605263157894737, "loss": 0.0199, "step": 3475 }, { "epoch": 30.97, "learning_rate": 0.00020526315789473683, "loss": 0.0207, "step": 3500 }, { "epoch": 31.19, "learning_rate": 0.0002044736842105263, "loss": 0.0221, "step": 3525 }, { "epoch": 31.42, "learning_rate": 0.00020368421052631576, "loss": 0.0234, "step": 3550 }, { "epoch": 31.64, "learning_rate": 0.00020289473684210525, "loss": 0.025, "step": 3575 }, { "epoch": 31.86, "learning_rate": 0.0002021052631578947, "loss": 0.0251, "step": 3600 }, { "epoch": 32.08, "learning_rate": 0.0002013157894736842, "loss": 0.0232, "step": 3625 }, { "epoch": 32.3, "learning_rate": 0.00020052631578947367, "loss": 0.0215, "step": 3650 }, { "epoch": 32.52, "learning_rate": 0.00019973684210526313, "loss": 0.019, "step": 3675 }, { "epoch": 32.74, "learning_rate": 0.0001989473684210526, "loss": 0.0215, "step": 3700 }, { "epoch": 32.96, "learning_rate": 0.0001981578947368421, "loss": 0.0209, "step": 3725 }, { "epoch": 33.19, "learning_rate": 0.00019736842105263157, "loss": 0.0258, "step": 3750 }, { "epoch": 33.41, "learning_rate": 0.00019657894736842106, "loss": 0.0206, "step": 3775 }, { "epoch": 33.63, "learning_rate": 0.0001957894736842105, "loss": 0.019, "step": 3800 }, { "epoch": 33.85, "learning_rate": 0.000195, "loss": 0.0205, "step": 3825 }, { "epoch": 34.07, "learning_rate": 0.00019421052631578945, "loss": 0.0243, "step": 3850 }, { "epoch": 34.29, "learning_rate": 0.00019342105263157894, "loss": 0.0166, "step": 3875 }, { "epoch": 34.51, "learning_rate": 0.0001926315789473684, "loss": 0.0159, "step": 3900 }, { "epoch": 34.73, "learning_rate": 0.00019184210526315787, "loss": 0.0194, "step": 3925 }, { "epoch": 34.96, "learning_rate": 0.00019105263157894736, "loss": 0.0207, "step": 3950 }, { "epoch": 35.18, "learning_rate": 0.0001902631578947368, "loss": 0.0212, "step": 3975 }, { "epoch": 35.4, "learning_rate": 0.0001894736842105263, "loss": 0.021, "step": 4000 }, { "epoch": 35.4, "eval_loss": 0.26615452766418457, "eval_runtime": 10.7513, "eval_samples_per_second": 9.301, "eval_steps_per_second": 1.209, "eval_wer": 0.07222288652397466, "step": 4000 }, { "epoch": 35.62, "learning_rate": 0.00018868421052631575, "loss": 0.0222, "step": 4025 }, { "epoch": 35.84, "learning_rate": 0.00018789473684210524, "loss": 0.0206, "step": 4050 }, { "epoch": 36.06, "learning_rate": 0.00018710526315789475, "loss": 0.0196, "step": 4075 }, { "epoch": 36.28, "learning_rate": 0.0001863157894736842, "loss": 0.0186, "step": 4100 }, { "epoch": 36.5, "learning_rate": 0.00018552631578947368, "loss": 0.0178, "step": 4125 }, { "epoch": 36.73, "learning_rate": 0.00018473684210526314, "loss": 0.0159, "step": 4150 }, { "epoch": 36.95, "learning_rate": 0.00018394736842105262, "loss": 0.0249, "step": 4175 }, { "epoch": 37.17, "learning_rate": 0.00018315789473684208, "loss": 0.0258, "step": 4200 }, { "epoch": 37.39, "learning_rate": 0.00018236842105263156, "loss": 0.0202, "step": 4225 }, { "epoch": 37.61, "learning_rate": 0.00018157894736842105, "loss": 0.022, "step": 4250 }, { "epoch": 37.83, "learning_rate": 0.0001807894736842105, "loss": 0.0266, "step": 4275 }, { "epoch": 38.05, "learning_rate": 0.00017999999999999998, "loss": 0.0183, "step": 4300 }, { "epoch": 38.27, "learning_rate": 0.00017921052631578944, "loss": 0.0159, "step": 4325 }, { "epoch": 38.5, "learning_rate": 0.00017842105263157892, "loss": 0.0167, "step": 4350 }, { "epoch": 38.72, "learning_rate": 0.00017763157894736838, "loss": 0.0163, "step": 4375 }, { "epoch": 38.94, "learning_rate": 0.00017684210526315786, "loss": 0.0151, "step": 4400 }, { "epoch": 39.16, "learning_rate": 0.00017605263157894737, "loss": 0.0172, "step": 4425 }, { "epoch": 39.38, "learning_rate": 0.00017526315789473683, "loss": 0.0175, "step": 4450 }, { "epoch": 39.6, "learning_rate": 0.0001744736842105263, "loss": 0.0193, "step": 4475 }, { "epoch": 39.82, "learning_rate": 0.0001736842105263158, "loss": 0.0241, "step": 4500 }, { "epoch": 40.04, "learning_rate": 0.00017289473684210525, "loss": 0.0192, "step": 4525 }, { "epoch": 40.27, "learning_rate": 0.00017210526315789473, "loss": 0.019, "step": 4550 }, { "epoch": 40.49, "learning_rate": 0.0001713157894736842, "loss": 0.0138, "step": 4575 }, { "epoch": 40.71, "learning_rate": 0.00017052631578947367, "loss": 0.0158, "step": 4600 }, { "epoch": 40.93, "learning_rate": 0.00016973684210526313, "loss": 0.0181, "step": 4625 }, { "epoch": 41.15, "learning_rate": 0.0001689473684210526, "loss": 0.0167, "step": 4650 }, { "epoch": 41.37, "learning_rate": 0.00016815789473684207, "loss": 0.0214, "step": 4675 }, { "epoch": 41.59, "learning_rate": 0.00016736842105263155, "loss": 0.0184, "step": 4700 }, { "epoch": 41.81, "learning_rate": 0.00016657894736842103, "loss": 0.0166, "step": 4725 }, { "epoch": 42.04, "learning_rate": 0.00016578947368421052, "loss": 0.0128, "step": 4750 }, { "epoch": 42.26, "learning_rate": 0.000165, "loss": 0.0156, "step": 4775 }, { "epoch": 42.48, "learning_rate": 0.00016421052631578948, "loss": 0.0131, "step": 4800 }, { "epoch": 42.7, "learning_rate": 0.00016342105263157894, "loss": 0.0146, "step": 4825 }, { "epoch": 42.92, "learning_rate": 0.00016263157894736842, "loss": 0.0157, "step": 4850 }, { "epoch": 43.14, "learning_rate": 0.00016184210526315788, "loss": 0.0158, "step": 4875 }, { "epoch": 43.36, "learning_rate": 0.00016105263157894736, "loss": 0.0157, "step": 4900 }, { "epoch": 43.58, "learning_rate": 0.00016026315789473682, "loss": 0.0159, "step": 4925 }, { "epoch": 43.81, "learning_rate": 0.0001594736842105263, "loss": 0.0217, "step": 4950 }, { "epoch": 44.03, "learning_rate": 0.00015868421052631578, "loss": 0.0178, "step": 4975 }, { "epoch": 44.25, "learning_rate": 0.00015789473684210524, "loss": 0.0177, "step": 5000 }, { "epoch": 44.25, "eval_loss": 0.24588559567928314, "eval_runtime": 10.7371, "eval_samples_per_second": 9.313, "eval_steps_per_second": 1.211, "eval_wer": 0.07437522420184145, "step": 5000 }, { "epoch": 44.47, "learning_rate": 0.00015710526315789472, "loss": 0.0163, "step": 5025 }, { "epoch": 44.69, "learning_rate": 0.00015631578947368418, "loss": 0.0196, "step": 5050 }, { "epoch": 44.91, "learning_rate": 0.00015552631578947366, "loss": 0.0161, "step": 5075 }, { "epoch": 45.13, "learning_rate": 0.00015473684210526317, "loss": 0.0158, "step": 5100 }, { "epoch": 45.35, "learning_rate": 0.00015394736842105263, "loss": 0.0159, "step": 5125 }, { "epoch": 45.58, "learning_rate": 0.0001531578947368421, "loss": 0.0163, "step": 5150 }, { "epoch": 45.8, "learning_rate": 0.00015236842105263156, "loss": 0.0137, "step": 5175 }, { "epoch": 46.02, "learning_rate": 0.00015157894736842105, "loss": 0.0131, "step": 5200 }, { "epoch": 46.24, "learning_rate": 0.0001507894736842105, "loss": 0.0125, "step": 5225 }, { "epoch": 46.46, "learning_rate": 0.00015, "loss": 0.0138, "step": 5250 }, { "epoch": 46.68, "learning_rate": 0.00014921052631578947, "loss": 0.0179, "step": 5275 }, { "epoch": 46.9, "learning_rate": 0.00014842105263157893, "loss": 0.0169, "step": 5300 }, { "epoch": 47.12, "learning_rate": 0.0001476315789473684, "loss": 0.0152, "step": 5325 }, { "epoch": 47.35, "learning_rate": 0.0001468421052631579, "loss": 0.0146, "step": 5350 }, { "epoch": 47.57, "learning_rate": 0.00014605263157894735, "loss": 0.0143, "step": 5375 }, { "epoch": 47.79, "learning_rate": 0.00014526315789473683, "loss": 0.0116, "step": 5400 }, { "epoch": 48.01, "learning_rate": 0.00014447368421052631, "loss": 0.0162, "step": 5425 }, { "epoch": 48.23, "learning_rate": 0.00014368421052631577, "loss": 0.0115, "step": 5450 }, { "epoch": 48.45, "learning_rate": 0.00014289473684210525, "loss": 0.0156, "step": 5475 }, { "epoch": 48.67, "learning_rate": 0.0001421052631578947, "loss": 0.0144, "step": 5500 }, { "epoch": 48.89, "learning_rate": 0.0001413157894736842, "loss": 0.0155, "step": 5525 }, { "epoch": 49.12, "learning_rate": 0.00014052631578947367, "loss": 0.0146, "step": 5550 }, { "epoch": 49.34, "learning_rate": 0.00013973684210526316, "loss": 0.0164, "step": 5575 }, { "epoch": 49.56, "learning_rate": 0.00013894736842105261, "loss": 0.0128, "step": 5600 }, { "epoch": 49.78, "learning_rate": 0.0001381578947368421, "loss": 0.0133, "step": 5625 }, { "epoch": 50.0, "learning_rate": 0.00013736842105263155, "loss": 0.0126, "step": 5650 }, { "epoch": 50.22, "learning_rate": 0.00013657894736842104, "loss": 0.011, "step": 5675 }, { "epoch": 50.44, "learning_rate": 0.00013578947368421052, "loss": 0.0154, "step": 5700 }, { "epoch": 50.66, "learning_rate": 0.000135, "loss": 0.0154, "step": 5725 }, { "epoch": 50.88, "learning_rate": 0.00013421052631578946, "loss": 0.0131, "step": 5750 }, { "epoch": 51.11, "learning_rate": 0.00013342105263157894, "loss": 0.0161, "step": 5775 }, { "epoch": 51.33, "learning_rate": 0.0001326315789473684, "loss": 0.0135, "step": 5800 }, { "epoch": 51.55, "learning_rate": 0.00013184210526315788, "loss": 0.0125, "step": 5825 }, { "epoch": 51.77, "learning_rate": 0.00013105263157894736, "loss": 0.0114, "step": 5850 }, { "epoch": 51.99, "learning_rate": 0.00013026315789473685, "loss": 0.0106, "step": 5875 }, { "epoch": 52.21, "learning_rate": 0.0001294736842105263, "loss": 0.0119, "step": 5900 }, { "epoch": 52.43, "learning_rate": 0.00012868421052631578, "loss": 0.0123, "step": 5925 }, { "epoch": 52.65, "learning_rate": 0.00012789473684210524, "loss": 0.0111, "step": 5950 }, { "epoch": 52.88, "learning_rate": 0.00012710526315789472, "loss": 0.0139, "step": 5975 }, { "epoch": 53.1, "learning_rate": 0.00012631578947368418, "loss": 0.0155, "step": 6000 }, { "epoch": 53.1, "eval_loss": 0.2688598036766052, "eval_runtime": 10.7568, "eval_samples_per_second": 9.296, "eval_steps_per_second": 1.209, "eval_wer": 0.06791821116824107, "step": 6000 }, { "epoch": 53.32, "learning_rate": 0.0001255263157894737, "loss": 0.0153, "step": 6025 }, { "epoch": 53.54, "learning_rate": 0.00012473684210526315, "loss": 0.013, "step": 6050 }, { "epoch": 53.76, "learning_rate": 0.00012394736842105263, "loss": 0.0093, "step": 6075 }, { "epoch": 53.98, "learning_rate": 0.00012315789473684208, "loss": 0.0125, "step": 6100 }, { "epoch": 54.2, "learning_rate": 0.00012236842105263157, "loss": 0.0122, "step": 6125 }, { "epoch": 54.42, "learning_rate": 0.00012157894736842104, "loss": 0.0131, "step": 6150 }, { "epoch": 54.65, "learning_rate": 0.0001207894736842105, "loss": 0.0129, "step": 6175 }, { "epoch": 54.87, "learning_rate": 0.00011999999999999999, "loss": 0.0119, "step": 6200 }, { "epoch": 55.09, "learning_rate": 0.00011921052631578947, "loss": 0.012, "step": 6225 }, { "epoch": 55.31, "learning_rate": 0.00011842105263157894, "loss": 0.0106, "step": 6250 }, { "epoch": 55.53, "learning_rate": 0.00011763157894736841, "loss": 0.0077, "step": 6275 }, { "epoch": 55.75, "learning_rate": 0.00011684210526315788, "loss": 0.0093, "step": 6300 }, { "epoch": 55.97, "learning_rate": 0.00011605263157894735, "loss": 0.0103, "step": 6325 }, { "epoch": 56.19, "learning_rate": 0.00011526315789473682, "loss": 0.0148, "step": 6350 }, { "epoch": 56.42, "learning_rate": 0.00011447368421052632, "loss": 0.0158, "step": 6375 }, { "epoch": 56.64, "learning_rate": 0.00011368421052631579, "loss": 0.0108, "step": 6400 }, { "epoch": 56.86, "learning_rate": 0.00011289473684210526, "loss": 0.0135, "step": 6425 }, { "epoch": 57.08, "learning_rate": 0.00011210526315789472, "loss": 0.0116, "step": 6450 }, { "epoch": 57.3, "learning_rate": 0.0001113157894736842, "loss": 0.0105, "step": 6475 }, { "epoch": 57.52, "learning_rate": 0.00011052631578947366, "loss": 0.0101, "step": 6500 }, { "epoch": 57.74, "learning_rate": 0.00010973684210526316, "loss": 0.0089, "step": 6525 }, { "epoch": 57.96, "learning_rate": 0.00010894736842105263, "loss": 0.0098, "step": 6550 }, { "epoch": 58.19, "learning_rate": 0.0001081578947368421, "loss": 0.0121, "step": 6575 }, { "epoch": 58.41, "learning_rate": 0.00010736842105263157, "loss": 0.013, "step": 6600 }, { "epoch": 58.63, "learning_rate": 0.00010657894736842104, "loss": 0.012, "step": 6625 }, { "epoch": 58.85, "learning_rate": 0.00010578947368421051, "loss": 0.0089, "step": 6650 }, { "epoch": 59.07, "learning_rate": 0.00010499999999999999, "loss": 0.0111, "step": 6675 }, { "epoch": 59.29, "learning_rate": 0.00010421052631578947, "loss": 0.0085, "step": 6700 }, { "epoch": 59.51, "learning_rate": 0.00010342105263157894, "loss": 0.0094, "step": 6725 }, { "epoch": 59.73, "learning_rate": 0.00010263157894736841, "loss": 0.0098, "step": 6750 }, { "epoch": 59.96, "learning_rate": 0.00010184210526315788, "loss": 0.0109, "step": 6775 }, { "epoch": 60.18, "learning_rate": 0.00010105263157894735, "loss": 0.0129, "step": 6800 }, { "epoch": 60.4, "learning_rate": 0.00010026315789473683, "loss": 0.01, "step": 6825 }, { "epoch": 60.62, "learning_rate": 9.94736842105263e-05, "loss": 0.0089, "step": 6850 }, { "epoch": 60.84, "learning_rate": 9.868421052631579e-05, "loss": 0.0122, "step": 6875 }, { "epoch": 61.06, "learning_rate": 9.789473684210526e-05, "loss": 0.0167, "step": 6900 }, { "epoch": 61.28, "learning_rate": 9.710526315789473e-05, "loss": 0.0105, "step": 6925 }, { "epoch": 61.5, "learning_rate": 9.63157894736842e-05, "loss": 0.0089, "step": 6950 }, { "epoch": 61.73, "learning_rate": 9.552631578947368e-05, "loss": 0.0102, "step": 6975 }, { "epoch": 61.95, "learning_rate": 9.473684210526315e-05, "loss": 0.0149, "step": 7000 }, { "epoch": 61.95, "eval_loss": 0.2759546935558319, "eval_runtime": 10.7835, "eval_samples_per_second": 9.273, "eval_steps_per_second": 1.206, "eval_wer": 0.07174458926222647, "step": 7000 }, { "epoch": 62.17, "learning_rate": 9.394736842105262e-05, "loss": 0.0116, "step": 7025 }, { "epoch": 62.39, "learning_rate": 9.31578947368421e-05, "loss": 0.0102, "step": 7050 }, { "epoch": 62.61, "learning_rate": 9.236842105263157e-05, "loss": 0.0118, "step": 7075 }, { "epoch": 62.83, "learning_rate": 9.157894736842104e-05, "loss": 0.0093, "step": 7100 }, { "epoch": 63.05, "learning_rate": 9.078947368421052e-05, "loss": 0.0074, "step": 7125 }, { "epoch": 63.27, "learning_rate": 8.999999999999999e-05, "loss": 0.0079, "step": 7150 }, { "epoch": 63.5, "learning_rate": 8.921052631578946e-05, "loss": 0.0075, "step": 7175 }, { "epoch": 63.72, "learning_rate": 8.842105263157893e-05, "loss": 0.0099, "step": 7200 }, { "epoch": 63.94, "learning_rate": 8.763157894736841e-05, "loss": 0.0098, "step": 7225 }, { "epoch": 64.16, "learning_rate": 8.68421052631579e-05, "loss": 0.0109, "step": 7250 }, { "epoch": 64.38, "learning_rate": 8.605263157894737e-05, "loss": 0.0103, "step": 7275 }, { "epoch": 64.6, "learning_rate": 8.526315789473684e-05, "loss": 0.0074, "step": 7300 }, { "epoch": 64.82, "learning_rate": 8.44736842105263e-05, "loss": 0.0101, "step": 7325 }, { "epoch": 65.04, "learning_rate": 8.368421052631578e-05, "loss": 0.009, "step": 7350 }, { "epoch": 65.27, "learning_rate": 8.289473684210526e-05, "loss": 0.0077, "step": 7375 }, { "epoch": 65.49, "learning_rate": 8.210526315789474e-05, "loss": 0.009, "step": 7400 }, { "epoch": 65.71, "learning_rate": 8.131578947368421e-05, "loss": 0.0098, "step": 7425 }, { "epoch": 65.93, "learning_rate": 8.052631578947368e-05, "loss": 0.0096, "step": 7450 }, { "epoch": 66.15, "learning_rate": 7.973684210526315e-05, "loss": 0.0136, "step": 7475 }, { "epoch": 66.37, "learning_rate": 7.894736842105262e-05, "loss": 0.0082, "step": 7500 }, { "epoch": 66.59, "learning_rate": 7.815789473684209e-05, "loss": 0.0089, "step": 7525 }, { "epoch": 66.81, "learning_rate": 7.736842105263159e-05, "loss": 0.0089, "step": 7550 }, { "epoch": 67.04, "learning_rate": 7.657894736842105e-05, "loss": 0.0084, "step": 7575 }, { "epoch": 67.26, "learning_rate": 7.578947368421052e-05, "loss": 0.0073, "step": 7600 }, { "epoch": 67.48, "learning_rate": 7.5e-05, "loss": 0.008, "step": 7625 }, { "epoch": 67.7, "learning_rate": 7.421052631578946e-05, "loss": 0.0083, "step": 7650 }, { "epoch": 67.92, "learning_rate": 7.342105263157895e-05, "loss": 0.0068, "step": 7675 }, { "epoch": 68.14, "learning_rate": 7.263157894736842e-05, "loss": 0.007, "step": 7700 }, { "epoch": 68.36, "learning_rate": 7.184210526315788e-05, "loss": 0.0086, "step": 7725 }, { "epoch": 68.58, "learning_rate": 7.105263157894735e-05, "loss": 0.0077, "step": 7750 }, { "epoch": 68.81, "learning_rate": 7.026315789473684e-05, "loss": 0.0076, "step": 7775 }, { "epoch": 69.03, "learning_rate": 6.947368421052631e-05, "loss": 0.006, "step": 7800 }, { "epoch": 69.25, "learning_rate": 6.868421052631578e-05, "loss": 0.006, "step": 7825 }, { "epoch": 69.47, "learning_rate": 6.789473684210526e-05, "loss": 0.0077, "step": 7850 }, { "epoch": 69.69, "learning_rate": 6.710526315789473e-05, "loss": 0.009, "step": 7875 }, { "epoch": 69.91, "learning_rate": 6.63157894736842e-05, "loss": 0.0081, "step": 7900 }, { "epoch": 70.13, "learning_rate": 6.552631578947368e-05, "loss": 0.0085, "step": 7925 }, { "epoch": 70.35, "learning_rate": 6.473684210526315e-05, "loss": 0.007, "step": 7950 }, { "epoch": 70.58, "learning_rate": 6.394736842105262e-05, "loss": 0.0088, "step": 7975 }, { "epoch": 70.8, "learning_rate": 6.315789473684209e-05, "loss": 0.0074, "step": 8000 }, { "epoch": 70.8, "eval_loss": 0.30043184757232666, "eval_runtime": 10.795, "eval_samples_per_second": 9.264, "eval_steps_per_second": 1.204, "eval_wer": 0.0680377854836781, "step": 8000 }, { "epoch": 71.02, "learning_rate": 6.236842105263157e-05, "loss": 0.0061, "step": 8025 }, { "epoch": 71.24, "learning_rate": 6.157894736842104e-05, "loss": 0.0059, "step": 8050 }, { "epoch": 71.46, "learning_rate": 6.078947368421052e-05, "loss": 0.008, "step": 8075 }, { "epoch": 71.68, "learning_rate": 5.9999999999999995e-05, "loss": 0.0105, "step": 8100 }, { "epoch": 71.9, "learning_rate": 5.921052631578947e-05, "loss": 0.0066, "step": 8125 }, { "epoch": 72.12, "learning_rate": 5.842105263157894e-05, "loss": 0.0075, "step": 8150 }, { "epoch": 72.35, "learning_rate": 5.763157894736841e-05, "loss": 0.01, "step": 8175 }, { "epoch": 72.57, "learning_rate": 5.684210526315789e-05, "loss": 0.0097, "step": 8200 }, { "epoch": 72.79, "learning_rate": 5.605263157894736e-05, "loss": 0.007, "step": 8225 }, { "epoch": 73.01, "learning_rate": 5.526315789473683e-05, "loss": 0.0047, "step": 8250 }, { "epoch": 73.23, "learning_rate": 5.4473684210526315e-05, "loss": 0.006, "step": 8275 }, { "epoch": 73.45, "learning_rate": 5.3684210526315784e-05, "loss": 0.0078, "step": 8300 }, { "epoch": 73.67, "learning_rate": 5.2894736842105254e-05, "loss": 0.0085, "step": 8325 }, { "epoch": 73.89, "learning_rate": 5.210526315789474e-05, "loss": 0.0075, "step": 8350 }, { "epoch": 74.12, "learning_rate": 5.1315789473684206e-05, "loss": 0.0062, "step": 8375 }, { "epoch": 74.34, "learning_rate": 5.0526315789473676e-05, "loss": 0.0073, "step": 8400 }, { "epoch": 74.56, "learning_rate": 4.973684210526315e-05, "loss": 0.0063, "step": 8425 }, { "epoch": 74.78, "learning_rate": 4.894736842105263e-05, "loss": 0.006, "step": 8450 }, { "epoch": 75.0, "learning_rate": 4.81578947368421e-05, "loss": 0.0064, "step": 8475 }, { "epoch": 75.22, "learning_rate": 4.7368421052631574e-05, "loss": 0.0058, "step": 8500 }, { "epoch": 75.44, "learning_rate": 4.657894736842105e-05, "loss": 0.0059, "step": 8525 }, { "epoch": 75.66, "learning_rate": 4.578947368421052e-05, "loss": 0.0064, "step": 8550 }, { "epoch": 75.88, "learning_rate": 4.4999999999999996e-05, "loss": 0.0085, "step": 8575 }, { "epoch": 76.11, "learning_rate": 4.4210526315789466e-05, "loss": 0.0065, "step": 8600 }, { "epoch": 76.33, "learning_rate": 4.342105263157895e-05, "loss": 0.0068, "step": 8625 }, { "epoch": 76.55, "learning_rate": 4.263157894736842e-05, "loss": 0.0084, "step": 8650 }, { "epoch": 76.77, "learning_rate": 4.184210526315789e-05, "loss": 0.0049, "step": 8675 }, { "epoch": 76.99, "learning_rate": 4.105263157894737e-05, "loss": 0.0055, "step": 8700 }, { "epoch": 77.21, "learning_rate": 4.026315789473684e-05, "loss": 0.0077, "step": 8725 }, { "epoch": 77.43, "learning_rate": 3.947368421052631e-05, "loss": 0.0071, "step": 8750 }, { "epoch": 77.65, "learning_rate": 3.868421052631579e-05, "loss": 0.005, "step": 8775 }, { "epoch": 77.88, "learning_rate": 3.789473684210526e-05, "loss": 0.0066, "step": 8800 }, { "epoch": 78.1, "learning_rate": 3.710526315789473e-05, "loss": 0.0067, "step": 8825 }, { "epoch": 78.32, "learning_rate": 3.631578947368421e-05, "loss": 0.0058, "step": 8850 }, { "epoch": 78.54, "learning_rate": 3.552631578947368e-05, "loss": 0.004, "step": 8875 }, { "epoch": 78.76, "learning_rate": 3.4736842105263153e-05, "loss": 0.0044, "step": 8900 }, { "epoch": 78.98, "learning_rate": 3.394736842105263e-05, "loss": 0.0052, "step": 8925 }, { "epoch": 79.2, "learning_rate": 3.31578947368421e-05, "loss": 0.0058, "step": 8950 }, { "epoch": 79.42, "learning_rate": 3.2368421052631575e-05, "loss": 0.0063, "step": 8975 }, { "epoch": 79.65, "learning_rate": 3.1578947368421045e-05, "loss": 0.0058, "step": 9000 }, { "epoch": 79.65, "eval_loss": 0.31133854389190674, "eval_runtime": 10.8234, "eval_samples_per_second": 9.239, "eval_steps_per_second": 1.201, "eval_wer": 0.065048427597752, "step": 9000 }, { "epoch": 79.87, "learning_rate": 3.078947368421052e-05, "loss": 0.0052, "step": 9025 }, { "epoch": 80.09, "learning_rate": 2.9999999999999997e-05, "loss": 0.0053, "step": 9050 }, { "epoch": 80.31, "learning_rate": 2.921052631578947e-05, "loss": 0.0049, "step": 9075 }, { "epoch": 80.53, "learning_rate": 2.8421052631578946e-05, "loss": 0.0049, "step": 9100 }, { "epoch": 80.75, "learning_rate": 2.7631578947368416e-05, "loss": 0.0045, "step": 9125 }, { "epoch": 80.97, "learning_rate": 2.6842105263157892e-05, "loss": 0.0053, "step": 9150 }, { "epoch": 81.19, "learning_rate": 2.605263157894737e-05, "loss": 0.0051, "step": 9175 }, { "epoch": 81.42, "learning_rate": 2.5263157894736838e-05, "loss": 0.0056, "step": 9200 }, { "epoch": 81.64, "learning_rate": 2.4473684210526314e-05, "loss": 0.0056, "step": 9225 }, { "epoch": 81.86, "learning_rate": 2.3684210526315787e-05, "loss": 0.0047, "step": 9250 }, { "epoch": 82.08, "learning_rate": 2.289473684210526e-05, "loss": 0.0072, "step": 9275 }, { "epoch": 82.3, "learning_rate": 2.2105263157894733e-05, "loss": 0.0055, "step": 9300 }, { "epoch": 82.52, "learning_rate": 2.131578947368421e-05, "loss": 0.005, "step": 9325 }, { "epoch": 82.74, "learning_rate": 2.0526315789473685e-05, "loss": 0.0035, "step": 9350 }, { "epoch": 82.96, "learning_rate": 1.9736842105263155e-05, "loss": 0.0052, "step": 9375 }, { "epoch": 83.19, "learning_rate": 1.894736842105263e-05, "loss": 0.0046, "step": 9400 }, { "epoch": 83.41, "learning_rate": 1.8157894736842104e-05, "loss": 0.0056, "step": 9425 }, { "epoch": 83.63, "learning_rate": 1.7368421052631577e-05, "loss": 0.005, "step": 9450 }, { "epoch": 83.85, "learning_rate": 1.657894736842105e-05, "loss": 0.0046, "step": 9475 }, { "epoch": 84.07, "learning_rate": 1.5789473684210522e-05, "loss": 0.0061, "step": 9500 }, { "epoch": 84.29, "learning_rate": 1.4999999999999999e-05, "loss": 0.0051, "step": 9525 }, { "epoch": 84.51, "learning_rate": 1.4210526315789473e-05, "loss": 0.0033, "step": 9550 }, { "epoch": 84.73, "learning_rate": 1.3421052631578946e-05, "loss": 0.0035, "step": 9575 }, { "epoch": 84.96, "learning_rate": 1.2631578947368419e-05, "loss": 0.0054, "step": 9600 }, { "epoch": 85.18, "learning_rate": 1.1842105263157894e-05, "loss": 0.005, "step": 9625 }, { "epoch": 85.4, "learning_rate": 1.1052631578947366e-05, "loss": 0.0039, "step": 9650 }, { "epoch": 85.62, "learning_rate": 1.0263157894736843e-05, "loss": 0.0044, "step": 9675 }, { "epoch": 85.84, "learning_rate": 9.473684210526315e-06, "loss": 0.0049, "step": 9700 }, { "epoch": 86.06, "learning_rate": 8.684210526315788e-06, "loss": 0.0042, "step": 9725 }, { "epoch": 86.28, "learning_rate": 7.894736842105261e-06, "loss": 0.0041, "step": 9750 }, { "epoch": 86.5, "learning_rate": 7.105263157894737e-06, "loss": 0.0036, "step": 9775 }, { "epoch": 86.73, "learning_rate": 6.3157894736842095e-06, "loss": 0.0037, "step": 9800 }, { "epoch": 86.95, "learning_rate": 5.526315789473683e-06, "loss": 0.005, "step": 9825 }, { "epoch": 87.17, "learning_rate": 4.736842105263158e-06, "loss": 0.0047, "step": 9850 }, { "epoch": 87.39, "learning_rate": 3.947368421052631e-06, "loss": 0.0039, "step": 9875 }, { "epoch": 87.61, "learning_rate": 3.1578947368421047e-06, "loss": 0.0041, "step": 9900 }, { "epoch": 87.83, "learning_rate": 2.368421052631579e-06, "loss": 0.0052, "step": 9925 }, { "epoch": 88.05, "learning_rate": 1.5789473684210524e-06, "loss": 0.0041, "step": 9950 }, { "epoch": 88.27, "learning_rate": 7.894736842105262e-07, "loss": 0.004, "step": 9975 }, { "epoch": 88.5, "learning_rate": 0.0, "loss": 0.0033, "step": 10000 }, { "epoch": 88.5, "eval_loss": 0.32122763991355896, "eval_runtime": 10.9251, "eval_samples_per_second": 9.153, "eval_steps_per_second": 1.19, "eval_wer": 0.06361353581250748, "step": 10000 }, { "epoch": 88.5, "step": 10000, "total_flos": 3.937659402470073e+19, "train_loss": 0.13773754513785244, "train_runtime": 20842.6742, "train_samples_per_second": 7.677, "train_steps_per_second": 0.48 } ], "max_steps": 10000, "num_train_epochs": 89, "total_flos": 3.937659402470073e+19, "trial_name": null, "trial_params": null }