{ "best_metric": null, "best_model_checkpoint": null, "epoch": 1.951092611862643, "global_step": 15000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.0, "learning_rate": 2e-05, "loss": 3.1319, "step": 25 }, { "epoch": 0.01, "learning_rate": 4.4999999999999996e-05, "loss": 0.7401, "step": 50 }, { "epoch": 0.01, "learning_rate": 7.000000000000001e-05, "loss": 0.4057, "step": 75 }, { "epoch": 0.01, "learning_rate": 9.5e-05, "loss": 0.4962, "step": 100 }, { "epoch": 0.02, "learning_rate": 0.00012, "loss": 0.5871, "step": 125 }, { "epoch": 0.02, "learning_rate": 0.000145, "loss": 0.7396, "step": 150 }, { "epoch": 0.02, "learning_rate": 0.00017, "loss": 0.8646, "step": 175 }, { "epoch": 0.03, "learning_rate": 0.00019500000000000002, "loss": 0.9588, "step": 200 }, { "epoch": 0.03, "learning_rate": 0.00022, "loss": 1.1775, "step": 225 }, { "epoch": 0.03, "learning_rate": 0.000245, "loss": 1.3102, "step": 250 }, { "epoch": 0.04, "learning_rate": 0.00027, "loss": 1.3424, "step": 275 }, { "epoch": 0.04, "learning_rate": 0.000295, "loss": 1.5322, "step": 300 }, { "epoch": 0.04, "learning_rate": 0.00032, "loss": 1.5378, "step": 325 }, { "epoch": 0.05, "learning_rate": 0.000345, "loss": 1.7126, "step": 350 }, { "epoch": 0.05, "learning_rate": 0.00037, "loss": 1.8599, "step": 375 }, { "epoch": 0.05, "learning_rate": 0.000395, "loss": 1.9935, "step": 400 }, { "epoch": 0.06, "learning_rate": 0.00042, "loss": 2.1787, "step": 425 }, { "epoch": 0.06, "learning_rate": 0.00044500000000000003, "loss": 2.2732, "step": 450 }, { "epoch": 0.06, "learning_rate": 0.00047, "loss": 2.3751, "step": 475 }, { "epoch": 0.07, "learning_rate": 0.000495, "loss": 2.3598, "step": 500 }, { "epoch": 0.07, "learning_rate": 0.0004993103448275862, "loss": 2.4274, "step": 525 }, { "epoch": 0.07, "learning_rate": 0.0004984482758620689, "loss": 2.4939, "step": 550 }, { "epoch": 0.07, "learning_rate": 0.0004975862068965517, "loss": 2.5545, "step": 575 }, { "epoch": 0.08, "learning_rate": 0.0004967241379310345, "loss": 2.4871, "step": 600 }, { "epoch": 0.08, "learning_rate": 0.0004958620689655172, "loss": 2.5471, "step": 625 }, { "epoch": 0.08, "learning_rate": 0.000495, "loss": 2.5501, "step": 650 }, { "epoch": 0.09, "learning_rate": 0.0004941379310344828, "loss": 2.5606, "step": 675 }, { "epoch": 0.09, "learning_rate": 0.0004932758620689655, "loss": 2.4623, "step": 700 }, { "epoch": 0.09, "learning_rate": 0.0004924137931034483, "loss": 2.5275, "step": 725 }, { "epoch": 0.1, "learning_rate": 0.000491551724137931, "loss": 2.5476, "step": 750 }, { "epoch": 0.1, "learning_rate": 0.0004906896551724138, "loss": 2.3168, "step": 775 }, { "epoch": 0.1, "learning_rate": 0.0004898275862068966, "loss": 2.5242, "step": 800 }, { "epoch": 0.11, "learning_rate": 0.0004889655172413793, "loss": 2.3614, "step": 825 }, { "epoch": 0.11, "learning_rate": 0.00048810344827586205, "loss": 2.3791, "step": 850 }, { "epoch": 0.11, "learning_rate": 0.00048724137931034487, "loss": 2.3276, "step": 875 }, { "epoch": 0.12, "learning_rate": 0.0004863793103448276, "loss": 2.2025, "step": 900 }, { "epoch": 0.12, "learning_rate": 0.00048551724137931035, "loss": 2.1829, "step": 925 }, { "epoch": 0.12, "learning_rate": 0.0004846551724137931, "loss": 2.1973, "step": 950 }, { "epoch": 0.13, "learning_rate": 0.0004837931034482759, "loss": 2.1836, "step": 975 }, { "epoch": 0.13, "learning_rate": 0.00048293103448275864, "loss": 2.2671, "step": 1000 }, { "epoch": 0.13, "eval_loss": 2.2107748985290527, "eval_runtime": 4394.6691, "eval_samples_per_second": 1.766, "eval_steps_per_second": 0.221, "eval_wer": 76.26673976035711, "step": 1000 }, { "epoch": 0.13, "learning_rate": 0.0004820689655172414, "loss": 2.0956, "step": 1025 }, { "epoch": 0.14, "learning_rate": 0.00048120689655172416, "loss": 2.1123, "step": 1050 }, { "epoch": 0.14, "learning_rate": 0.0004803448275862069, "loss": 2.1222, "step": 1075 }, { "epoch": 0.14, "learning_rate": 0.0004794827586206897, "loss": 2.0646, "step": 1100 }, { "epoch": 0.15, "learning_rate": 0.0004786206896551724, "loss": 2.0799, "step": 1125 }, { "epoch": 0.15, "learning_rate": 0.0004777586206896552, "loss": 2.1963, "step": 1150 }, { "epoch": 0.15, "learning_rate": 0.00047689655172413793, "loss": 2.083, "step": 1175 }, { "epoch": 0.16, "learning_rate": 0.0004760344827586207, "loss": 2.2261, "step": 1200 }, { "epoch": 0.16, "learning_rate": 0.00047517241379310346, "loss": 1.9378, "step": 1225 }, { "epoch": 0.16, "learning_rate": 0.0004743103448275862, "loss": 2.0096, "step": 1250 }, { "epoch": 0.17, "learning_rate": 0.00047344827586206893, "loss": 2.054, "step": 1275 }, { "epoch": 0.17, "learning_rate": 0.00047258620689655175, "loss": 1.9917, "step": 1300 }, { "epoch": 0.17, "learning_rate": 0.00047172413793103446, "loss": 1.9676, "step": 1325 }, { "epoch": 0.18, "learning_rate": 0.0004708620689655173, "loss": 1.9742, "step": 1350 }, { "epoch": 0.18, "learning_rate": 0.00047, "loss": 1.9385, "step": 1375 }, { "epoch": 0.18, "learning_rate": 0.00046913793103448275, "loss": 2.0315, "step": 1400 }, { "epoch": 0.19, "learning_rate": 0.00046827586206896557, "loss": 2.0112, "step": 1425 }, { "epoch": 0.19, "learning_rate": 0.0004674137931034483, "loss": 1.9804, "step": 1450 }, { "epoch": 0.19, "learning_rate": 0.00046655172413793104, "loss": 1.9427, "step": 1475 }, { "epoch": 0.2, "learning_rate": 0.0004656896551724138, "loss": 1.9456, "step": 1500 }, { "epoch": 0.2, "learning_rate": 0.00046482758620689657, "loss": 1.945, "step": 1525 }, { "epoch": 0.2, "learning_rate": 0.00046396551724137933, "loss": 1.994, "step": 1550 }, { "epoch": 0.2, "learning_rate": 0.0004631034482758621, "loss": 1.8514, "step": 1575 }, { "epoch": 0.21, "learning_rate": 0.0004622413793103448, "loss": 1.8197, "step": 1600 }, { "epoch": 0.21, "learning_rate": 0.0004613793103448276, "loss": 1.7805, "step": 1625 }, { "epoch": 0.21, "learning_rate": 0.00046051724137931033, "loss": 1.6539, "step": 1650 }, { "epoch": 0.22, "learning_rate": 0.0004596551724137931, "loss": 1.7395, "step": 1675 }, { "epoch": 0.22, "learning_rate": 0.00045879310344827586, "loss": 1.668, "step": 1700 }, { "epoch": 0.22, "learning_rate": 0.0004579310344827586, "loss": 1.7438, "step": 1725 }, { "epoch": 0.23, "learning_rate": 0.0004570689655172414, "loss": 1.6567, "step": 1750 }, { "epoch": 0.23, "learning_rate": 0.00045620689655172415, "loss": 1.7735, "step": 1775 }, { "epoch": 0.23, "learning_rate": 0.0004553448275862069, "loss": 1.6275, "step": 1800 }, { "epoch": 0.24, "learning_rate": 0.0004544827586206897, "loss": 1.6819, "step": 1825 }, { "epoch": 0.24, "learning_rate": 0.00045362068965517244, "loss": 1.6431, "step": 1850 }, { "epoch": 0.24, "learning_rate": 0.00045275862068965515, "loss": 1.6827, "step": 1875 }, { "epoch": 0.25, "learning_rate": 0.00045189655172413797, "loss": 1.5828, "step": 1900 }, { "epoch": 0.25, "learning_rate": 0.0004510344827586207, "loss": 1.5717, "step": 1925 }, { "epoch": 0.25, "learning_rate": 0.0004501724137931035, "loss": 1.5818, "step": 1950 }, { "epoch": 0.26, "learning_rate": 0.0004493103448275862, "loss": 1.5846, "step": 1975 }, { "epoch": 0.26, "learning_rate": 0.000448448275862069, "loss": 1.4465, "step": 2000 }, { "epoch": 0.26, "eval_loss": 1.6056782007217407, "eval_runtime": 4475.9307, "eval_samples_per_second": 1.734, "eval_steps_per_second": 0.217, "eval_wer": 67.87532304800689, "step": 2000 }, { "epoch": 0.26, "learning_rate": 0.00044758620689655174, "loss": 1.5138, "step": 2025 }, { "epoch": 0.27, "learning_rate": 0.0004467241379310345, "loss": 1.4, "step": 2050 }, { "epoch": 0.27, "learning_rate": 0.0004458620689655172, "loss": 1.5535, "step": 2075 }, { "epoch": 0.27, "learning_rate": 0.00044500000000000003, "loss": 1.4993, "step": 2100 }, { "epoch": 0.28, "learning_rate": 0.00044413793103448274, "loss": 1.4173, "step": 2125 }, { "epoch": 0.28, "learning_rate": 0.00044327586206896556, "loss": 1.4179, "step": 2150 }, { "epoch": 0.28, "learning_rate": 0.00044241379310344827, "loss": 1.4607, "step": 2175 }, { "epoch": 0.29, "learning_rate": 0.00044155172413793103, "loss": 1.3994, "step": 2200 }, { "epoch": 0.29, "learning_rate": 0.00044068965517241385, "loss": 1.3698, "step": 2225 }, { "epoch": 0.29, "learning_rate": 0.00043982758620689656, "loss": 1.3497, "step": 2250 }, { "epoch": 0.3, "learning_rate": 0.0004389655172413793, "loss": 1.3208, "step": 2275 }, { "epoch": 0.3, "learning_rate": 0.0004381034482758621, "loss": 1.3414, "step": 2300 }, { "epoch": 0.3, "learning_rate": 0.00043724137931034485, "loss": 1.3346, "step": 2325 }, { "epoch": 0.31, "learning_rate": 0.00043637931034482756, "loss": 1.3796, "step": 2350 }, { "epoch": 0.31, "learning_rate": 0.0004355172413793104, "loss": 1.3769, "step": 2375 }, { "epoch": 0.31, "learning_rate": 0.0004346551724137931, "loss": 1.3307, "step": 2400 }, { "epoch": 0.32, "learning_rate": 0.0004337931034482759, "loss": 1.3188, "step": 2425 }, { "epoch": 0.32, "learning_rate": 0.0004329310344827586, "loss": 1.2849, "step": 2450 }, { "epoch": 0.32, "learning_rate": 0.0004320689655172414, "loss": 1.2577, "step": 2475 }, { "epoch": 0.33, "learning_rate": 0.00043120689655172414, "loss": 1.2991, "step": 2500 }, { "epoch": 0.33, "learning_rate": 0.0004303448275862069, "loss": 1.2529, "step": 2525 }, { "epoch": 0.33, "learning_rate": 0.0004294827586206896, "loss": 1.2531, "step": 2550 }, { "epoch": 0.33, "learning_rate": 0.00042862068965517243, "loss": 1.2895, "step": 2575 }, { "epoch": 0.34, "learning_rate": 0.0004277586206896552, "loss": 1.2486, "step": 2600 }, { "epoch": 0.34, "learning_rate": 0.00042689655172413796, "loss": 1.2277, "step": 2625 }, { "epoch": 0.34, "learning_rate": 0.0004260344827586207, "loss": 1.2134, "step": 2650 }, { "epoch": 0.35, "learning_rate": 0.00042517241379310343, "loss": 1.2484, "step": 2675 }, { "epoch": 0.35, "learning_rate": 0.00042431034482758625, "loss": 1.1762, "step": 2700 }, { "epoch": 0.35, "learning_rate": 0.00042344827586206896, "loss": 1.1954, "step": 2725 }, { "epoch": 0.36, "learning_rate": 0.0004225862068965517, "loss": 1.1791, "step": 2750 }, { "epoch": 0.36, "learning_rate": 0.0004217241379310345, "loss": 1.2085, "step": 2775 }, { "epoch": 0.36, "learning_rate": 0.00042086206896551725, "loss": 1.1596, "step": 2800 }, { "epoch": 0.37, "learning_rate": 0.00042, "loss": 1.2118, "step": 2825 }, { "epoch": 0.37, "learning_rate": 0.0004191379310344828, "loss": 1.1266, "step": 2850 }, { "epoch": 0.37, "learning_rate": 0.0004182758620689655, "loss": 1.1305, "step": 2875 }, { "epoch": 0.38, "learning_rate": 0.0004174137931034483, "loss": 1.2088, "step": 2900 }, { "epoch": 0.38, "learning_rate": 0.000416551724137931, "loss": 1.2252, "step": 2925 }, { "epoch": 0.38, "learning_rate": 0.0004156896551724138, "loss": 1.1721, "step": 2950 }, { "epoch": 0.39, "learning_rate": 0.00041482758620689655, "loss": 1.2093, "step": 2975 }, { "epoch": 0.39, "learning_rate": 0.0004139655172413793, "loss": 1.0997, "step": 3000 }, { "epoch": 0.39, "eval_loss": 1.1928162574768066, "eval_runtime": 4384.3449, "eval_samples_per_second": 1.77, "eval_steps_per_second": 0.221, "eval_wer": 54.243349779413684, "step": 3000 }, { "epoch": 0.39, "learning_rate": 0.00041310344827586213, "loss": 1.1659, "step": 3025 }, { "epoch": 0.4, "learning_rate": 0.00041224137931034484, "loss": 1.0854, "step": 3050 }, { "epoch": 0.4, "learning_rate": 0.0004113793103448276, "loss": 1.1169, "step": 3075 }, { "epoch": 0.4, "learning_rate": 0.00041051724137931037, "loss": 1.1124, "step": 3100 }, { "epoch": 0.41, "learning_rate": 0.00040965517241379313, "loss": 1.1342, "step": 3125 }, { "epoch": 0.41, "learning_rate": 0.00040879310344827584, "loss": 1.151, "step": 3150 }, { "epoch": 0.41, "learning_rate": 0.00040793103448275866, "loss": 1.0429, "step": 3175 }, { "epoch": 0.42, "learning_rate": 0.00040706896551724137, "loss": 1.0763, "step": 3200 }, { "epoch": 0.42, "learning_rate": 0.0004062068965517242, "loss": 1.0854, "step": 3225 }, { "epoch": 0.42, "learning_rate": 0.0004053448275862069, "loss": 1.0552, "step": 3250 }, { "epoch": 0.43, "learning_rate": 0.00040448275862068966, "loss": 1.1546, "step": 3275 }, { "epoch": 0.43, "learning_rate": 0.0004036206896551724, "loss": 1.0574, "step": 3300 }, { "epoch": 0.43, "learning_rate": 0.0004027586206896552, "loss": 1.0368, "step": 3325 }, { "epoch": 0.44, "learning_rate": 0.0004018965517241379, "loss": 1.0203, "step": 3350 }, { "epoch": 0.44, "learning_rate": 0.0004010344827586207, "loss": 1.0167, "step": 3375 }, { "epoch": 0.44, "learning_rate": 0.0004001724137931034, "loss": 1.0448, "step": 3400 }, { "epoch": 0.45, "learning_rate": 0.0003993103448275862, "loss": 1.0953, "step": 3425 }, { "epoch": 0.45, "learning_rate": 0.000398448275862069, "loss": 1.0192, "step": 3450 }, { "epoch": 0.45, "learning_rate": 0.0003975862068965517, "loss": 1.0239, "step": 3475 }, { "epoch": 0.46, "learning_rate": 0.00039672413793103453, "loss": 1.0651, "step": 3500 }, { "epoch": 0.46, "learning_rate": 0.00039586206896551724, "loss": 1.0226, "step": 3525 }, { "epoch": 0.46, "learning_rate": 0.000395, "loss": 1.0377, "step": 3550 }, { "epoch": 0.47, "learning_rate": 0.00039413793103448277, "loss": 0.9857, "step": 3575 }, { "epoch": 0.47, "learning_rate": 0.00039327586206896553, "loss": 1.0303, "step": 3600 }, { "epoch": 0.47, "learning_rate": 0.00039241379310344824, "loss": 0.9793, "step": 3625 }, { "epoch": 0.47, "learning_rate": 0.00039155172413793106, "loss": 0.9847, "step": 3650 }, { "epoch": 0.48, "learning_rate": 0.00039068965517241377, "loss": 0.9959, "step": 3675 }, { "epoch": 0.48, "learning_rate": 0.0003898275862068966, "loss": 0.9824, "step": 3700 }, { "epoch": 0.48, "learning_rate": 0.0003889655172413793, "loss": 0.9516, "step": 3725 }, { "epoch": 0.49, "learning_rate": 0.00038810344827586206, "loss": 0.9663, "step": 3750 }, { "epoch": 0.49, "learning_rate": 0.0003872413793103448, "loss": 0.931, "step": 3775 }, { "epoch": 0.49, "learning_rate": 0.0003863793103448276, "loss": 1.0623, "step": 3800 }, { "epoch": 0.5, "learning_rate": 0.00038551724137931035, "loss": 0.9715, "step": 3825 }, { "epoch": 0.5, "learning_rate": 0.0003846551724137931, "loss": 0.9888, "step": 3850 }, { "epoch": 0.5, "learning_rate": 0.0003837931034482759, "loss": 0.9638, "step": 3875 }, { "epoch": 0.51, "learning_rate": 0.00038293103448275865, "loss": 1.0095, "step": 3900 }, { "epoch": 0.51, "learning_rate": 0.0003820689655172414, "loss": 0.9327, "step": 3925 }, { "epoch": 0.51, "learning_rate": 0.0003812068965517241, "loss": 0.9805, "step": 3950 }, { "epoch": 0.52, "learning_rate": 0.00038034482758620694, "loss": 0.9266, "step": 3975 }, { "epoch": 0.52, "learning_rate": 0.00037948275862068965, "loss": 0.9389, "step": 4000 }, { "epoch": 0.52, "eval_loss": 1.0020068883895874, "eval_runtime": 4390.2968, "eval_samples_per_second": 1.768, "eval_steps_per_second": 0.221, "eval_wer": 47.83068368705459, "step": 4000 }, { "epoch": 0.52, "learning_rate": 0.0003786206896551724, "loss": 0.9285, "step": 4025 }, { "epoch": 0.53, "learning_rate": 0.0003777586206896552, "loss": 0.9294, "step": 4050 }, { "epoch": 0.53, "learning_rate": 0.00037689655172413794, "loss": 1.0058, "step": 4075 }, { "epoch": 0.53, "learning_rate": 0.0003760344827586207, "loss": 0.967, "step": 4100 }, { "epoch": 0.54, "learning_rate": 0.00037517241379310347, "loss": 0.961, "step": 4125 }, { "epoch": 0.54, "learning_rate": 0.0003743103448275862, "loss": 0.9719, "step": 4150 }, { "epoch": 0.54, "learning_rate": 0.000373448275862069, "loss": 0.9453, "step": 4175 }, { "epoch": 0.55, "learning_rate": 0.0003725862068965517, "loss": 0.9381, "step": 4200 }, { "epoch": 0.55, "learning_rate": 0.00037172413793103447, "loss": 0.8868, "step": 4225 }, { "epoch": 0.55, "learning_rate": 0.0003708620689655173, "loss": 0.9326, "step": 4250 }, { "epoch": 0.56, "learning_rate": 0.00037, "loss": 0.8897, "step": 4275 }, { "epoch": 0.56, "learning_rate": 0.0003691379310344828, "loss": 0.8873, "step": 4300 }, { "epoch": 0.56, "learning_rate": 0.0003682758620689655, "loss": 0.9535, "step": 4325 }, { "epoch": 0.57, "learning_rate": 0.0003674137931034483, "loss": 0.9418, "step": 4350 }, { "epoch": 0.57, "learning_rate": 0.00036655172413793105, "loss": 0.9324, "step": 4375 }, { "epoch": 0.57, "learning_rate": 0.0003656896551724138, "loss": 0.8473, "step": 4400 }, { "epoch": 0.58, "learning_rate": 0.0003648275862068965, "loss": 0.8834, "step": 4425 }, { "epoch": 0.58, "learning_rate": 0.00036396551724137934, "loss": 0.8658, "step": 4450 }, { "epoch": 0.58, "learning_rate": 0.00036310344827586205, "loss": 0.8875, "step": 4475 }, { "epoch": 0.59, "learning_rate": 0.00036224137931034487, "loss": 0.9032, "step": 4500 }, { "epoch": 0.59, "learning_rate": 0.0003613793103448276, "loss": 0.8691, "step": 4525 }, { "epoch": 0.59, "learning_rate": 0.00036051724137931034, "loss": 0.9029, "step": 4550 }, { "epoch": 0.6, "learning_rate": 0.0003596551724137931, "loss": 0.9048, "step": 4575 }, { "epoch": 0.6, "learning_rate": 0.00035879310344827587, "loss": 0.8637, "step": 4600 }, { "epoch": 0.6, "learning_rate": 0.00035793103448275863, "loss": 0.9253, "step": 4625 }, { "epoch": 0.6, "learning_rate": 0.0003570689655172414, "loss": 0.8934, "step": 4650 }, { "epoch": 0.61, "learning_rate": 0.00035620689655172416, "loss": 0.812, "step": 4675 }, { "epoch": 0.61, "learning_rate": 0.00035534482758620687, "loss": 0.8735, "step": 4700 }, { "epoch": 0.61, "learning_rate": 0.0003544827586206897, "loss": 0.8427, "step": 4725 }, { "epoch": 0.62, "learning_rate": 0.0003536206896551724, "loss": 0.8414, "step": 4750 }, { "epoch": 0.62, "learning_rate": 0.0003527586206896552, "loss": 0.8734, "step": 4775 }, { "epoch": 0.62, "learning_rate": 0.0003518965517241379, "loss": 0.8303, "step": 4800 }, { "epoch": 0.63, "learning_rate": 0.0003510344827586207, "loss": 0.8463, "step": 4825 }, { "epoch": 0.63, "learning_rate": 0.00035017241379310345, "loss": 0.821, "step": 4850 }, { "epoch": 0.63, "learning_rate": 0.0003493103448275862, "loss": 0.7878, "step": 4875 }, { "epoch": 0.64, "learning_rate": 0.00034844827586206893, "loss": 0.8106, "step": 4900 }, { "epoch": 0.64, "learning_rate": 0.00034758620689655175, "loss": 0.8263, "step": 4925 }, { "epoch": 0.64, "learning_rate": 0.00034672413793103446, "loss": 0.8215, "step": 4950 }, { "epoch": 0.65, "learning_rate": 0.0003458620689655173, "loss": 0.8063, "step": 4975 }, { "epoch": 0.65, "learning_rate": 0.000345, "loss": 0.7881, "step": 5000 }, { "epoch": 0.65, "eval_loss": 0.8933315873146057, "eval_runtime": 4334.0003, "eval_samples_per_second": 1.79, "eval_steps_per_second": 0.224, "eval_wer": 46.00464667032135, "step": 5000 }, { "epoch": 0.65, "learning_rate": 0.00034413793103448275, "loss": 0.775, "step": 5025 }, { "epoch": 0.66, "learning_rate": 0.00034327586206896557, "loss": 0.842, "step": 5050 }, { "epoch": 0.66, "learning_rate": 0.0003424137931034483, "loss": 0.8452, "step": 5075 }, { "epoch": 0.66, "learning_rate": 0.00034155172413793104, "loss": 0.8184, "step": 5100 }, { "epoch": 0.67, "learning_rate": 0.0003406896551724138, "loss": 0.8292, "step": 5125 }, { "epoch": 0.67, "learning_rate": 0.00033982758620689657, "loss": 0.7263, "step": 5150 }, { "epoch": 0.67, "learning_rate": 0.00033896551724137933, "loss": 0.7959, "step": 5175 }, { "epoch": 0.68, "learning_rate": 0.0003381034482758621, "loss": 0.7924, "step": 5200 }, { "epoch": 0.68, "learning_rate": 0.0003372413793103448, "loss": 0.8546, "step": 5225 }, { "epoch": 0.68, "learning_rate": 0.0003363793103448276, "loss": 0.7946, "step": 5250 }, { "epoch": 0.69, "learning_rate": 0.00033551724137931033, "loss": 0.7834, "step": 5275 }, { "epoch": 0.69, "learning_rate": 0.0003346551724137931, "loss": 0.7676, "step": 5300 }, { "epoch": 0.69, "learning_rate": 0.00033379310344827586, "loss": 0.8122, "step": 5325 }, { "epoch": 0.7, "learning_rate": 0.0003329310344827586, "loss": 0.795, "step": 5350 }, { "epoch": 0.7, "learning_rate": 0.0003320689655172414, "loss": 0.8275, "step": 5375 }, { "epoch": 0.7, "learning_rate": 0.00033120689655172415, "loss": 0.7469, "step": 5400 }, { "epoch": 0.71, "learning_rate": 0.0003303448275862069, "loss": 0.7791, "step": 5425 }, { "epoch": 0.71, "learning_rate": 0.0003294827586206897, "loss": 0.735, "step": 5450 }, { "epoch": 0.71, "learning_rate": 0.00032862068965517244, "loss": 0.8282, "step": 5475 }, { "epoch": 0.72, "learning_rate": 0.00032775862068965515, "loss": 0.7646, "step": 5500 }, { "epoch": 0.72, "learning_rate": 0.00032689655172413797, "loss": 0.7744, "step": 5525 }, { "epoch": 0.72, "learning_rate": 0.0003260344827586207, "loss": 0.6798, "step": 5550 }, { "epoch": 0.73, "learning_rate": 0.0003251724137931035, "loss": 0.728, "step": 5575 }, { "epoch": 0.73, "learning_rate": 0.0003243103448275862, "loss": 0.715, "step": 5600 }, { "epoch": 0.73, "learning_rate": 0.00032344827586206897, "loss": 0.789, "step": 5625 }, { "epoch": 0.73, "learning_rate": 0.00032258620689655173, "loss": 0.7733, "step": 5650 }, { "epoch": 0.74, "learning_rate": 0.0003217241379310345, "loss": 0.7132, "step": 5675 }, { "epoch": 0.74, "learning_rate": 0.0003208620689655172, "loss": 0.7397, "step": 5700 }, { "epoch": 0.74, "learning_rate": 0.00032, "loss": 0.7236, "step": 5725 }, { "epoch": 0.75, "learning_rate": 0.00031913793103448274, "loss": 0.7283, "step": 5750 }, { "epoch": 0.75, "learning_rate": 0.00031827586206896555, "loss": 0.7346, "step": 5775 }, { "epoch": 0.75, "learning_rate": 0.00031741379310344826, "loss": 0.8194, "step": 5800 }, { "epoch": 0.76, "learning_rate": 0.00031655172413793103, "loss": 0.7943, "step": 5825 }, { "epoch": 0.76, "learning_rate": 0.00031568965517241385, "loss": 0.7206, "step": 5850 }, { "epoch": 0.76, "learning_rate": 0.00031482758620689655, "loss": 0.7185, "step": 5875 }, { "epoch": 0.77, "learning_rate": 0.0003139655172413793, "loss": 0.6939, "step": 5900 }, { "epoch": 0.77, "learning_rate": 0.0003131034482758621, "loss": 0.6927, "step": 5925 }, { "epoch": 0.77, "learning_rate": 0.00031224137931034485, "loss": 0.7656, "step": 5950 }, { "epoch": 0.78, "learning_rate": 0.00031137931034482756, "loss": 0.7598, "step": 5975 }, { "epoch": 0.78, "learning_rate": 0.0003105172413793104, "loss": 0.7596, "step": 6000 }, { "epoch": 0.78, "eval_loss": 0.7720913290977478, "eval_runtime": 4465.1172, "eval_samples_per_second": 1.738, "eval_steps_per_second": 0.217, "eval_wer": 38.55953220038113, "step": 6000 }, { "epoch": 0.78, "learning_rate": 0.0003096551724137931, "loss": 0.7287, "step": 6025 }, { "epoch": 0.79, "learning_rate": 0.0003087931034482759, "loss": 0.7175, "step": 6050 }, { "epoch": 0.79, "learning_rate": 0.0003079310344827586, "loss": 0.704, "step": 6075 }, { "epoch": 0.79, "learning_rate": 0.0003070689655172414, "loss": 0.6944, "step": 6100 }, { "epoch": 0.8, "learning_rate": 0.00030620689655172414, "loss": 0.6961, "step": 6125 }, { "epoch": 0.8, "learning_rate": 0.0003053448275862069, "loss": 0.6878, "step": 6150 }, { "epoch": 0.8, "learning_rate": 0.0003044827586206896, "loss": 0.6977, "step": 6175 }, { "epoch": 0.81, "learning_rate": 0.00030362068965517243, "loss": 0.711, "step": 6200 }, { "epoch": 0.81, "learning_rate": 0.0003027586206896552, "loss": 0.7098, "step": 6225 }, { "epoch": 0.81, "learning_rate": 0.00030189655172413796, "loss": 0.6776, "step": 6250 }, { "epoch": 0.82, "learning_rate": 0.0003010344827586207, "loss": 0.7545, "step": 6275 }, { "epoch": 0.82, "learning_rate": 0.00030017241379310343, "loss": 0.6484, "step": 6300 }, { "epoch": 0.82, "learning_rate": 0.00029931034482758625, "loss": 0.6543, "step": 6325 }, { "epoch": 0.83, "learning_rate": 0.00029844827586206896, "loss": 0.705, "step": 6350 }, { "epoch": 0.83, "learning_rate": 0.0002975862068965517, "loss": 0.6692, "step": 6375 }, { "epoch": 0.83, "learning_rate": 0.0002967241379310345, "loss": 0.7034, "step": 6400 }, { "epoch": 0.84, "learning_rate": 0.00029586206896551725, "loss": 0.7017, "step": 6425 }, { "epoch": 0.84, "learning_rate": 0.000295, "loss": 0.6735, "step": 6450 }, { "epoch": 0.84, "learning_rate": 0.0002941379310344828, "loss": 0.7277, "step": 6475 }, { "epoch": 0.85, "learning_rate": 0.0002932758620689655, "loss": 0.6576, "step": 6500 }, { "epoch": 0.85, "learning_rate": 0.0002924137931034483, "loss": 0.6608, "step": 6525 }, { "epoch": 0.85, "learning_rate": 0.000291551724137931, "loss": 0.7124, "step": 6550 }, { "epoch": 0.86, "learning_rate": 0.0002906896551724138, "loss": 0.7117, "step": 6575 }, { "epoch": 0.86, "learning_rate": 0.00028982758620689654, "loss": 0.6616, "step": 6600 }, { "epoch": 0.86, "learning_rate": 0.0002889655172413793, "loss": 0.6366, "step": 6625 }, { "epoch": 0.86, "learning_rate": 0.0002881034482758621, "loss": 0.6889, "step": 6650 }, { "epoch": 0.87, "learning_rate": 0.00028724137931034484, "loss": 0.6519, "step": 6675 }, { "epoch": 0.87, "learning_rate": 0.0002863793103448276, "loss": 0.6387, "step": 6700 }, { "epoch": 0.87, "learning_rate": 0.00028551724137931036, "loss": 0.6623, "step": 6725 }, { "epoch": 0.88, "learning_rate": 0.0002846551724137931, "loss": 0.6319, "step": 6750 }, { "epoch": 0.88, "learning_rate": 0.00028379310344827584, "loss": 0.6237, "step": 6775 }, { "epoch": 0.88, "learning_rate": 0.00028293103448275865, "loss": 0.6792, "step": 6800 }, { "epoch": 0.89, "learning_rate": 0.00028206896551724136, "loss": 0.6469, "step": 6825 }, { "epoch": 0.89, "learning_rate": 0.0002812068965517242, "loss": 0.6552, "step": 6850 }, { "epoch": 0.89, "learning_rate": 0.0002803448275862069, "loss": 0.6866, "step": 6875 }, { "epoch": 0.9, "learning_rate": 0.00027948275862068966, "loss": 0.6912, "step": 6900 }, { "epoch": 0.9, "learning_rate": 0.0002786206896551724, "loss": 0.6707, "step": 6925 }, { "epoch": 0.9, "learning_rate": 0.0002777586206896552, "loss": 0.6205, "step": 6950 }, { "epoch": 0.91, "learning_rate": 0.0002768965517241379, "loss": 0.6686, "step": 6975 }, { "epoch": 0.91, "learning_rate": 0.0002760344827586207, "loss": 0.5678, "step": 7000 }, { "epoch": 0.91, "eval_loss": 0.6903061866760254, "eval_runtime": 4363.4092, "eval_samples_per_second": 1.778, "eval_steps_per_second": 0.222, "eval_wer": 36.289712063069416, "step": 7000 }, { "epoch": 0.91, "learning_rate": 0.0002751724137931034, "loss": 0.6476, "step": 7025 }, { "epoch": 0.92, "learning_rate": 0.0002743103448275862, "loss": 0.5989, "step": 7050 }, { "epoch": 0.92, "learning_rate": 0.000273448275862069, "loss": 0.6408, "step": 7075 }, { "epoch": 0.92, "learning_rate": 0.0002725862068965517, "loss": 0.6082, "step": 7100 }, { "epoch": 0.93, "learning_rate": 0.00027172413793103453, "loss": 0.6294, "step": 7125 }, { "epoch": 0.93, "learning_rate": 0.00027086206896551724, "loss": 0.6513, "step": 7150 }, { "epoch": 0.93, "learning_rate": 0.00027, "loss": 0.635, "step": 7175 }, { "epoch": 0.94, "learning_rate": 0.00026913793103448277, "loss": 0.6699, "step": 7200 }, { "epoch": 0.94, "learning_rate": 0.00026827586206896553, "loss": 0.6685, "step": 7225 }, { "epoch": 0.94, "learning_rate": 0.00026741379310344824, "loss": 0.6326, "step": 7250 }, { "epoch": 0.95, "learning_rate": 0.00026655172413793106, "loss": 0.6095, "step": 7275 }, { "epoch": 0.95, "learning_rate": 0.00026568965517241377, "loss": 0.6083, "step": 7300 }, { "epoch": 0.95, "learning_rate": 0.0002648275862068966, "loss": 0.6832, "step": 7325 }, { "epoch": 0.96, "learning_rate": 0.0002639655172413793, "loss": 0.6248, "step": 7350 }, { "epoch": 0.96, "learning_rate": 0.00026310344827586206, "loss": 0.5977, "step": 7375 }, { "epoch": 0.96, "learning_rate": 0.0002622413793103448, "loss": 0.5741, "step": 7400 }, { "epoch": 0.97, "learning_rate": 0.0002613793103448276, "loss": 0.6256, "step": 7425 }, { "epoch": 0.97, "learning_rate": 0.00026051724137931035, "loss": 0.599, "step": 7450 }, { "epoch": 0.97, "learning_rate": 0.0002596551724137931, "loss": 0.6052, "step": 7475 }, { "epoch": 0.98, "learning_rate": 0.0002587931034482759, "loss": 0.653, "step": 7500 }, { "epoch": 0.98, "learning_rate": 0.00025793103448275864, "loss": 0.5683, "step": 7525 }, { "epoch": 0.98, "learning_rate": 0.0002570689655172414, "loss": 0.5726, "step": 7550 }, { "epoch": 0.99, "learning_rate": 0.0002562068965517241, "loss": 0.5458, "step": 7575 }, { "epoch": 0.99, "learning_rate": 0.00025534482758620693, "loss": 0.631, "step": 7600 }, { "epoch": 0.99, "learning_rate": 0.00025448275862068964, "loss": 0.6076, "step": 7625 }, { "epoch": 1.0, "learning_rate": 0.0002536206896551724, "loss": 0.6211, "step": 7650 }, { "epoch": 1.0, "learning_rate": 0.00025275862068965517, "loss": 0.587, "step": 7675 }, { "epoch": 1.0, "learning_rate": 0.00025189655172413794, "loss": 0.5228, "step": 7700 }, { "epoch": 1.0, "learning_rate": 0.0002510344827586207, "loss": 0.4133, "step": 7725 }, { "epoch": 1.01, "learning_rate": 0.00025017241379310346, "loss": 0.3992, "step": 7750 }, { "epoch": 1.01, "learning_rate": 0.00024931034482758623, "loss": 0.4355, "step": 7775 }, { "epoch": 1.01, "learning_rate": 0.000248448275862069, "loss": 0.43, "step": 7800 }, { "epoch": 1.02, "learning_rate": 0.00024758620689655175, "loss": 0.4121, "step": 7825 }, { "epoch": 1.02, "learning_rate": 0.00024672413793103446, "loss": 0.4258, "step": 7850 }, { "epoch": 1.02, "learning_rate": 0.00024586206896551723, "loss": 0.4448, "step": 7875 }, { "epoch": 1.03, "learning_rate": 0.000245, "loss": 0.4889, "step": 7900 }, { "epoch": 1.03, "learning_rate": 0.00024413793103448276, "loss": 0.4237, "step": 7925 }, { "epoch": 1.03, "learning_rate": 0.00024327586206896552, "loss": 0.4333, "step": 7950 }, { "epoch": 1.04, "learning_rate": 0.00024241379310344828, "loss": 0.4656, "step": 7975 }, { "epoch": 1.04, "learning_rate": 0.00024155172413793102, "loss": 0.4412, "step": 8000 }, { "epoch": 1.04, "eval_loss": 0.6475529074668884, "eval_runtime": 4254.9114, "eval_samples_per_second": 1.824, "eval_steps_per_second": 0.228, "eval_wer": 32.74727856527528, "step": 8000 }, { "epoch": 1.04, "learning_rate": 0.00024068965517241378, "loss": 0.417, "step": 8025 }, { "epoch": 1.05, "learning_rate": 0.00023982758620689658, "loss": 0.4202, "step": 8050 }, { "epoch": 1.05, "learning_rate": 0.00023896551724137934, "loss": 0.4492, "step": 8075 }, { "epoch": 1.05, "learning_rate": 0.00023810344827586208, "loss": 0.4164, "step": 8100 }, { "epoch": 1.06, "learning_rate": 0.00023724137931034484, "loss": 0.4167, "step": 8125 }, { "epoch": 1.06, "learning_rate": 0.0002363793103448276, "loss": 0.4335, "step": 8150 }, { "epoch": 1.06, "learning_rate": 0.00023551724137931037, "loss": 0.4287, "step": 8175 }, { "epoch": 1.07, "learning_rate": 0.0002346551724137931, "loss": 0.4858, "step": 8200 }, { "epoch": 1.07, "learning_rate": 0.00023379310344827587, "loss": 0.4382, "step": 8225 }, { "epoch": 1.07, "learning_rate": 0.00023293103448275863, "loss": 0.4262, "step": 8250 }, { "epoch": 1.08, "learning_rate": 0.00023206896551724137, "loss": 0.4313, "step": 8275 }, { "epoch": 1.08, "learning_rate": 0.00023120689655172413, "loss": 0.4018, "step": 8300 }, { "epoch": 1.08, "learning_rate": 0.0002303448275862069, "loss": 0.437, "step": 8325 }, { "epoch": 1.09, "learning_rate": 0.00022948275862068966, "loss": 0.4006, "step": 8350 }, { "epoch": 1.09, "learning_rate": 0.0002286206896551724, "loss": 0.4307, "step": 8375 }, { "epoch": 1.09, "learning_rate": 0.00022775862068965516, "loss": 0.4145, "step": 8400 }, { "epoch": 1.1, "learning_rate": 0.00022689655172413792, "loss": 0.4171, "step": 8425 }, { "epoch": 1.1, "learning_rate": 0.00022603448275862072, "loss": 0.4135, "step": 8450 }, { "epoch": 1.1, "learning_rate": 0.00022517241379310345, "loss": 0.3852, "step": 8475 }, { "epoch": 1.11, "learning_rate": 0.00022431034482758622, "loss": 0.4196, "step": 8500 }, { "epoch": 1.11, "learning_rate": 0.00022344827586206898, "loss": 0.4233, "step": 8525 }, { "epoch": 1.11, "learning_rate": 0.00022258620689655174, "loss": 0.4414, "step": 8550 }, { "epoch": 1.12, "learning_rate": 0.00022172413793103448, "loss": 0.4032, "step": 8575 }, { "epoch": 1.12, "learning_rate": 0.00022086206896551724, "loss": 0.4294, "step": 8600 }, { "epoch": 1.12, "learning_rate": 0.00022, "loss": 0.3927, "step": 8625 }, { "epoch": 1.13, "learning_rate": 0.00021913793103448277, "loss": 0.4101, "step": 8650 }, { "epoch": 1.13, "learning_rate": 0.0002182758620689655, "loss": 0.3791, "step": 8675 }, { "epoch": 1.13, "learning_rate": 0.00021741379310344827, "loss": 0.4372, "step": 8700 }, { "epoch": 1.13, "learning_rate": 0.00021655172413793104, "loss": 0.45, "step": 8725 }, { "epoch": 1.14, "learning_rate": 0.0002156896551724138, "loss": 0.4376, "step": 8750 }, { "epoch": 1.14, "learning_rate": 0.00021482758620689654, "loss": 0.3919, "step": 8775 }, { "epoch": 1.14, "learning_rate": 0.0002139655172413793, "loss": 0.4215, "step": 8800 }, { "epoch": 1.15, "learning_rate": 0.00021310344827586206, "loss": 0.3651, "step": 8825 }, { "epoch": 1.15, "learning_rate": 0.00021224137931034486, "loss": 0.4094, "step": 8850 }, { "epoch": 1.15, "learning_rate": 0.0002113793103448276, "loss": 0.4074, "step": 8875 }, { "epoch": 1.16, "learning_rate": 0.00021051724137931036, "loss": 0.4215, "step": 8900 }, { "epoch": 1.16, "learning_rate": 0.00020965517241379312, "loss": 0.4052, "step": 8925 }, { "epoch": 1.16, "learning_rate": 0.00020879310344827588, "loss": 0.4191, "step": 8950 }, { "epoch": 1.17, "learning_rate": 0.00020793103448275862, "loss": 0.4341, "step": 8975 }, { "epoch": 1.17, "learning_rate": 0.00020706896551724138, "loss": 0.4239, "step": 9000 }, { "epoch": 1.17, "eval_loss": 0.5972723960876465, "eval_runtime": 4291.4577, "eval_samples_per_second": 1.808, "eval_steps_per_second": 0.226, "eval_wer": 30.81421150181429, "step": 9000 }, { "epoch": 1.17, "learning_rate": 0.00020620689655172415, "loss": 0.4005, "step": 9025 }, { "epoch": 1.18, "learning_rate": 0.0002053448275862069, "loss": 0.3868, "step": 9050 }, { "epoch": 1.18, "learning_rate": 0.00020448275862068965, "loss": 0.39, "step": 9075 }, { "epoch": 1.18, "learning_rate": 0.0002036206896551724, "loss": 0.4323, "step": 9100 }, { "epoch": 1.19, "learning_rate": 0.00020275862068965518, "loss": 0.4045, "step": 9125 }, { "epoch": 1.19, "learning_rate": 0.00020189655172413794, "loss": 0.4103, "step": 9150 }, { "epoch": 1.19, "learning_rate": 0.00020103448275862068, "loss": 0.3935, "step": 9175 }, { "epoch": 1.2, "learning_rate": 0.00020017241379310344, "loss": 0.4184, "step": 9200 }, { "epoch": 1.2, "learning_rate": 0.0001993103448275862, "loss": 0.4026, "step": 9225 }, { "epoch": 1.2, "learning_rate": 0.000198448275862069, "loss": 0.3878, "step": 9250 }, { "epoch": 1.21, "learning_rate": 0.00019758620689655173, "loss": 0.3757, "step": 9275 }, { "epoch": 1.21, "learning_rate": 0.0001967241379310345, "loss": 0.3906, "step": 9300 }, { "epoch": 1.21, "learning_rate": 0.00019586206896551726, "loss": 0.3943, "step": 9325 }, { "epoch": 1.22, "learning_rate": 0.00019500000000000002, "loss": 0.4365, "step": 9350 }, { "epoch": 1.22, "learning_rate": 0.00019413793103448276, "loss": 0.4127, "step": 9375 }, { "epoch": 1.22, "learning_rate": 0.00019327586206896552, "loss": 0.3919, "step": 9400 }, { "epoch": 1.23, "learning_rate": 0.0001924137931034483, "loss": 0.3684, "step": 9425 }, { "epoch": 1.23, "learning_rate": 0.00019155172413793102, "loss": 0.3723, "step": 9450 }, { "epoch": 1.23, "learning_rate": 0.0001906896551724138, "loss": 0.431, "step": 9475 }, { "epoch": 1.24, "learning_rate": 0.00018982758620689655, "loss": 0.4256, "step": 9500 }, { "epoch": 1.24, "learning_rate": 0.00018896551724137932, "loss": 0.3996, "step": 9525 }, { "epoch": 1.24, "learning_rate": 0.00018810344827586205, "loss": 0.3945, "step": 9550 }, { "epoch": 1.25, "learning_rate": 0.00018724137931034482, "loss": 0.378, "step": 9575 }, { "epoch": 1.25, "learning_rate": 0.00018637931034482758, "loss": 0.386, "step": 9600 }, { "epoch": 1.25, "learning_rate": 0.00018551724137931034, "loss": 0.3649, "step": 9625 }, { "epoch": 1.26, "learning_rate": 0.0001846551724137931, "loss": 0.384, "step": 9650 }, { "epoch": 1.26, "learning_rate": 0.00018379310344827587, "loss": 0.417, "step": 9675 }, { "epoch": 1.26, "learning_rate": 0.00018293103448275864, "loss": 0.4006, "step": 9700 }, { "epoch": 1.26, "learning_rate": 0.0001820689655172414, "loss": 0.3675, "step": 9725 }, { "epoch": 1.27, "learning_rate": 0.00018120689655172414, "loss": 0.3888, "step": 9750 }, { "epoch": 1.27, "learning_rate": 0.0001803448275862069, "loss": 0.3747, "step": 9775 }, { "epoch": 1.27, "learning_rate": 0.00017948275862068966, "loss": 0.3842, "step": 9800 }, { "epoch": 1.28, "learning_rate": 0.00017862068965517243, "loss": 0.3928, "step": 9825 }, { "epoch": 1.28, "learning_rate": 0.00017775862068965516, "loss": 0.37, "step": 9850 }, { "epoch": 1.28, "learning_rate": 0.00017689655172413793, "loss": 0.3633, "step": 9875 }, { "epoch": 1.29, "learning_rate": 0.0001760344827586207, "loss": 0.3973, "step": 9900 }, { "epoch": 1.29, "learning_rate": 0.00017517241379310346, "loss": 0.3864, "step": 9925 }, { "epoch": 1.29, "learning_rate": 0.0001743103448275862, "loss": 0.3784, "step": 9950 }, { "epoch": 1.3, "learning_rate": 0.00017344827586206896, "loss": 0.3646, "step": 9975 }, { "epoch": 1.3, "learning_rate": 0.00017258620689655172, "loss": 0.3935, "step": 10000 }, { "epoch": 1.3, "eval_loss": 0.54443359375, "eval_runtime": 4568.2387, "eval_samples_per_second": 1.699, "eval_steps_per_second": 0.212, "eval_wer": 29.020805596888298, "step": 10000 }, { "epoch": 1.3, "learning_rate": 0.00017172413793103448, "loss": 0.3519, "step": 10025 }, { "epoch": 1.31, "learning_rate": 0.00017086206896551722, "loss": 0.3706, "step": 10050 }, { "epoch": 1.31, "learning_rate": 0.00017, "loss": 0.3678, "step": 10075 }, { "epoch": 1.31, "learning_rate": 0.00016913793103448278, "loss": 0.3964, "step": 10100 }, { "epoch": 1.32, "learning_rate": 0.00016827586206896554, "loss": 0.3489, "step": 10125 }, { "epoch": 1.32, "learning_rate": 0.00016741379310344828, "loss": 0.3341, "step": 10150 }, { "epoch": 1.32, "learning_rate": 0.00016655172413793104, "loss": 0.3754, "step": 10175 }, { "epoch": 1.33, "learning_rate": 0.0001656896551724138, "loss": 0.3617, "step": 10200 }, { "epoch": 1.33, "learning_rate": 0.00016482758620689657, "loss": 0.3565, "step": 10225 }, { "epoch": 1.33, "learning_rate": 0.000164, "loss": 0.37, "step": 10250 }, { "epoch": 1.34, "learning_rate": 0.00016313793103448277, "loss": 0.334, "step": 10275 }, { "epoch": 1.34, "learning_rate": 0.00016227586206896553, "loss": 0.3508, "step": 10300 }, { "epoch": 1.34, "learning_rate": 0.0001614137931034483, "loss": 0.351, "step": 10325 }, { "epoch": 1.35, "learning_rate": 0.00016055172413793103, "loss": 0.3416, "step": 10350 }, { "epoch": 1.35, "learning_rate": 0.0001596896551724138, "loss": 0.38, "step": 10375 }, { "epoch": 1.35, "learning_rate": 0.00015882758620689656, "loss": 0.3593, "step": 10400 }, { "epoch": 1.36, "learning_rate": 0.00015796551724137932, "loss": 0.3445, "step": 10425 }, { "epoch": 1.36, "learning_rate": 0.00015710344827586206, "loss": 0.3626, "step": 10450 }, { "epoch": 1.36, "learning_rate": 0.00015624137931034482, "loss": 0.3461, "step": 10475 }, { "epoch": 1.37, "learning_rate": 0.00015537931034482759, "loss": 0.351, "step": 10500 }, { "epoch": 1.37, "learning_rate": 0.00015451724137931032, "loss": 0.3158, "step": 10525 }, { "epoch": 1.37, "learning_rate": 0.0001536551724137931, "loss": 0.3363, "step": 10550 }, { "epoch": 1.38, "learning_rate": 0.00015279310344827588, "loss": 0.3392, "step": 10575 }, { "epoch": 1.38, "learning_rate": 0.00015193103448275864, "loss": 0.3629, "step": 10600 }, { "epoch": 1.38, "learning_rate": 0.00015106896551724138, "loss": 0.3693, "step": 10625 }, { "epoch": 1.39, "learning_rate": 0.00015020689655172414, "loss": 0.3253, "step": 10650 }, { "epoch": 1.39, "learning_rate": 0.0001493448275862069, "loss": 0.3799, "step": 10675 }, { "epoch": 1.39, "learning_rate": 0.00014848275862068967, "loss": 0.3498, "step": 10700 }, { "epoch": 1.4, "learning_rate": 0.0001476206896551724, "loss": 0.3476, "step": 10725 }, { "epoch": 1.4, "learning_rate": 0.00014675862068965517, "loss": 0.3617, "step": 10750 }, { "epoch": 1.4, "learning_rate": 0.00014589655172413793, "loss": 0.3356, "step": 10775 }, { "epoch": 1.4, "learning_rate": 0.0001450344827586207, "loss": 0.3586, "step": 10800 }, { "epoch": 1.41, "learning_rate": 0.00014417241379310343, "loss": 0.3718, "step": 10825 }, { "epoch": 1.41, "learning_rate": 0.0001433103448275862, "loss": 0.3636, "step": 10850 }, { "epoch": 1.41, "learning_rate": 0.00014244827586206896, "loss": 0.3336, "step": 10875 }, { "epoch": 1.42, "learning_rate": 0.00014158620689655173, "loss": 0.3223, "step": 10900 }, { "epoch": 1.42, "learning_rate": 0.00014072413793103446, "loss": 0.3768, "step": 10925 }, { "epoch": 1.42, "learning_rate": 0.00013986206896551723, "loss": 0.3507, "step": 10950 }, { "epoch": 1.43, "learning_rate": 0.00013900000000000002, "loss": 0.3418, "step": 10975 }, { "epoch": 1.43, "learning_rate": 0.00013813793103448278, "loss": 0.3307, "step": 11000 }, { "epoch": 1.43, "eval_loss": 0.5023880004882812, "eval_runtime": 4328.4809, "eval_samples_per_second": 1.793, "eval_steps_per_second": 0.224, "eval_wer": 27.04336022136946, "step": 11000 }, { "epoch": 1.43, "learning_rate": 0.00013727586206896552, "loss": 0.337, "step": 11025 }, { "epoch": 1.44, "learning_rate": 0.00013641379310344828, "loss": 0.3461, "step": 11050 }, { "epoch": 1.44, "learning_rate": 0.00013555172413793105, "loss": 0.3155, "step": 11075 }, { "epoch": 1.44, "learning_rate": 0.0001346896551724138, "loss": 0.3749, "step": 11100 }, { "epoch": 1.45, "learning_rate": 0.00013382758620689655, "loss": 0.3423, "step": 11125 }, { "epoch": 1.45, "learning_rate": 0.0001329655172413793, "loss": 0.3545, "step": 11150 }, { "epoch": 1.45, "learning_rate": 0.00013210344827586207, "loss": 0.345, "step": 11175 }, { "epoch": 1.46, "learning_rate": 0.00013124137931034484, "loss": 0.3202, "step": 11200 }, { "epoch": 1.46, "learning_rate": 0.00013037931034482757, "loss": 0.3508, "step": 11225 }, { "epoch": 1.46, "learning_rate": 0.00012951724137931034, "loss": 0.3255, "step": 11250 }, { "epoch": 1.47, "learning_rate": 0.0001286551724137931, "loss": 0.2894, "step": 11275 }, { "epoch": 1.47, "learning_rate": 0.00012779310344827587, "loss": 0.3045, "step": 11300 }, { "epoch": 1.47, "learning_rate": 0.0001269310344827586, "loss": 0.3172, "step": 11325 }, { "epoch": 1.48, "learning_rate": 0.00012606896551724137, "loss": 0.3262, "step": 11350 }, { "epoch": 1.48, "learning_rate": 0.00012520689655172416, "loss": 0.3419, "step": 11375 }, { "epoch": 1.48, "learning_rate": 0.0001243448275862069, "loss": 0.3341, "step": 11400 }, { "epoch": 1.49, "learning_rate": 0.00012348275862068966, "loss": 0.2947, "step": 11425 }, { "epoch": 1.49, "learning_rate": 0.00012262068965517242, "loss": 0.3074, "step": 11450 }, { "epoch": 1.49, "learning_rate": 0.00012175862068965519, "loss": 0.3026, "step": 11475 }, { "epoch": 1.5, "learning_rate": 0.00012089655172413794, "loss": 0.3012, "step": 11500 }, { "epoch": 1.5, "learning_rate": 0.0001200344827586207, "loss": 0.3195, "step": 11525 }, { "epoch": 1.5, "learning_rate": 0.00011917241379310345, "loss": 0.2823, "step": 11550 }, { "epoch": 1.51, "learning_rate": 0.00011831034482758621, "loss": 0.3612, "step": 11575 }, { "epoch": 1.51, "learning_rate": 0.00011744827586206896, "loss": 0.3005, "step": 11600 }, { "epoch": 1.51, "learning_rate": 0.00011658620689655173, "loss": 0.3733, "step": 11625 }, { "epoch": 1.52, "learning_rate": 0.00011572413793103448, "loss": 0.3113, "step": 11650 }, { "epoch": 1.52, "learning_rate": 0.00011486206896551726, "loss": 0.3026, "step": 11675 }, { "epoch": 1.52, "learning_rate": 0.000114, "loss": 0.3372, "step": 11700 }, { "epoch": 1.53, "learning_rate": 0.00011313793103448277, "loss": 0.3303, "step": 11725 }, { "epoch": 1.53, "learning_rate": 0.00011227586206896552, "loss": 0.3109, "step": 11750 }, { "epoch": 1.53, "learning_rate": 0.00011141379310344827, "loss": 0.3581, "step": 11775 }, { "epoch": 1.53, "learning_rate": 0.00011055172413793103, "loss": 0.3221, "step": 11800 }, { "epoch": 1.54, "learning_rate": 0.00010968965517241378, "loss": 0.3066, "step": 11825 }, { "epoch": 1.54, "learning_rate": 0.00010882758620689655, "loss": 0.3128, "step": 11850 }, { "epoch": 1.54, "learning_rate": 0.00010796551724137931, "loss": 0.3117, "step": 11875 }, { "epoch": 1.55, "learning_rate": 0.00010710344827586208, "loss": 0.2919, "step": 11900 }, { "epoch": 1.55, "learning_rate": 0.00010624137931034483, "loss": 0.3158, "step": 11925 }, { "epoch": 1.55, "learning_rate": 0.00010537931034482759, "loss": 0.2745, "step": 11950 }, { "epoch": 1.56, "learning_rate": 0.00010451724137931034, "loss": 0.3358, "step": 11975 }, { "epoch": 1.56, "learning_rate": 0.0001036551724137931, "loss": 0.2937, "step": 12000 }, { "epoch": 1.56, "eval_loss": 0.46081113815307617, "eval_runtime": 4356.6638, "eval_samples_per_second": 1.781, "eval_steps_per_second": 0.223, "eval_wer": 24.731772260944478, "step": 12000 }, { "epoch": 1.56, "learning_rate": 0.00010279310344827585, "loss": 0.3444, "step": 12025 }, { "epoch": 1.57, "learning_rate": 0.00010193103448275862, "loss": 0.3275, "step": 12050 }, { "epoch": 1.57, "learning_rate": 0.00010106896551724138, "loss": 0.3283, "step": 12075 }, { "epoch": 1.57, "learning_rate": 0.00010020689655172415, "loss": 0.314, "step": 12100 }, { "epoch": 1.58, "learning_rate": 9.93448275862069e-05, "loss": 0.3032, "step": 12125 }, { "epoch": 1.58, "learning_rate": 9.848275862068966e-05, "loss": 0.2962, "step": 12150 }, { "epoch": 1.58, "learning_rate": 9.762068965517241e-05, "loss": 0.2966, "step": 12175 }, { "epoch": 1.59, "learning_rate": 9.675862068965517e-05, "loss": 0.2742, "step": 12200 }, { "epoch": 1.59, "learning_rate": 9.589655172413792e-05, "loss": 0.2963, "step": 12225 }, { "epoch": 1.59, "learning_rate": 9.503448275862069e-05, "loss": 0.337, "step": 12250 }, { "epoch": 1.6, "learning_rate": 9.417241379310345e-05, "loss": 0.2903, "step": 12275 }, { "epoch": 1.6, "learning_rate": 9.331034482758622e-05, "loss": 0.291, "step": 12300 }, { "epoch": 1.6, "learning_rate": 9.244827586206897e-05, "loss": 0.2931, "step": 12325 }, { "epoch": 1.61, "learning_rate": 9.158620689655173e-05, "loss": 0.2939, "step": 12350 }, { "epoch": 1.61, "learning_rate": 9.072413793103448e-05, "loss": 0.3276, "step": 12375 }, { "epoch": 1.61, "learning_rate": 8.986206896551724e-05, "loss": 0.2787, "step": 12400 }, { "epoch": 1.62, "learning_rate": 8.9e-05, "loss": 0.2883, "step": 12425 }, { "epoch": 1.62, "learning_rate": 8.813793103448276e-05, "loss": 0.2871, "step": 12450 }, { "epoch": 1.62, "learning_rate": 8.727586206896552e-05, "loss": 0.281, "step": 12475 }, { "epoch": 1.63, "learning_rate": 8.641379310344829e-05, "loss": 0.28, "step": 12500 }, { "epoch": 1.63, "learning_rate": 8.555172413793104e-05, "loss": 0.3186, "step": 12525 }, { "epoch": 1.63, "learning_rate": 8.46896551724138e-05, "loss": 0.2893, "step": 12550 }, { "epoch": 1.64, "learning_rate": 8.382758620689655e-05, "loss": 0.3071, "step": 12575 }, { "epoch": 1.64, "learning_rate": 8.296551724137931e-05, "loss": 0.2608, "step": 12600 }, { "epoch": 1.64, "learning_rate": 8.210344827586206e-05, "loss": 0.2681, "step": 12625 }, { "epoch": 1.65, "learning_rate": 8.124137931034483e-05, "loss": 0.2982, "step": 12650 }, { "epoch": 1.65, "learning_rate": 8.037931034482759e-05, "loss": 0.296, "step": 12675 }, { "epoch": 1.65, "learning_rate": 7.951724137931036e-05, "loss": 0.2987, "step": 12700 }, { "epoch": 1.66, "learning_rate": 7.86551724137931e-05, "loss": 0.3035, "step": 12725 }, { "epoch": 1.66, "learning_rate": 7.779310344827587e-05, "loss": 0.2744, "step": 12750 }, { "epoch": 1.66, "learning_rate": 7.693103448275862e-05, "loss": 0.3077, "step": 12775 }, { "epoch": 1.66, "learning_rate": 7.606896551724138e-05, "loss": 0.265, "step": 12800 }, { "epoch": 1.67, "learning_rate": 7.520689655172413e-05, "loss": 0.3041, "step": 12825 }, { "epoch": 1.67, "learning_rate": 7.43448275862069e-05, "loss": 0.248, "step": 12850 }, { "epoch": 1.67, "learning_rate": 7.348275862068966e-05, "loss": 0.2656, "step": 12875 }, { "epoch": 1.68, "learning_rate": 7.262068965517243e-05, "loss": 0.2983, "step": 12900 }, { "epoch": 1.68, "learning_rate": 7.175862068965518e-05, "loss": 0.3198, "step": 12925 }, { "epoch": 1.68, "learning_rate": 7.089655172413793e-05, "loss": 0.2788, "step": 12950 }, { "epoch": 1.69, "learning_rate": 7.003448275862069e-05, "loss": 0.3102, "step": 12975 }, { "epoch": 1.69, "learning_rate": 6.917241379310344e-05, "loss": 0.2471, "step": 13000 }, { "epoch": 1.69, "eval_loss": 0.4258999526500702, "eval_runtime": 4325.4065, "eval_samples_per_second": 1.794, "eval_steps_per_second": 0.224, "eval_wer": 22.893988043960633, "step": 13000 }, { "epoch": 1.69, "learning_rate": 6.83103448275862e-05, "loss": 0.2669, "step": 13025 }, { "epoch": 1.7, "learning_rate": 6.744827586206896e-05, "loss": 0.2777, "step": 13050 }, { "epoch": 1.7, "learning_rate": 6.658620689655173e-05, "loss": 0.2804, "step": 13075 }, { "epoch": 1.7, "learning_rate": 6.572413793103448e-05, "loss": 0.3107, "step": 13100 }, { "epoch": 1.71, "learning_rate": 6.486206896551725e-05, "loss": 0.2575, "step": 13125 }, { "epoch": 1.71, "learning_rate": 6.4e-05, "loss": 0.2762, "step": 13150 }, { "epoch": 1.71, "learning_rate": 6.313793103448276e-05, "loss": 0.2814, "step": 13175 }, { "epoch": 1.72, "learning_rate": 6.227586206896551e-05, "loss": 0.2872, "step": 13200 }, { "epoch": 1.72, "learning_rate": 6.141379310344828e-05, "loss": 0.2988, "step": 13225 }, { "epoch": 1.72, "learning_rate": 6.055172413793104e-05, "loss": 0.2728, "step": 13250 }, { "epoch": 1.73, "learning_rate": 5.9689655172413796e-05, "loss": 0.2777, "step": 13275 }, { "epoch": 1.73, "learning_rate": 5.882758620689655e-05, "loss": 0.272, "step": 13300 }, { "epoch": 1.73, "learning_rate": 5.796551724137932e-05, "loss": 0.2952, "step": 13325 }, { "epoch": 1.74, "learning_rate": 5.7103448275862074e-05, "loss": 0.2512, "step": 13350 }, { "epoch": 1.74, "learning_rate": 5.624137931034483e-05, "loss": 0.2648, "step": 13375 }, { "epoch": 1.74, "learning_rate": 5.537931034482759e-05, "loss": 0.2759, "step": 13400 }, { "epoch": 1.75, "learning_rate": 5.451724137931035e-05, "loss": 0.2689, "step": 13425 }, { "epoch": 1.75, "learning_rate": 5.365517241379311e-05, "loss": 0.2624, "step": 13450 }, { "epoch": 1.75, "learning_rate": 5.2793103448275866e-05, "loss": 0.2853, "step": 13475 }, { "epoch": 1.76, "learning_rate": 5.193103448275862e-05, "loss": 0.2773, "step": 13500 }, { "epoch": 1.76, "learning_rate": 5.106896551724139e-05, "loss": 0.2625, "step": 13525 }, { "epoch": 1.76, "learning_rate": 5.020689655172414e-05, "loss": 0.2648, "step": 13550 }, { "epoch": 1.77, "learning_rate": 4.9344827586206894e-05, "loss": 0.2718, "step": 13575 }, { "epoch": 1.77, "learning_rate": 4.848275862068965e-05, "loss": 0.2734, "step": 13600 }, { "epoch": 1.77, "learning_rate": 4.7620689655172415e-05, "loss": 0.2511, "step": 13625 }, { "epoch": 1.78, "learning_rate": 4.675862068965517e-05, "loss": 0.2552, "step": 13650 }, { "epoch": 1.78, "learning_rate": 4.589655172413793e-05, "loss": 0.2476, "step": 13675 }, { "epoch": 1.78, "learning_rate": 4.5034482758620686e-05, "loss": 0.2851, "step": 13700 }, { "epoch": 1.79, "learning_rate": 4.417241379310345e-05, "loss": 0.2863, "step": 13725 }, { "epoch": 1.79, "learning_rate": 4.331034482758621e-05, "loss": 0.2669, "step": 13750 }, { "epoch": 1.79, "learning_rate": 4.2448275862068964e-05, "loss": 0.2711, "step": 13775 }, { "epoch": 1.8, "learning_rate": 4.158620689655172e-05, "loss": 0.2463, "step": 13800 }, { "epoch": 1.8, "learning_rate": 4.0724137931034485e-05, "loss": 0.2726, "step": 13825 }, { "epoch": 1.8, "learning_rate": 3.986206896551724e-05, "loss": 0.2507, "step": 13850 }, { "epoch": 1.8, "learning_rate": 3.9e-05, "loss": 0.2827, "step": 13875 }, { "epoch": 1.81, "learning_rate": 3.8137931034482756e-05, "loss": 0.2541, "step": 13900 }, { "epoch": 1.81, "learning_rate": 3.727586206896552e-05, "loss": 0.2527, "step": 13925 }, { "epoch": 1.81, "learning_rate": 3.641379310344828e-05, "loss": 0.2555, "step": 13950 }, { "epoch": 1.82, "learning_rate": 3.5551724137931034e-05, "loss": 0.2882, "step": 13975 }, { "epoch": 1.82, "learning_rate": 3.468965517241379e-05, "loss": 0.2357, "step": 14000 }, { "epoch": 1.82, "eval_loss": 0.3936106860637665, "eval_runtime": 4292.3141, "eval_samples_per_second": 1.808, "eval_steps_per_second": 0.226, "eval_wer": 21.60179601639387, "step": 14000 }, { "epoch": 1.82, "learning_rate": 3.382758620689655e-05, "loss": 0.2675, "step": 14025 }, { "epoch": 1.83, "learning_rate": 3.296551724137931e-05, "loss": 0.2382, "step": 14050 }, { "epoch": 1.83, "learning_rate": 3.210344827586207e-05, "loss": 0.2706, "step": 14075 }, { "epoch": 1.83, "learning_rate": 3.1241379310344826e-05, "loss": 0.2518, "step": 14100 }, { "epoch": 1.84, "learning_rate": 3.0379310344827587e-05, "loss": 0.2612, "step": 14125 }, { "epoch": 1.84, "learning_rate": 2.9517241379310344e-05, "loss": 0.2029, "step": 14150 }, { "epoch": 1.84, "learning_rate": 2.8655172413793104e-05, "loss": 0.2558, "step": 14175 }, { "epoch": 1.85, "learning_rate": 2.779310344827586e-05, "loss": 0.266, "step": 14200 }, { "epoch": 1.85, "learning_rate": 2.6931034482758622e-05, "loss": 0.2729, "step": 14225 }, { "epoch": 1.85, "learning_rate": 2.606896551724138e-05, "loss": 0.2423, "step": 14250 }, { "epoch": 1.86, "learning_rate": 2.520689655172414e-05, "loss": 0.2539, "step": 14275 }, { "epoch": 1.86, "learning_rate": 2.4344827586206896e-05, "loss": 0.2264, "step": 14300 }, { "epoch": 1.86, "learning_rate": 2.3482758620689657e-05, "loss": 0.2423, "step": 14325 }, { "epoch": 1.87, "learning_rate": 2.2620689655172414e-05, "loss": 0.2669, "step": 14350 }, { "epoch": 1.87, "learning_rate": 2.1758620689655174e-05, "loss": 0.2393, "step": 14375 }, { "epoch": 1.87, "learning_rate": 2.089655172413793e-05, "loss": 0.2602, "step": 14400 }, { "epoch": 1.88, "learning_rate": 2.0034482758620692e-05, "loss": 0.2433, "step": 14425 }, { "epoch": 1.88, "learning_rate": 1.9172413793103446e-05, "loss": 0.268, "step": 14450 }, { "epoch": 1.88, "learning_rate": 1.8344827586206896e-05, "loss": 0.2483, "step": 14475 }, { "epoch": 1.89, "learning_rate": 1.7482758620689653e-05, "loss": 0.2439, "step": 14500 }, { "epoch": 1.89, "learning_rate": 1.6620689655172414e-05, "loss": 0.2539, "step": 14525 }, { "epoch": 1.89, "learning_rate": 1.575862068965517e-05, "loss": 0.2593, "step": 14550 }, { "epoch": 1.9, "learning_rate": 1.4896551724137931e-05, "loss": 0.2572, "step": 14575 }, { "epoch": 1.9, "learning_rate": 1.403448275862069e-05, "loss": 0.2422, "step": 14600 }, { "epoch": 1.9, "learning_rate": 1.3172413793103449e-05, "loss": 0.2499, "step": 14625 }, { "epoch": 1.91, "learning_rate": 1.2310344827586208e-05, "loss": 0.2799, "step": 14650 }, { "epoch": 1.91, "learning_rate": 1.1448275862068966e-05, "loss": 0.24, "step": 14675 }, { "epoch": 1.91, "learning_rate": 1.0586206896551725e-05, "loss": 0.2371, "step": 14700 }, { "epoch": 1.92, "learning_rate": 9.724137931034484e-06, "loss": 0.2306, "step": 14725 }, { "epoch": 1.92, "learning_rate": 8.862068965517241e-06, "loss": 0.2336, "step": 14750 }, { "epoch": 1.92, "learning_rate": 8e-06, "loss": 0.2333, "step": 14775 }, { "epoch": 1.93, "learning_rate": 7.137931034482759e-06, "loss": 0.2281, "step": 14800 }, { "epoch": 1.93, "learning_rate": 6.275862068965517e-06, "loss": 0.2424, "step": 14825 }, { "epoch": 1.93, "learning_rate": 5.413793103448276e-06, "loss": 0.2271, "step": 14850 }, { "epoch": 1.93, "learning_rate": 4.551724137931035e-06, "loss": 0.2713, "step": 14875 }, { "epoch": 1.94, "learning_rate": 3.689655172413793e-06, "loss": 0.2567, "step": 14900 }, { "epoch": 1.94, "learning_rate": 2.8275862068965517e-06, "loss": 0.2722, "step": 14925 }, { "epoch": 1.94, "learning_rate": 1.96551724137931e-06, "loss": 0.2542, "step": 14950 }, { "epoch": 1.95, "learning_rate": 1.103448275862069e-06, "loss": 0.2503, "step": 14975 }, { "epoch": 1.95, "learning_rate": 2.413793103448276e-07, "loss": 0.2292, "step": 15000 }, { "epoch": 1.95, "eval_loss": 0.37756896018981934, "eval_runtime": 4311.5582, "eval_samples_per_second": 1.8, "eval_steps_per_second": 0.225, "eval_wer": 20.80037591040802, "step": 15000 }, { "epoch": 1.95, "step": 15000, "total_flos": 6.925645588119552e+19, "train_loss": 0.7551171837806702, "train_runtime": 171060.4893, "train_samples_per_second": 1.403, "train_steps_per_second": 0.088 } ], "max_steps": 15000, "num_train_epochs": 2, "total_flos": 6.925645588119552e+19, "trial_name": null, "trial_params": null }