{ "best_metric": 5.195940767715421, "best_model_checkpoint": "../whisper/whisper-medium-pt-500h/checkpoint-19000", "epoch": 5.097920858484239, "global_step": 19000, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.01, "learning_rate": 4.4e-08, "loss": 3.2635, "step": 25 }, { "epoch": 0.01, "learning_rate": 9e-08, "loss": 2.9227, "step": 50 }, { "epoch": 0.02, "learning_rate": 1.4e-07, "loss": 2.4941, "step": 75 }, { "epoch": 0.03, "learning_rate": 1.8999999999999998e-07, "loss": 2.3036, "step": 100 }, { "epoch": 0.03, "learning_rate": 2.4e-07, "loss": 2.1442, "step": 125 }, { "epoch": 0.04, "learning_rate": 2.9e-07, "loss": 1.9714, "step": 150 }, { "epoch": 0.05, "learning_rate": 3.4000000000000003e-07, "loss": 1.8159, "step": 175 }, { "epoch": 0.05, "learning_rate": 3.8999999999999997e-07, "loss": 1.6164, "step": 200 }, { "epoch": 0.06, "learning_rate": 4.3999999999999997e-07, "loss": 1.4693, "step": 225 }, { "epoch": 0.07, "learning_rate": 4.9e-07, "loss": 1.3286, "step": 250 }, { "epoch": 0.07, "learning_rate": 5.4e-07, "loss": 1.2613, "step": 275 }, { "epoch": 0.08, "learning_rate": 5.9e-07, "loss": 1.1551, "step": 300 }, { "epoch": 0.09, "learning_rate": 6.4e-07, "loss": 1.0567, "step": 325 }, { "epoch": 0.09, "learning_rate": 6.9e-07, "loss": 0.8704, "step": 350 }, { "epoch": 0.1, "learning_rate": 7.4e-07, "loss": 0.4781, "step": 375 }, { "epoch": 0.11, "learning_rate": 7.9e-07, "loss": 0.3163, "step": 400 }, { "epoch": 0.11, "learning_rate": 8.399999999999999e-07, "loss": 0.3128, "step": 425 }, { "epoch": 0.12, "learning_rate": 8.9e-07, "loss": 0.2811, "step": 450 }, { "epoch": 0.13, "learning_rate": 9.399999999999999e-07, "loss": 0.2667, "step": 475 }, { "epoch": 0.13, "learning_rate": 9.9e-07, "loss": 0.2719, "step": 500 }, { "epoch": 0.13, "eval_loss": 0.27221569418907166, "eval_runtime": 12409.0095, "eval_samples_per_second": 1.602, "eval_steps_per_second": 0.067, "eval_wer": 10.173798097942095, "step": 500 }, { "epoch": 0.14, "learning_rate": 9.978947368421053e-07, "loss": 0.265, "step": 525 }, { "epoch": 0.15, "learning_rate": 9.952631578947369e-07, "loss": 0.2669, "step": 550 }, { "epoch": 0.15, "learning_rate": 9.926315789473685e-07, "loss": 0.2514, "step": 575 }, { "epoch": 0.16, "learning_rate": 9.9e-07, "loss": 0.2517, "step": 600 }, { "epoch": 0.17, "learning_rate": 9.873684210526314e-07, "loss": 0.2411, "step": 625 }, { "epoch": 0.17, "learning_rate": 9.84736842105263e-07, "loss": 0.2517, "step": 650 }, { "epoch": 0.18, "learning_rate": 9.821052631578946e-07, "loss": 0.2447, "step": 675 }, { "epoch": 0.19, "learning_rate": 9.794736842105262e-07, "loss": 0.2426, "step": 700 }, { "epoch": 0.19, "learning_rate": 9.768421052631577e-07, "loss": 0.2414, "step": 725 }, { "epoch": 0.2, "learning_rate": 9.742105263157895e-07, "loss": 0.2221, "step": 750 }, { "epoch": 0.21, "learning_rate": 9.715789473684211e-07, "loss": 0.2327, "step": 775 }, { "epoch": 0.21, "learning_rate": 9.689473684210527e-07, "loss": 0.233, "step": 800 }, { "epoch": 0.22, "learning_rate": 9.663157894736843e-07, "loss": 0.224, "step": 825 }, { "epoch": 0.23, "learning_rate": 9.636842105263159e-07, "loss": 0.242, "step": 850 }, { "epoch": 0.23, "learning_rate": 9.610526315789472e-07, "loss": 0.2196, "step": 875 }, { "epoch": 0.24, "learning_rate": 9.584210526315788e-07, "loss": 0.2205, "step": 900 }, { "epoch": 0.25, "learning_rate": 9.557894736842104e-07, "loss": 0.2214, "step": 925 }, { "epoch": 0.25, "learning_rate": 9.531578947368421e-07, "loss": 0.2133, "step": 950 }, { "epoch": 0.26, "learning_rate": 9.505263157894737e-07, "loss": 0.2086, "step": 975 }, { "epoch": 0.27, "learning_rate": 9.478947368421051e-07, "loss": 0.2286, "step": 1000 }, { "epoch": 0.27, "eval_loss": 0.21437391638755798, "eval_runtime": 12725.5029, "eval_samples_per_second": 1.562, "eval_steps_per_second": 0.065, "eval_wer": 8.258881499426503, "step": 1000 }, { "epoch": 0.27, "learning_rate": 9.452631578947367e-07, "loss": 0.2166, "step": 1025 }, { "epoch": 0.28, "learning_rate": 9.426315789473683e-07, "loss": 0.2158, "step": 1050 }, { "epoch": 0.29, "learning_rate": 9.399999999999999e-07, "loss": 0.2082, "step": 1075 }, { "epoch": 0.3, "learning_rate": 9.373684210526316e-07, "loss": 0.202, "step": 1100 }, { "epoch": 0.3, "learning_rate": 9.347368421052632e-07, "loss": 0.2098, "step": 1125 }, { "epoch": 0.31, "learning_rate": 9.321052631578947e-07, "loss": 0.2201, "step": 1150 }, { "epoch": 0.32, "learning_rate": 9.294736842105263e-07, "loss": 0.2255, "step": 1175 }, { "epoch": 0.32, "learning_rate": 9.268421052631579e-07, "loss": 0.21, "step": 1200 }, { "epoch": 0.33, "learning_rate": 9.242105263157895e-07, "loss": 0.2068, "step": 1225 }, { "epoch": 0.34, "learning_rate": 9.21578947368421e-07, "loss": 0.2158, "step": 1250 }, { "epoch": 0.34, "learning_rate": 9.189473684210525e-07, "loss": 0.1901, "step": 1275 }, { "epoch": 0.35, "learning_rate": 9.163157894736841e-07, "loss": 0.2015, "step": 1300 }, { "epoch": 0.36, "learning_rate": 9.136842105263157e-07, "loss": 0.1992, "step": 1325 }, { "epoch": 0.36, "learning_rate": 9.110526315789473e-07, "loss": 0.1935, "step": 1350 }, { "epoch": 0.37, "learning_rate": 9.084210526315789e-07, "loss": 0.1985, "step": 1375 }, { "epoch": 0.38, "learning_rate": 9.057894736842105e-07, "loss": 0.183, "step": 1400 }, { "epoch": 0.38, "learning_rate": 9.031578947368421e-07, "loss": 0.1963, "step": 1425 }, { "epoch": 0.39, "learning_rate": 9.005263157894737e-07, "loss": 0.1888, "step": 1450 }, { "epoch": 0.4, "learning_rate": 8.978947368421053e-07, "loss": 0.1935, "step": 1475 }, { "epoch": 0.4, "learning_rate": 8.952631578947368e-07, "loss": 0.1889, "step": 1500 }, { "epoch": 0.4, "eval_loss": 0.19333799183368683, "eval_runtime": 12729.4999, "eval_samples_per_second": 1.562, "eval_steps_per_second": 0.065, "eval_wer": 7.502275988704923, "step": 1500 }, { "epoch": 0.41, "learning_rate": 8.926315789473684e-07, "loss": 0.1923, "step": 1525 }, { "epoch": 0.42, "learning_rate": 8.9e-07, "loss": 0.1895, "step": 1550 }, { "epoch": 0.42, "learning_rate": 8.873684210526315e-07, "loss": 0.1947, "step": 1575 }, { "epoch": 0.43, "learning_rate": 8.847368421052631e-07, "loss": 0.1835, "step": 1600 }, { "epoch": 0.44, "learning_rate": 8.821052631578947e-07, "loss": 0.1854, "step": 1625 }, { "epoch": 0.44, "learning_rate": 8.794736842105263e-07, "loss": 0.1863, "step": 1650 }, { "epoch": 0.45, "learning_rate": 8.768421052631579e-07, "loss": 0.1881, "step": 1675 }, { "epoch": 0.46, "learning_rate": 8.742105263157894e-07, "loss": 0.1946, "step": 1700 }, { "epoch": 0.46, "learning_rate": 8.71578947368421e-07, "loss": 0.174, "step": 1725 }, { "epoch": 0.47, "learning_rate": 8.689473684210525e-07, "loss": 0.2037, "step": 1750 }, { "epoch": 0.48, "learning_rate": 8.663157894736842e-07, "loss": 0.1741, "step": 1775 }, { "epoch": 0.48, "learning_rate": 8.636842105263158e-07, "loss": 0.1782, "step": 1800 }, { "epoch": 0.49, "learning_rate": 8.610526315789473e-07, "loss": 0.1886, "step": 1825 }, { "epoch": 0.5, "learning_rate": 8.584210526315789e-07, "loss": 0.1802, "step": 1850 }, { "epoch": 0.5, "learning_rate": 8.557894736842105e-07, "loss": 0.176, "step": 1875 }, { "epoch": 0.51, "learning_rate": 8.531578947368421e-07, "loss": 0.1913, "step": 1900 }, { "epoch": 0.52, "learning_rate": 8.505263157894737e-07, "loss": 0.1783, "step": 1925 }, { "epoch": 0.52, "learning_rate": 8.478947368421053e-07, "loss": 0.1934, "step": 1950 }, { "epoch": 0.53, "learning_rate": 8.452631578947368e-07, "loss": 0.1849, "step": 1975 }, { "epoch": 0.54, "learning_rate": 8.426315789473683e-07, "loss": 0.1711, "step": 2000 }, { "epoch": 0.54, "eval_loss": 0.18054527044296265, "eval_runtime": 12625.2245, "eval_samples_per_second": 1.575, "eval_steps_per_second": 0.066, "eval_wer": 7.031133468092439, "step": 2000 }, { "epoch": 0.54, "learning_rate": 8.399999999999999e-07, "loss": 0.1687, "step": 2025 }, { "epoch": 0.55, "learning_rate": 8.373684210526315e-07, "loss": 0.1774, "step": 2050 }, { "epoch": 0.56, "learning_rate": 8.347368421052631e-07, "loss": 0.1589, "step": 2075 }, { "epoch": 0.56, "learning_rate": 8.321052631578947e-07, "loss": 0.1813, "step": 2100 }, { "epoch": 0.57, "learning_rate": 8.294736842105263e-07, "loss": 0.1797, "step": 2125 }, { "epoch": 0.58, "learning_rate": 8.268421052631579e-07, "loss": 0.1787, "step": 2150 }, { "epoch": 0.58, "learning_rate": 8.242105263157895e-07, "loss": 0.1764, "step": 2175 }, { "epoch": 0.59, "learning_rate": 8.215789473684211e-07, "loss": 0.1763, "step": 2200 }, { "epoch": 0.6, "learning_rate": 8.189473684210527e-07, "loss": 0.1744, "step": 2225 }, { "epoch": 0.6, "learning_rate": 8.163157894736841e-07, "loss": 0.1706, "step": 2250 }, { "epoch": 0.61, "learning_rate": 8.136842105263157e-07, "loss": 0.1748, "step": 2275 }, { "epoch": 0.62, "learning_rate": 8.110526315789473e-07, "loss": 0.1673, "step": 2300 }, { "epoch": 0.62, "learning_rate": 8.084210526315789e-07, "loss": 0.1779, "step": 2325 }, { "epoch": 0.63, "learning_rate": 8.057894736842105e-07, "loss": 0.1785, "step": 2350 }, { "epoch": 0.64, "learning_rate": 8.03157894736842e-07, "loss": 0.1759, "step": 2375 }, { "epoch": 0.64, "learning_rate": 8.005263157894736e-07, "loss": 0.1669, "step": 2400 }, { "epoch": 0.65, "learning_rate": 7.978947368421052e-07, "loss": 0.1782, "step": 2425 }, { "epoch": 0.66, "learning_rate": 7.952631578947369e-07, "loss": 0.1691, "step": 2450 }, { "epoch": 0.66, "learning_rate": 7.926315789473685e-07, "loss": 0.1783, "step": 2475 }, { "epoch": 0.67, "learning_rate": 7.9e-07, "loss": 0.1634, "step": 2500 }, { "epoch": 0.67, "eval_loss": 0.1707507222890854, "eval_runtime": 12645.5618, "eval_samples_per_second": 1.572, "eval_steps_per_second": 0.066, "eval_wer": 6.7173813528374, "step": 2500 }, { "epoch": 0.68, "learning_rate": 7.873684210526315e-07, "loss": 0.1646, "step": 2525 }, { "epoch": 0.68, "learning_rate": 7.847368421052631e-07, "loss": 0.1645, "step": 2550 }, { "epoch": 0.69, "learning_rate": 7.821052631578947e-07, "loss": 0.1682, "step": 2575 }, { "epoch": 0.7, "learning_rate": 7.794736842105263e-07, "loss": 0.1843, "step": 2600 }, { "epoch": 0.7, "learning_rate": 7.768421052631579e-07, "loss": 0.1645, "step": 2625 }, { "epoch": 0.71, "learning_rate": 7.742105263157894e-07, "loss": 0.1592, "step": 2650 }, { "epoch": 0.72, "learning_rate": 7.71578947368421e-07, "loss": 0.1588, "step": 2675 }, { "epoch": 0.72, "learning_rate": 7.689473684210526e-07, "loss": 0.1722, "step": 2700 }, { "epoch": 0.73, "learning_rate": 7.663157894736842e-07, "loss": 0.1658, "step": 2725 }, { "epoch": 0.74, "learning_rate": 7.636842105263157e-07, "loss": 0.167, "step": 2750 }, { "epoch": 0.74, "learning_rate": 7.610526315789472e-07, "loss": 0.1553, "step": 2775 }, { "epoch": 0.75, "learning_rate": 7.584210526315789e-07, "loss": 0.1791, "step": 2800 }, { "epoch": 0.76, "learning_rate": 7.557894736842105e-07, "loss": 0.1631, "step": 2825 }, { "epoch": 0.76, "learning_rate": 7.531578947368421e-07, "loss": 0.172, "step": 2850 }, { "epoch": 0.77, "learning_rate": 7.505263157894737e-07, "loss": 0.162, "step": 2875 }, { "epoch": 0.78, "learning_rate": 7.478947368421053e-07, "loss": 0.1685, "step": 2900 }, { "epoch": 0.78, "learning_rate": 7.452631578947368e-07, "loss": 0.1639, "step": 2925 }, { "epoch": 0.79, "learning_rate": 7.426315789473684e-07, "loss": 0.1641, "step": 2950 }, { "epoch": 0.8, "learning_rate": 7.4e-07, "loss": 0.1525, "step": 2975 }, { "epoch": 0.8, "learning_rate": 7.373684210526315e-07, "loss": 0.1573, "step": 3000 }, { "epoch": 0.8, "eval_loss": 0.16384626924991608, "eval_runtime": 12676.1367, "eval_samples_per_second": 1.568, "eval_steps_per_second": 0.065, "eval_wer": 6.492097047129683, "step": 3000 }, { "epoch": 0.81, "learning_rate": 7.347368421052631e-07, "loss": 0.1605, "step": 3025 }, { "epoch": 0.82, "learning_rate": 7.321052631578946e-07, "loss": 0.1634, "step": 3050 }, { "epoch": 0.82, "learning_rate": 7.294736842105262e-07, "loss": 0.1743, "step": 3075 }, { "epoch": 0.83, "learning_rate": 7.268421052631578e-07, "loss": 0.1579, "step": 3100 }, { "epoch": 0.84, "learning_rate": 7.242105263157895e-07, "loss": 0.1638, "step": 3125 }, { "epoch": 0.85, "learning_rate": 7.215789473684211e-07, "loss": 0.1676, "step": 3150 }, { "epoch": 0.85, "learning_rate": 7.189473684210527e-07, "loss": 0.1524, "step": 3175 }, { "epoch": 0.86, "learning_rate": 7.163157894736842e-07, "loss": 0.1602, "step": 3200 }, { "epoch": 0.87, "learning_rate": 7.136842105263158e-07, "loss": 0.158, "step": 3225 }, { "epoch": 0.87, "learning_rate": 7.110526315789473e-07, "loss": 0.1701, "step": 3250 }, { "epoch": 0.88, "learning_rate": 7.084210526315789e-07, "loss": 0.164, "step": 3275 }, { "epoch": 0.89, "learning_rate": 7.057894736842105e-07, "loss": 0.1619, "step": 3300 }, { "epoch": 0.89, "learning_rate": 7.03157894736842e-07, "loss": 0.1661, "step": 3325 }, { "epoch": 0.9, "learning_rate": 7.005263157894736e-07, "loss": 0.1617, "step": 3350 }, { "epoch": 0.91, "learning_rate": 6.978947368421052e-07, "loss": 0.167, "step": 3375 }, { "epoch": 0.91, "learning_rate": 6.952631578947368e-07, "loss": 0.176, "step": 3400 }, { "epoch": 0.92, "learning_rate": 6.926315789473684e-07, "loss": 0.1512, "step": 3425 }, { "epoch": 0.93, "learning_rate": 6.9e-07, "loss": 0.1526, "step": 3450 }, { "epoch": 0.93, "learning_rate": 6.873684210526316e-07, "loss": 0.1549, "step": 3475 }, { "epoch": 0.94, "learning_rate": 6.847368421052631e-07, "loss": 0.1502, "step": 3500 }, { "epoch": 0.94, "eval_loss": 0.15862466394901276, "eval_runtime": 12664.1726, "eval_samples_per_second": 1.57, "eval_steps_per_second": 0.065, "eval_wer": 6.300759691597101, "step": 3500 }, { "epoch": 0.95, "learning_rate": 6.821052631578947e-07, "loss": 0.165, "step": 3525 }, { "epoch": 0.95, "learning_rate": 6.794736842105263e-07, "loss": 0.1563, "step": 3550 }, { "epoch": 0.96, "learning_rate": 6.768421052631579e-07, "loss": 0.1481, "step": 3575 }, { "epoch": 0.97, "learning_rate": 6.742105263157894e-07, "loss": 0.1563, "step": 3600 }, { "epoch": 0.97, "learning_rate": 6.71578947368421e-07, "loss": 0.1482, "step": 3625 }, { "epoch": 0.98, "learning_rate": 6.689473684210526e-07, "loss": 0.1577, "step": 3650 }, { "epoch": 0.99, "learning_rate": 6.663157894736842e-07, "loss": 0.149, "step": 3675 }, { "epoch": 0.99, "learning_rate": 6.636842105263158e-07, "loss": 0.1519, "step": 3700 }, { "epoch": 1.0, "learning_rate": 6.610526315789474e-07, "loss": 0.1456, "step": 3725 }, { "epoch": 1.01, "learning_rate": 6.584210526315788e-07, "loss": 0.1497, "step": 3750 }, { "epoch": 1.01, "learning_rate": 6.557894736842104e-07, "loss": 0.1415, "step": 3775 }, { "epoch": 1.02, "learning_rate": 6.531578947368421e-07, "loss": 0.1373, "step": 3800 }, { "epoch": 1.03, "learning_rate": 6.505263157894737e-07, "loss": 0.1265, "step": 3825 }, { "epoch": 1.03, "learning_rate": 6.478947368421053e-07, "loss": 0.1336, "step": 3850 }, { "epoch": 1.04, "learning_rate": 6.452631578947368e-07, "loss": 0.1271, "step": 3875 }, { "epoch": 1.05, "learning_rate": 6.426315789473684e-07, "loss": 0.1304, "step": 3900 }, { "epoch": 1.05, "learning_rate": 6.4e-07, "loss": 0.1273, "step": 3925 }, { "epoch": 1.06, "learning_rate": 6.373684210526316e-07, "loss": 0.131, "step": 3950 }, { "epoch": 1.07, "learning_rate": 6.347368421052632e-07, "loss": 0.1389, "step": 3975 }, { "epoch": 1.07, "learning_rate": 6.321052631578947e-07, "loss": 0.1255, "step": 4000 }, { "epoch": 1.07, "eval_loss": 0.15485483407974243, "eval_runtime": 12746.2936, "eval_samples_per_second": 1.56, "eval_steps_per_second": 0.065, "eval_wer": 6.126395811152087, "step": 4000 }, { "epoch": 1.08, "learning_rate": 6.294736842105262e-07, "loss": 0.1401, "step": 4025 }, { "epoch": 1.09, "learning_rate": 6.268421052631578e-07, "loss": 0.1338, "step": 4050 }, { "epoch": 1.09, "learning_rate": 6.242105263157894e-07, "loss": 0.1355, "step": 4075 }, { "epoch": 1.1, "learning_rate": 6.21578947368421e-07, "loss": 0.1293, "step": 4100 }, { "epoch": 1.11, "learning_rate": 6.189473684210526e-07, "loss": 0.1315, "step": 4125 }, { "epoch": 1.11, "learning_rate": 6.163157894736843e-07, "loss": 0.1361, "step": 4150 }, { "epoch": 1.12, "learning_rate": 6.136842105263158e-07, "loss": 0.1248, "step": 4175 }, { "epoch": 1.13, "learning_rate": 6.110526315789474e-07, "loss": 0.1279, "step": 4200 }, { "epoch": 1.13, "learning_rate": 6.08421052631579e-07, "loss": 0.1481, "step": 4225 }, { "epoch": 1.14, "learning_rate": 6.057894736842105e-07, "loss": 0.1309, "step": 4250 }, { "epoch": 1.15, "learning_rate": 6.031578947368421e-07, "loss": 0.1238, "step": 4275 }, { "epoch": 1.15, "learning_rate": 6.005263157894736e-07, "loss": 0.1291, "step": 4300 }, { "epoch": 1.16, "learning_rate": 5.978947368421052e-07, "loss": 0.1473, "step": 4325 }, { "epoch": 1.17, "learning_rate": 5.952631578947368e-07, "loss": 0.1425, "step": 4350 }, { "epoch": 1.17, "learning_rate": 5.926315789473684e-07, "loss": 0.1275, "step": 4375 }, { "epoch": 1.18, "learning_rate": 5.9e-07, "loss": 0.1353, "step": 4400 }, { "epoch": 1.19, "learning_rate": 5.873684210526315e-07, "loss": 0.1245, "step": 4425 }, { "epoch": 1.19, "learning_rate": 5.847368421052631e-07, "loss": 0.1384, "step": 4450 }, { "epoch": 1.2, "learning_rate": 5.821052631578948e-07, "loss": 0.1301, "step": 4475 }, { "epoch": 1.21, "learning_rate": 5.794736842105263e-07, "loss": 0.1307, "step": 4500 }, { "epoch": 1.21, "eval_loss": 0.15187527239322662, "eval_runtime": 12760.4138, "eval_samples_per_second": 1.558, "eval_steps_per_second": 0.065, "eval_wer": 6.114565813363783, "step": 4500 }, { "epoch": 1.21, "learning_rate": 5.768421052631579e-07, "loss": 0.1338, "step": 4525 }, { "epoch": 1.22, "learning_rate": 5.742105263157895e-07, "loss": 0.1349, "step": 4550 }, { "epoch": 1.23, "learning_rate": 5.71578947368421e-07, "loss": 0.124, "step": 4575 }, { "epoch": 1.23, "learning_rate": 5.689473684210526e-07, "loss": 0.1315, "step": 4600 }, { "epoch": 1.24, "learning_rate": 5.663157894736842e-07, "loss": 0.1171, "step": 4625 }, { "epoch": 1.25, "learning_rate": 5.636842105263158e-07, "loss": 0.1322, "step": 4650 }, { "epoch": 1.25, "learning_rate": 5.610526315789474e-07, "loss": 0.1252, "step": 4675 }, { "epoch": 1.26, "learning_rate": 5.584210526315789e-07, "loss": 0.1306, "step": 4700 }, { "epoch": 1.27, "learning_rate": 5.557894736842105e-07, "loss": 0.1272, "step": 4725 }, { "epoch": 1.27, "learning_rate": 5.53157894736842e-07, "loss": 0.1254, "step": 4750 }, { "epoch": 1.28, "learning_rate": 5.505263157894736e-07, "loss": 0.125, "step": 4775 }, { "epoch": 1.29, "learning_rate": 5.478947368421052e-07, "loss": 0.1268, "step": 4800 }, { "epoch": 1.29, "learning_rate": 5.452631578947369e-07, "loss": 0.1293, "step": 4825 }, { "epoch": 1.3, "learning_rate": 5.426315789473684e-07, "loss": 0.1301, "step": 4850 }, { "epoch": 1.31, "learning_rate": 5.4e-07, "loss": 0.1199, "step": 4875 }, { "epoch": 1.31, "learning_rate": 5.373684210526316e-07, "loss": 0.1269, "step": 4900 }, { "epoch": 1.32, "learning_rate": 5.347368421052632e-07, "loss": 0.1302, "step": 4925 }, { "epoch": 1.33, "learning_rate": 5.321052631578948e-07, "loss": 0.1262, "step": 4950 }, { "epoch": 1.33, "learning_rate": 5.294736842105263e-07, "loss": 0.1376, "step": 4975 }, { "epoch": 1.34, "learning_rate": 5.268421052631578e-07, "loss": 0.1173, "step": 5000 }, { "epoch": 1.34, "eval_loss": 0.14950169622898102, "eval_runtime": 12705.6418, "eval_samples_per_second": 1.565, "eval_steps_per_second": 0.065, "eval_wer": 5.9643762762253045, "step": 5000 }, { "epoch": 1.35, "learning_rate": 5.242105263157894e-07, "loss": 0.1204, "step": 5025 }, { "epoch": 1.35, "learning_rate": 5.21578947368421e-07, "loss": 0.1349, "step": 5050 }, { "epoch": 1.36, "learning_rate": 5.189473684210526e-07, "loss": 0.1267, "step": 5075 }, { "epoch": 1.37, "learning_rate": 5.163157894736841e-07, "loss": 0.1272, "step": 5100 }, { "epoch": 1.38, "learning_rate": 5.136842105263157e-07, "loss": 0.1386, "step": 5125 }, { "epoch": 1.38, "learning_rate": 5.110526315789473e-07, "loss": 0.119, "step": 5150 }, { "epoch": 1.39, "learning_rate": 5.08421052631579e-07, "loss": 0.1265, "step": 5175 }, { "epoch": 1.4, "learning_rate": 5.057894736842106e-07, "loss": 0.1228, "step": 5200 }, { "epoch": 1.4, "learning_rate": 5.031578947368422e-07, "loss": 0.1255, "step": 5225 }, { "epoch": 1.41, "learning_rate": 5.005263157894736e-07, "loss": 0.1328, "step": 5250 }, { "epoch": 1.42, "learning_rate": 4.978947368421052e-07, "loss": 0.1246, "step": 5275 }, { "epoch": 1.42, "learning_rate": 4.952631578947368e-07, "loss": 0.1345, "step": 5300 }, { "epoch": 1.43, "learning_rate": 4.926315789473684e-07, "loss": 0.1245, "step": 5325 }, { "epoch": 1.44, "learning_rate": 4.9e-07, "loss": 0.1273, "step": 5350 }, { "epoch": 1.44, "learning_rate": 4.873684210526315e-07, "loss": 0.1224, "step": 5375 }, { "epoch": 1.45, "learning_rate": 4.847368421052631e-07, "loss": 0.1435, "step": 5400 }, { "epoch": 1.46, "learning_rate": 4.821052631578947e-07, "loss": 0.1274, "step": 5425 }, { "epoch": 1.46, "learning_rate": 4.794736842105263e-07, "loss": 0.1272, "step": 5450 }, { "epoch": 1.47, "learning_rate": 4.768421052631579e-07, "loss": 0.1283, "step": 5475 }, { "epoch": 1.48, "learning_rate": 4.7421052631578946e-07, "loss": 0.1283, "step": 5500 }, { "epoch": 1.48, "eval_loss": 0.14696012437343597, "eval_runtime": 12749.031, "eval_samples_per_second": 1.559, "eval_steps_per_second": 0.065, "eval_wer": 5.853791514291152, "step": 5500 }, { "epoch": 1.48, "learning_rate": 4.7157894736842104e-07, "loss": 0.1224, "step": 5525 }, { "epoch": 1.49, "learning_rate": 4.6894736842105257e-07, "loss": 0.1326, "step": 5550 }, { "epoch": 1.5, "learning_rate": 4.6631578947368415e-07, "loss": 0.1252, "step": 5575 }, { "epoch": 1.5, "learning_rate": 4.636842105263158e-07, "loss": 0.1256, "step": 5600 }, { "epoch": 1.51, "learning_rate": 4.6105263157894737e-07, "loss": 0.1158, "step": 5625 }, { "epoch": 1.52, "learning_rate": 4.5842105263157895e-07, "loss": 0.1239, "step": 5650 }, { "epoch": 1.52, "learning_rate": 4.557894736842105e-07, "loss": 0.1332, "step": 5675 }, { "epoch": 1.53, "learning_rate": 4.5315789473684206e-07, "loss": 0.1324, "step": 5700 }, { "epoch": 1.54, "learning_rate": 4.5052631578947364e-07, "loss": 0.1366, "step": 5725 }, { "epoch": 1.54, "learning_rate": 4.478947368421053e-07, "loss": 0.1235, "step": 5750 }, { "epoch": 1.55, "learning_rate": 4.4526315789473686e-07, "loss": 0.1309, "step": 5775 }, { "epoch": 1.56, "learning_rate": 4.426315789473684e-07, "loss": 0.1381, "step": 5800 }, { "epoch": 1.56, "learning_rate": 4.3999999999999997e-07, "loss": 0.1172, "step": 5825 }, { "epoch": 1.57, "learning_rate": 4.3736842105263155e-07, "loss": 0.1239, "step": 5850 }, { "epoch": 1.58, "learning_rate": 4.3473684210526313e-07, "loss": 0.1235, "step": 5875 }, { "epoch": 1.58, "learning_rate": 4.321052631578947e-07, "loss": 0.1274, "step": 5900 }, { "epoch": 1.59, "learning_rate": 4.294736842105263e-07, "loss": 0.128, "step": 5925 }, { "epoch": 1.6, "learning_rate": 4.268421052631579e-07, "loss": 0.1244, "step": 5950 }, { "epoch": 1.6, "learning_rate": 4.2421052631578946e-07, "loss": 0.1304, "step": 5975 }, { "epoch": 1.61, "learning_rate": 4.2157894736842104e-07, "loss": 0.1279, "step": 6000 }, { "epoch": 1.61, "eval_loss": 0.14495572447776794, "eval_runtime": 12863.3039, "eval_samples_per_second": 1.545, "eval_steps_per_second": 0.064, "eval_wer": 5.83733238693351, "step": 6000 }, { "epoch": 1.62, "learning_rate": 4.189473684210526e-07, "loss": 0.1312, "step": 6025 }, { "epoch": 1.62, "learning_rate": 4.1631578947368415e-07, "loss": 0.1255, "step": 6050 }, { "epoch": 1.63, "learning_rate": 4.1368421052631574e-07, "loss": 0.1231, "step": 6075 }, { "epoch": 1.64, "learning_rate": 4.1105263157894737e-07, "loss": 0.1314, "step": 6100 }, { "epoch": 1.64, "learning_rate": 4.0842105263157895e-07, "loss": 0.1238, "step": 6125 }, { "epoch": 1.65, "learning_rate": 4.0578947368421053e-07, "loss": 0.1245, "step": 6150 }, { "epoch": 1.66, "learning_rate": 4.0315789473684206e-07, "loss": 0.1225, "step": 6175 }, { "epoch": 1.66, "learning_rate": 4.0052631578947364e-07, "loss": 0.1237, "step": 6200 }, { "epoch": 1.67, "learning_rate": 3.978947368421052e-07, "loss": 0.1233, "step": 6225 }, { "epoch": 1.68, "learning_rate": 3.952631578947368e-07, "loss": 0.12, "step": 6250 }, { "epoch": 1.68, "learning_rate": 3.9263157894736844e-07, "loss": 0.1252, "step": 6275 }, { "epoch": 1.69, "learning_rate": 3.8999999999999997e-07, "loss": 0.1194, "step": 6300 }, { "epoch": 1.7, "learning_rate": 3.8736842105263155e-07, "loss": 0.1292, "step": 6325 }, { "epoch": 1.7, "learning_rate": 3.8473684210526314e-07, "loss": 0.1211, "step": 6350 }, { "epoch": 1.71, "learning_rate": 3.821052631578947e-07, "loss": 0.1305, "step": 6375 }, { "epoch": 1.72, "learning_rate": 3.794736842105263e-07, "loss": 0.1248, "step": 6400 }, { "epoch": 1.72, "learning_rate": 3.7684210526315783e-07, "loss": 0.1188, "step": 6425 }, { "epoch": 1.73, "learning_rate": 3.7421052631578946e-07, "loss": 0.1235, "step": 6450 }, { "epoch": 1.74, "learning_rate": 3.7157894736842104e-07, "loss": 0.1269, "step": 6475 }, { "epoch": 1.74, "learning_rate": 3.6894736842105263e-07, "loss": 0.1172, "step": 6500 }, { "epoch": 1.74, "eval_loss": 0.14329521358013153, "eval_runtime": 12792.4004, "eval_samples_per_second": 1.554, "eval_steps_per_second": 0.065, "eval_wer": 5.712860236291347, "step": 6500 }, { "epoch": 1.75, "learning_rate": 3.663157894736842e-07, "loss": 0.1172, "step": 6525 }, { "epoch": 1.76, "learning_rate": 3.6368421052631574e-07, "loss": 0.1199, "step": 6550 }, { "epoch": 1.76, "learning_rate": 3.610526315789473e-07, "loss": 0.1163, "step": 6575 }, { "epoch": 1.77, "learning_rate": 3.5842105263157895e-07, "loss": 0.1261, "step": 6600 }, { "epoch": 1.78, "learning_rate": 3.5578947368421054e-07, "loss": 0.1266, "step": 6625 }, { "epoch": 1.78, "learning_rate": 3.531578947368421e-07, "loss": 0.125, "step": 6650 }, { "epoch": 1.79, "learning_rate": 3.5052631578947365e-07, "loss": 0.1187, "step": 6675 }, { "epoch": 1.8, "learning_rate": 3.4789473684210523e-07, "loss": 0.1306, "step": 6700 }, { "epoch": 1.8, "learning_rate": 3.452631578947368e-07, "loss": 0.1377, "step": 6725 }, { "epoch": 1.81, "learning_rate": 3.426315789473684e-07, "loss": 0.1224, "step": 6750 }, { "epoch": 1.82, "learning_rate": 3.4000000000000003e-07, "loss": 0.1301, "step": 6775 }, { "epoch": 1.82, "learning_rate": 3.3736842105263156e-07, "loss": 0.1222, "step": 6800 }, { "epoch": 1.83, "learning_rate": 3.3473684210526314e-07, "loss": 0.1323, "step": 6825 }, { "epoch": 1.84, "learning_rate": 3.321052631578947e-07, "loss": 0.1245, "step": 6850 }, { "epoch": 1.84, "learning_rate": 3.294736842105263e-07, "loss": 0.1133, "step": 6875 }, { "epoch": 1.85, "learning_rate": 3.268421052631579e-07, "loss": 0.1245, "step": 6900 }, { "epoch": 1.86, "learning_rate": 3.242105263157894e-07, "loss": 0.1214, "step": 6925 }, { "epoch": 1.86, "learning_rate": 3.2157894736842105e-07, "loss": 0.1199, "step": 6950 }, { "epoch": 1.87, "learning_rate": 3.1894736842105263e-07, "loss": 0.1289, "step": 6975 }, { "epoch": 1.88, "learning_rate": 3.163157894736842e-07, "loss": 0.1169, "step": 7000 }, { "epoch": 1.88, "eval_loss": 0.14210045337677002, "eval_runtime": 12770.8704, "eval_samples_per_second": 1.557, "eval_steps_per_second": 0.065, "eval_wer": 5.669140679247612, "step": 7000 }, { "epoch": 1.88, "learning_rate": 3.136842105263158e-07, "loss": 0.127, "step": 7025 }, { "epoch": 1.89, "learning_rate": 3.110526315789473e-07, "loss": 0.1229, "step": 7050 }, { "epoch": 1.9, "learning_rate": 3.084210526315789e-07, "loss": 0.1275, "step": 7075 }, { "epoch": 1.9, "learning_rate": 3.057894736842105e-07, "loss": 0.116, "step": 7100 }, { "epoch": 1.91, "learning_rate": 3.031578947368421e-07, "loss": 0.1235, "step": 7125 }, { "epoch": 1.92, "learning_rate": 3.005263157894737e-07, "loss": 0.114, "step": 7150 }, { "epoch": 1.93, "learning_rate": 2.9789473684210523e-07, "loss": 0.1143, "step": 7175 }, { "epoch": 1.93, "learning_rate": 2.952631578947368e-07, "loss": 0.127, "step": 7200 }, { "epoch": 1.94, "learning_rate": 2.926315789473684e-07, "loss": 0.1222, "step": 7225 }, { "epoch": 1.95, "learning_rate": 2.9e-07, "loss": 0.1212, "step": 7250 }, { "epoch": 1.95, "learning_rate": 2.873684210526316e-07, "loss": 0.1156, "step": 7275 }, { "epoch": 1.96, "learning_rate": 2.8473684210526314e-07, "loss": 0.131, "step": 7300 }, { "epoch": 1.97, "learning_rate": 2.821052631578947e-07, "loss": 0.1281, "step": 7325 }, { "epoch": 1.97, "learning_rate": 2.794736842105263e-07, "loss": 0.1219, "step": 7350 }, { "epoch": 1.98, "learning_rate": 2.768421052631579e-07, "loss": 0.1256, "step": 7375 }, { "epoch": 1.99, "learning_rate": 2.7421052631578947e-07, "loss": 0.124, "step": 7400 }, { "epoch": 1.99, "learning_rate": 2.71578947368421e-07, "loss": 0.132, "step": 7425 }, { "epoch": 2.0, "learning_rate": 2.6894736842105263e-07, "loss": 0.1077, "step": 7450 }, { "epoch": 2.01, "learning_rate": 2.663157894736842e-07, "loss": 0.1045, "step": 7475 }, { "epoch": 2.01, "learning_rate": 2.636842105263158e-07, "loss": 0.1042, "step": 7500 }, { "epoch": 2.01, "eval_loss": 0.14147816598415375, "eval_runtime": 12653.1516, "eval_samples_per_second": 1.571, "eval_steps_per_second": 0.066, "eval_wer": 5.626964165393656, "step": 7500 }, { "epoch": 2.02, "learning_rate": 2.610526315789474e-07, "loss": 0.115, "step": 7525 }, { "epoch": 2.03, "learning_rate": 2.584210526315789e-07, "loss": 0.1028, "step": 7550 }, { "epoch": 2.03, "learning_rate": 2.557894736842105e-07, "loss": 0.1103, "step": 7575 }, { "epoch": 2.04, "learning_rate": 2.5315789473684207e-07, "loss": 0.1187, "step": 7600 }, { "epoch": 2.05, "learning_rate": 2.505263157894737e-07, "loss": 0.1095, "step": 7625 }, { "epoch": 2.05, "learning_rate": 2.4789473684210523e-07, "loss": 0.1151, "step": 7650 }, { "epoch": 2.06, "learning_rate": 2.452631578947368e-07, "loss": 0.1068, "step": 7675 }, { "epoch": 2.07, "learning_rate": 2.426315789473684e-07, "loss": 0.1085, "step": 7700 }, { "epoch": 2.07, "learning_rate": 2.4e-07, "loss": 0.109, "step": 7725 }, { "epoch": 2.08, "learning_rate": 2.373684210526316e-07, "loss": 0.111, "step": 7750 }, { "epoch": 2.09, "learning_rate": 2.3473684210526314e-07, "loss": 0.1051, "step": 7775 }, { "epoch": 2.09, "learning_rate": 2.3210526315789472e-07, "loss": 0.1075, "step": 7800 }, { "epoch": 2.1, "learning_rate": 2.2947368421052628e-07, "loss": 0.1055, "step": 7825 }, { "epoch": 2.11, "learning_rate": 2.268421052631579e-07, "loss": 0.1115, "step": 7850 }, { "epoch": 2.11, "learning_rate": 2.2421052631578947e-07, "loss": 0.1101, "step": 7875 }, { "epoch": 2.12, "learning_rate": 2.2157894736842103e-07, "loss": 0.1098, "step": 7900 }, { "epoch": 2.13, "learning_rate": 2.1894736842105263e-07, "loss": 0.1157, "step": 7925 }, { "epoch": 2.13, "learning_rate": 2.163157894736842e-07, "loss": 0.1098, "step": 7950 }, { "epoch": 2.14, "learning_rate": 2.1368421052631577e-07, "loss": 0.121, "step": 7975 }, { "epoch": 2.15, "learning_rate": 2.1105263157894738e-07, "loss": 0.1083, "step": 8000 }, { "epoch": 2.15, "eval_loss": 0.14116224646568298, "eval_runtime": 12636.1586, "eval_samples_per_second": 1.573, "eval_steps_per_second": 0.066, "eval_wer": 5.569357219641911, "step": 8000 }, { "epoch": 2.15, "learning_rate": 2.0842105263157893e-07, "loss": 0.1026, "step": 8025 }, { "epoch": 2.16, "learning_rate": 2.0578947368421052e-07, "loss": 0.1079, "step": 8050 }, { "epoch": 2.17, "learning_rate": 2.0315789473684207e-07, "loss": 0.1089, "step": 8075 }, { "epoch": 2.17, "learning_rate": 2.0052631578947368e-07, "loss": 0.1063, "step": 8100 }, { "epoch": 2.18, "learning_rate": 1.9789473684210526e-07, "loss": 0.1205, "step": 8125 }, { "epoch": 2.19, "learning_rate": 1.9526315789473682e-07, "loss": 0.1138, "step": 8150 }, { "epoch": 2.19, "learning_rate": 1.9263157894736843e-07, "loss": 0.1147, "step": 8175 }, { "epoch": 2.2, "learning_rate": 1.8999999999999998e-07, "loss": 0.1129, "step": 8200 }, { "epoch": 2.21, "learning_rate": 1.8736842105263156e-07, "loss": 0.1106, "step": 8225 }, { "epoch": 2.21, "learning_rate": 1.8473684210526317e-07, "loss": 0.1039, "step": 8250 }, { "epoch": 2.22, "learning_rate": 1.8210526315789473e-07, "loss": 0.115, "step": 8275 }, { "epoch": 2.23, "learning_rate": 1.794736842105263e-07, "loss": 0.1023, "step": 8300 }, { "epoch": 2.23, "learning_rate": 1.7684210526315786e-07, "loss": 0.1066, "step": 8325 }, { "epoch": 2.24, "learning_rate": 1.7421052631578947e-07, "loss": 0.1172, "step": 8350 }, { "epoch": 2.25, "learning_rate": 1.7157894736842105e-07, "loss": 0.1122, "step": 8375 }, { "epoch": 2.25, "learning_rate": 1.689473684210526e-07, "loss": 0.1106, "step": 8400 }, { "epoch": 2.26, "learning_rate": 1.6631578947368422e-07, "loss": 0.104, "step": 8425 }, { "epoch": 2.27, "learning_rate": 1.6368421052631577e-07, "loss": 0.1025, "step": 8450 }, { "epoch": 2.27, "learning_rate": 1.6105263157894735e-07, "loss": 0.1108, "step": 8475 }, { "epoch": 2.28, "learning_rate": 1.5842105263157894e-07, "loss": 0.1122, "step": 8500 }, { "epoch": 2.28, "eval_loss": 0.14060835540294647, "eval_runtime": 12599.7607, "eval_samples_per_second": 1.578, "eval_steps_per_second": 0.066, "eval_wer": 5.574500696941174, "step": 8500 }, { "epoch": 2.29, "learning_rate": 1.5578947368421052e-07, "loss": 0.1135, "step": 8525 }, { "epoch": 2.29, "learning_rate": 1.531578947368421e-07, "loss": 0.1168, "step": 8550 }, { "epoch": 2.3, "learning_rate": 1.5052631578947366e-07, "loss": 0.1046, "step": 8575 }, { "epoch": 2.31, "learning_rate": 1.4789473684210526e-07, "loss": 0.1109, "step": 8600 }, { "epoch": 2.31, "learning_rate": 1.4526315789473685e-07, "loss": 0.1092, "step": 8625 }, { "epoch": 2.32, "learning_rate": 1.426315789473684e-07, "loss": 0.1093, "step": 8650 }, { "epoch": 2.33, "learning_rate": 1.4e-07, "loss": 0.1048, "step": 8675 }, { "epoch": 2.33, "learning_rate": 1.3736842105263156e-07, "loss": 0.1052, "step": 8700 }, { "epoch": 2.34, "learning_rate": 1.3473684210526315e-07, "loss": 0.1125, "step": 8725 }, { "epoch": 2.35, "learning_rate": 1.3210526315789473e-07, "loss": 0.1132, "step": 8750 }, { "epoch": 2.35, "learning_rate": 1.294736842105263e-07, "loss": 0.108, "step": 8775 }, { "epoch": 2.36, "learning_rate": 1.268421052631579e-07, "loss": 0.1072, "step": 8800 }, { "epoch": 2.37, "learning_rate": 1.2421052631578947e-07, "loss": 0.1064, "step": 8825 }, { "epoch": 2.37, "learning_rate": 1.2157894736842106e-07, "loss": 0.1006, "step": 8850 }, { "epoch": 2.38, "learning_rate": 1.1894736842105262e-07, "loss": 0.1123, "step": 8875 }, { "epoch": 2.39, "learning_rate": 1.163157894736842e-07, "loss": 0.1095, "step": 8900 }, { "epoch": 2.39, "learning_rate": 1.1368421052631579e-07, "loss": 0.1155, "step": 8925 }, { "epoch": 2.4, "learning_rate": 1.1105263157894737e-07, "loss": 0.1095, "step": 8950 }, { "epoch": 2.41, "learning_rate": 1.0842105263157894e-07, "loss": 0.099, "step": 8975 }, { "epoch": 2.41, "learning_rate": 1.0578947368421052e-07, "loss": 0.0958, "step": 9000 }, { "epoch": 2.41, "eval_loss": 0.14025309681892395, "eval_runtime": 12563.2725, "eval_samples_per_second": 1.582, "eval_steps_per_second": 0.066, "eval_wer": 5.567814176452132, "step": 9000 }, { "epoch": 2.42, "learning_rate": 1.031578947368421e-07, "loss": 0.1099, "step": 9025 }, { "epoch": 2.43, "learning_rate": 1.0052631578947368e-07, "loss": 0.1121, "step": 9050 }, { "epoch": 2.43, "learning_rate": 9.789473684210525e-08, "loss": 0.1051, "step": 9075 }, { "epoch": 2.44, "learning_rate": 9.526315789473683e-08, "loss": 0.1112, "step": 9100 }, { "epoch": 2.45, "learning_rate": 9.263157894736842e-08, "loss": 0.104, "step": 9125 }, { "epoch": 2.45, "learning_rate": 9e-08, "loss": 0.1087, "step": 9150 }, { "epoch": 2.46, "learning_rate": 8.736842105263158e-08, "loss": 0.1007, "step": 9175 }, { "epoch": 2.47, "learning_rate": 8.473684210526315e-08, "loss": 0.1068, "step": 9200 }, { "epoch": 2.48, "learning_rate": 8.210526315789473e-08, "loss": 0.1048, "step": 9225 }, { "epoch": 2.48, "learning_rate": 7.947368421052631e-08, "loss": 0.1116, "step": 9250 }, { "epoch": 2.49, "learning_rate": 7.68421052631579e-08, "loss": 0.1111, "step": 9275 }, { "epoch": 2.5, "learning_rate": 7.421052631578948e-08, "loss": 0.1064, "step": 9300 }, { "epoch": 2.5, "learning_rate": 7.157894736842105e-08, "loss": 0.1032, "step": 9325 }, { "epoch": 2.51, "learning_rate": 6.894736842105263e-08, "loss": 0.1067, "step": 9350 }, { "epoch": 2.52, "learning_rate": 6.631578947368421e-08, "loss": 0.1129, "step": 9375 }, { "epoch": 2.52, "learning_rate": 6.368421052631579e-08, "loss": 0.1115, "step": 9400 }, { "epoch": 2.53, "learning_rate": 6.105263157894736e-08, "loss": 0.1059, "step": 9425 }, { "epoch": 2.54, "learning_rate": 5.842105263157895e-08, "loss": 0.1136, "step": 9450 }, { "epoch": 2.54, "learning_rate": 5.578947368421052e-08, "loss": 0.1092, "step": 9475 }, { "epoch": 2.55, "learning_rate": 5.3157894736842105e-08, "loss": 0.1046, "step": 9500 }, { "epoch": 2.55, "eval_loss": 0.1400066465139389, "eval_runtime": 12501.569, "eval_samples_per_second": 1.59, "eval_steps_per_second": 0.066, "eval_wer": 5.569357219641911, "step": 9500 }, { "epoch": 2.56, "learning_rate": 5.052631578947368e-08, "loss": 0.1018, "step": 9525 }, { "epoch": 2.56, "learning_rate": 4.789473684210526e-08, "loss": 0.1055, "step": 9550 }, { "epoch": 2.57, "learning_rate": 4.5263157894736844e-08, "loss": 0.1112, "step": 9575 }, { "epoch": 2.58, "learning_rate": 4.263157894736842e-08, "loss": 0.1049, "step": 9600 }, { "epoch": 2.58, "learning_rate": 4e-08, "loss": 0.1099, "step": 9625 }, { "epoch": 2.59, "learning_rate": 3.7368421052631576e-08, "loss": 0.099, "step": 9650 }, { "epoch": 2.6, "learning_rate": 3.473684210526316e-08, "loss": 0.1102, "step": 9675 }, { "epoch": 2.6, "learning_rate": 3.210526315789473e-08, "loss": 0.1114, "step": 9700 }, { "epoch": 2.61, "learning_rate": 2.9473684210526315e-08, "loss": 0.106, "step": 9725 }, { "epoch": 2.62, "learning_rate": 2.6842105263157894e-08, "loss": 0.1151, "step": 9750 }, { "epoch": 2.62, "learning_rate": 2.4210526315789472e-08, "loss": 0.1141, "step": 9775 }, { "epoch": 2.63, "learning_rate": 2.157894736842105e-08, "loss": 0.1032, "step": 9800 }, { "epoch": 2.64, "learning_rate": 1.8947368421052632e-08, "loss": 0.1123, "step": 9825 }, { "epoch": 2.64, "learning_rate": 1.631578947368421e-08, "loss": 0.1068, "step": 9850 }, { "epoch": 2.65, "learning_rate": 1.368421052631579e-08, "loss": 0.1122, "step": 9875 }, { "epoch": 2.66, "learning_rate": 1.1052631578947366e-08, "loss": 0.1034, "step": 9900 }, { "epoch": 2.66, "learning_rate": 8.421052631578947e-09, "loss": 0.1081, "step": 9925 }, { "epoch": 2.67, "learning_rate": 5.789473684210526e-09, "loss": 0.1045, "step": 9950 }, { "epoch": 2.68, "learning_rate": 3.157894736842105e-09, "loss": 0.1051, "step": 9975 }, { "epoch": 2.68, "learning_rate": 5.263157894736842e-10, "loss": 0.1086, "step": 10000 }, { "epoch": 2.68, "eval_loss": 0.13990232348442078, "eval_runtime": 12547.9361, "eval_samples_per_second": 1.584, "eval_steps_per_second": 0.066, "eval_wer": 5.560613308233164, "step": 10000 }, { "epoch": 2.69, "learning_rate": 5.117948717948718e-07, "loss": 0.1052, "step": 10025 }, { "epoch": 2.7, "learning_rate": 5.105128205128205e-07, "loss": 0.1085, "step": 10050 }, { "epoch": 2.7, "learning_rate": 5.092307692307692e-07, "loss": 0.0999, "step": 10075 }, { "epoch": 2.71, "learning_rate": 5.079487179487179e-07, "loss": 0.1005, "step": 10100 }, { "epoch": 2.72, "learning_rate": 5.066666666666667e-07, "loss": 0.1041, "step": 10125 }, { "epoch": 2.72, "learning_rate": 5.053846153846154e-07, "loss": 0.1053, "step": 10150 }, { "epoch": 2.73, "learning_rate": 5.041025641025641e-07, "loss": 0.1074, "step": 10175 }, { "epoch": 2.74, "learning_rate": 5.028205128205127e-07, "loss": 0.1136, "step": 10200 }, { "epoch": 2.74, "learning_rate": 5.015384615384614e-07, "loss": 0.1137, "step": 10225 }, { "epoch": 2.75, "learning_rate": 5.002564102564102e-07, "loss": 0.1087, "step": 10250 }, { "epoch": 2.76, "learning_rate": 4.98974358974359e-07, "loss": 0.1154, "step": 10275 }, { "epoch": 2.76, "learning_rate": 4.976923076923077e-07, "loss": 0.0969, "step": 10300 }, { "epoch": 2.77, "learning_rate": 4.964102564102564e-07, "loss": 0.1105, "step": 10325 }, { "epoch": 2.78, "learning_rate": 4.951282051282051e-07, "loss": 0.1063, "step": 10350 }, { "epoch": 2.78, "learning_rate": 4.938461538461538e-07, "loss": 0.1052, "step": 10375 }, { "epoch": 2.79, "learning_rate": 4.925641025641025e-07, "loss": 0.1085, "step": 10400 }, { "epoch": 2.8, "learning_rate": 4.912820512820512e-07, "loss": 0.1037, "step": 10425 }, { "epoch": 2.8, "learning_rate": 4.9e-07, "loss": 0.1043, "step": 10450 }, { "epoch": 2.81, "learning_rate": 4.887179487179487e-07, "loss": 0.1037, "step": 10475 }, { "epoch": 2.82, "learning_rate": 4.874358974358974e-07, "loss": 0.1181, "step": 10500 }, { "epoch": 2.82, "eval_loss": 0.13893495500087738, "eval_runtime": 12469.4463, "eval_samples_per_second": 1.594, "eval_steps_per_second": 0.066, "eval_wer": 5.522037228488692, "step": 10500 }, { "epoch": 2.82, "learning_rate": 4.861538461538461e-07, "loss": 0.1154, "step": 10525 }, { "epoch": 2.83, "learning_rate": 4.848717948717949e-07, "loss": 0.1084, "step": 10550 }, { "epoch": 2.84, "learning_rate": 4.835897435897436e-07, "loss": 0.1113, "step": 10575 }, { "epoch": 2.84, "learning_rate": 4.823076923076923e-07, "loss": 0.1122, "step": 10600 }, { "epoch": 2.85, "learning_rate": 4.81025641025641e-07, "loss": 0.097, "step": 10625 }, { "epoch": 2.86, "learning_rate": 4.797435897435898e-07, "loss": 0.1073, "step": 10650 }, { "epoch": 2.86, "learning_rate": 4.784615384615384e-07, "loss": 0.103, "step": 10675 }, { "epoch": 2.87, "learning_rate": 4.771794871794871e-07, "loss": 0.1061, "step": 10700 }, { "epoch": 2.88, "learning_rate": 4.758974358974359e-07, "loss": 0.1067, "step": 10725 }, { "epoch": 2.88, "learning_rate": 4.7461538461538456e-07, "loss": 0.1095, "step": 10750 }, { "epoch": 2.89, "learning_rate": 4.733333333333333e-07, "loss": 0.1086, "step": 10775 }, { "epoch": 2.9, "learning_rate": 4.72051282051282e-07, "loss": 0.1037, "step": 10800 }, { "epoch": 2.9, "learning_rate": 4.7076923076923073e-07, "loss": 0.1111, "step": 10825 }, { "epoch": 2.91, "learning_rate": 4.6948717948717946e-07, "loss": 0.1092, "step": 10850 }, { "epoch": 2.92, "learning_rate": 4.682051282051282e-07, "loss": 0.1053, "step": 10875 }, { "epoch": 2.92, "learning_rate": 4.669230769230769e-07, "loss": 0.1032, "step": 10900 }, { "epoch": 2.93, "learning_rate": 4.6564102564102563e-07, "loss": 0.1022, "step": 10925 }, { "epoch": 2.94, "learning_rate": 4.6435897435897436e-07, "loss": 0.1053, "step": 10950 }, { "epoch": 2.94, "learning_rate": 4.6307692307692303e-07, "loss": 0.1136, "step": 10975 }, { "epoch": 2.95, "learning_rate": 4.6179487179487175e-07, "loss": 0.1152, "step": 11000 }, { "epoch": 2.95, "eval_loss": 0.13824784755706787, "eval_runtime": 12480.7034, "eval_samples_per_second": 1.593, "eval_steps_per_second": 0.066, "eval_wer": 5.504549405671198, "step": 11000 }, { "epoch": 2.96, "learning_rate": 4.605128205128205e-07, "loss": 0.1089, "step": 11025 }, { "epoch": 2.96, "learning_rate": 4.592307692307692e-07, "loss": 0.0989, "step": 11050 }, { "epoch": 2.97, "learning_rate": 4.5794871794871793e-07, "loss": 0.1003, "step": 11075 }, { "epoch": 2.98, "learning_rate": 4.5666666666666665e-07, "loss": 0.1034, "step": 11100 }, { "epoch": 2.98, "learning_rate": 4.553846153846154e-07, "loss": 0.1066, "step": 11125 }, { "epoch": 2.99, "learning_rate": 4.541025641025641e-07, "loss": 0.1065, "step": 11150 }, { "epoch": 3.0, "learning_rate": 4.5282051282051283e-07, "loss": 0.1095, "step": 11175 }, { "epoch": 3.01, "learning_rate": 4.5153846153846155e-07, "loss": 0.108, "step": 11200 }, { "epoch": 3.01, "learning_rate": 4.502564102564103e-07, "loss": 0.0952, "step": 11225 }, { "epoch": 3.02, "learning_rate": 4.48974358974359e-07, "loss": 0.102, "step": 11250 }, { "epoch": 3.03, "learning_rate": 4.476923076923076e-07, "loss": 0.1073, "step": 11275 }, { "epoch": 3.03, "learning_rate": 4.4641025641025635e-07, "loss": 0.1031, "step": 11300 }, { "epoch": 3.04, "learning_rate": 4.4512820512820507e-07, "loss": 0.0969, "step": 11325 }, { "epoch": 3.05, "learning_rate": 4.438461538461538e-07, "loss": 0.0972, "step": 11350 }, { "epoch": 3.05, "learning_rate": 4.425641025641025e-07, "loss": 0.114, "step": 11375 }, { "epoch": 3.06, "learning_rate": 4.4128205128205124e-07, "loss": 0.092, "step": 11400 }, { "epoch": 3.07, "learning_rate": 4.3999999999999997e-07, "loss": 0.0981, "step": 11425 }, { "epoch": 3.07, "learning_rate": 4.387179487179487e-07, "loss": 0.1024, "step": 11450 }, { "epoch": 3.08, "learning_rate": 4.374358974358974e-07, "loss": 0.0922, "step": 11475 }, { "epoch": 3.09, "learning_rate": 4.3615384615384614e-07, "loss": 0.1075, "step": 11500 }, { "epoch": 3.09, "eval_loss": 0.13769695162773132, "eval_runtime": 12481.6798, "eval_samples_per_second": 1.593, "eval_steps_per_second": 0.066, "eval_wer": 5.445399416729675, "step": 11500 }, { "epoch": 3.09, "learning_rate": 4.3487179487179487e-07, "loss": 0.1117, "step": 11525 }, { "epoch": 3.1, "learning_rate": 4.335897435897436e-07, "loss": 0.0971, "step": 11550 }, { "epoch": 3.11, "learning_rate": 4.3230769230769227e-07, "loss": 0.0947, "step": 11575 }, { "epoch": 3.11, "learning_rate": 4.31025641025641e-07, "loss": 0.1104, "step": 11600 }, { "epoch": 3.12, "learning_rate": 4.297435897435897e-07, "loss": 0.1037, "step": 11625 }, { "epoch": 3.13, "learning_rate": 4.2846153846153844e-07, "loss": 0.0994, "step": 11650 }, { "epoch": 3.13, "learning_rate": 4.2717948717948716e-07, "loss": 0.099, "step": 11675 }, { "epoch": 3.14, "learning_rate": 4.258974358974359e-07, "loss": 0.0935, "step": 11700 }, { "epoch": 3.15, "learning_rate": 4.246153846153846e-07, "loss": 0.1005, "step": 11725 }, { "epoch": 3.15, "learning_rate": 4.2333333333333334e-07, "loss": 0.1038, "step": 11750 }, { "epoch": 3.16, "learning_rate": 4.2205128205128206e-07, "loss": 0.0944, "step": 11775 }, { "epoch": 3.17, "learning_rate": 4.207692307692308e-07, "loss": 0.1006, "step": 11800 }, { "epoch": 3.17, "learning_rate": 4.194871794871795e-07, "loss": 0.1054, "step": 11825 }, { "epoch": 3.18, "learning_rate": 4.182051282051282e-07, "loss": 0.1003, "step": 11850 }, { "epoch": 3.19, "learning_rate": 4.1692307692307686e-07, "loss": 0.0904, "step": 11875 }, { "epoch": 3.19, "learning_rate": 4.156410256410256e-07, "loss": 0.0955, "step": 11900 }, { "epoch": 3.2, "learning_rate": 4.143589743589743e-07, "loss": 0.098, "step": 11925 }, { "epoch": 3.21, "learning_rate": 4.1307692307692303e-07, "loss": 0.0962, "step": 11950 }, { "epoch": 3.21, "learning_rate": 4.1179487179487176e-07, "loss": 0.0974, "step": 11975 }, { "epoch": 3.22, "learning_rate": 4.105128205128205e-07, "loss": 0.0974, "step": 12000 }, { "epoch": 3.22, "eval_loss": 0.1371021270751953, "eval_runtime": 12563.4615, "eval_samples_per_second": 1.582, "eval_steps_per_second": 0.066, "eval_wer": 5.4176246393136545, "step": 12000 }, { "epoch": 3.23, "learning_rate": 4.092307692307692e-07, "loss": 0.1024, "step": 12025 }, { "epoch": 3.23, "learning_rate": 4.0794871794871793e-07, "loss": 0.1001, "step": 12050 }, { "epoch": 3.24, "learning_rate": 4.0666666666666666e-07, "loss": 0.0998, "step": 12075 }, { "epoch": 3.25, "learning_rate": 4.053846153846154e-07, "loss": 0.1013, "step": 12100 }, { "epoch": 3.25, "learning_rate": 4.041025641025641e-07, "loss": 0.1021, "step": 12125 }, { "epoch": 3.26, "learning_rate": 4.0282051282051283e-07, "loss": 0.098, "step": 12150 }, { "epoch": 3.27, "learning_rate": 4.015384615384615e-07, "loss": 0.0942, "step": 12175 }, { "epoch": 3.27, "learning_rate": 4.002564102564102e-07, "loss": 0.0984, "step": 12200 }, { "epoch": 3.28, "learning_rate": 3.9897435897435895e-07, "loss": 0.0981, "step": 12225 }, { "epoch": 3.29, "learning_rate": 3.976923076923077e-07, "loss": 0.0966, "step": 12250 }, { "epoch": 3.29, "learning_rate": 3.964102564102564e-07, "loss": 0.0992, "step": 12275 }, { "epoch": 3.3, "learning_rate": 3.951282051282051e-07, "loss": 0.1219, "step": 12300 }, { "epoch": 3.31, "learning_rate": 3.9384615384615385e-07, "loss": 0.104, "step": 12325 }, { "epoch": 3.31, "learning_rate": 3.925641025641026e-07, "loss": 0.1059, "step": 12350 }, { "epoch": 3.32, "learning_rate": 3.9128205128205125e-07, "loss": 0.0958, "step": 12375 }, { "epoch": 3.33, "learning_rate": 3.8999999999999997e-07, "loss": 0.0964, "step": 12400 }, { "epoch": 3.33, "learning_rate": 3.887179487179487e-07, "loss": 0.0979, "step": 12425 }, { "epoch": 3.34, "learning_rate": 3.874358974358974e-07, "loss": 0.0995, "step": 12450 }, { "epoch": 3.35, "learning_rate": 3.8615384615384615e-07, "loss": 0.1015, "step": 12475 }, { "epoch": 3.35, "learning_rate": 3.848717948717948e-07, "loss": 0.0991, "step": 12500 }, { "epoch": 3.35, "eval_loss": 0.13588958978652954, "eval_runtime": 12780.0758, "eval_samples_per_second": 1.555, "eval_steps_per_second": 0.065, "eval_wer": 5.383677689138519, "step": 12500 }, { "epoch": 3.36, "learning_rate": 3.8358974358974354e-07, "loss": 0.0994, "step": 12525 }, { "epoch": 3.37, "learning_rate": 3.8230769230769227e-07, "loss": 0.0949, "step": 12550 }, { "epoch": 3.37, "learning_rate": 3.81025641025641e-07, "loss": 0.1044, "step": 12575 }, { "epoch": 3.38, "learning_rate": 3.797435897435897e-07, "loss": 0.0957, "step": 12600 }, { "epoch": 3.39, "learning_rate": 3.7846153846153844e-07, "loss": 0.0968, "step": 12625 }, { "epoch": 3.39, "learning_rate": 3.7717948717948717e-07, "loss": 0.089, "step": 12650 }, { "epoch": 3.4, "learning_rate": 3.758974358974359e-07, "loss": 0.1012, "step": 12675 }, { "epoch": 3.41, "learning_rate": 3.746153846153846e-07, "loss": 0.0943, "step": 12700 }, { "epoch": 3.41, "learning_rate": 3.7333333333333334e-07, "loss": 0.0996, "step": 12725 }, { "epoch": 3.42, "learning_rate": 3.7205128205128207e-07, "loss": 0.0962, "step": 12750 }, { "epoch": 3.43, "learning_rate": 3.707692307692308e-07, "loss": 0.1005, "step": 12775 }, { "epoch": 3.43, "learning_rate": 3.6953846153846153e-07, "loss": 0.0994, "step": 12800 }, { "epoch": 3.44, "learning_rate": 3.6825641025641026e-07, "loss": 0.0981, "step": 12825 }, { "epoch": 3.45, "learning_rate": 3.66974358974359e-07, "loss": 0.094, "step": 12850 }, { "epoch": 3.45, "learning_rate": 3.656923076923077e-07, "loss": 0.1006, "step": 12875 }, { "epoch": 3.46, "learning_rate": 3.6441025641025643e-07, "loss": 0.097, "step": 12900 }, { "epoch": 3.47, "learning_rate": 3.6312820512820516e-07, "loss": 0.0975, "step": 12925 }, { "epoch": 3.47, "learning_rate": 3.618461538461539e-07, "loss": 0.1024, "step": 12950 }, { "epoch": 3.48, "learning_rate": 3.605641025641025e-07, "loss": 0.0977, "step": 12975 }, { "epoch": 3.49, "learning_rate": 3.592820512820512e-07, "loss": 0.1025, "step": 13000 }, { "epoch": 3.49, "eval_loss": 0.13534535467624664, "eval_runtime": 12747.167, "eval_samples_per_second": 1.559, "eval_steps_per_second": 0.065, "eval_wer": 5.373390734539993, "step": 13000 }, { "epoch": 3.49, "learning_rate": 3.5799999999999995e-07, "loss": 0.0962, "step": 13025 }, { "epoch": 3.5, "learning_rate": 3.567179487179487e-07, "loss": 0.0923, "step": 13050 }, { "epoch": 3.51, "learning_rate": 3.554358974358974e-07, "loss": 0.0983, "step": 13075 }, { "epoch": 3.51, "learning_rate": 3.541538461538461e-07, "loss": 0.0981, "step": 13100 }, { "epoch": 3.52, "learning_rate": 3.5287179487179485e-07, "loss": 0.0936, "step": 13125 }, { "epoch": 3.53, "learning_rate": 3.515897435897436e-07, "loss": 0.1052, "step": 13150 }, { "epoch": 3.53, "learning_rate": 3.503076923076923e-07, "loss": 0.0964, "step": 13175 }, { "epoch": 3.54, "learning_rate": 3.49025641025641e-07, "loss": 0.098, "step": 13200 }, { "epoch": 3.55, "learning_rate": 3.4774358974358975e-07, "loss": 0.1038, "step": 13225 }, { "epoch": 3.56, "learning_rate": 3.4646153846153847e-07, "loss": 0.106, "step": 13250 }, { "epoch": 3.56, "learning_rate": 3.4517948717948714e-07, "loss": 0.0992, "step": 13275 }, { "epoch": 3.57, "learning_rate": 3.4389743589743587e-07, "loss": 0.1083, "step": 13300 }, { "epoch": 3.58, "learning_rate": 3.426153846153846e-07, "loss": 0.1033, "step": 13325 }, { "epoch": 3.58, "learning_rate": 3.413333333333333e-07, "loss": 0.0947, "step": 13350 }, { "epoch": 3.59, "learning_rate": 3.4005128205128204e-07, "loss": 0.1011, "step": 13375 }, { "epoch": 3.6, "learning_rate": 3.3876923076923077e-07, "loss": 0.0897, "step": 13400 }, { "epoch": 3.6, "learning_rate": 3.374871794871795e-07, "loss": 0.0937, "step": 13425 }, { "epoch": 3.61, "learning_rate": 3.362051282051282e-07, "loss": 0.0965, "step": 13450 }, { "epoch": 3.62, "learning_rate": 3.3492307692307694e-07, "loss": 0.1017, "step": 13475 }, { "epoch": 3.62, "learning_rate": 3.3364102564102567e-07, "loss": 0.0973, "step": 13500 }, { "epoch": 3.62, "eval_loss": 0.1346551775932312, "eval_runtime": 12748.9576, "eval_samples_per_second": 1.559, "eval_steps_per_second": 0.065, "eval_wer": 5.331214220686037, "step": 13500 }, { "epoch": 3.63, "learning_rate": 3.3235897435897434e-07, "loss": 0.0971, "step": 13525 }, { "epoch": 3.64, "learning_rate": 3.3107692307692306e-07, "loss": 0.1029, "step": 13550 }, { "epoch": 3.64, "learning_rate": 3.2979487179487174e-07, "loss": 0.0955, "step": 13575 }, { "epoch": 3.65, "learning_rate": 3.2851282051282046e-07, "loss": 0.0907, "step": 13600 }, { "epoch": 3.66, "learning_rate": 3.272307692307692e-07, "loss": 0.086, "step": 13625 }, { "epoch": 3.66, "learning_rate": 3.259487179487179e-07, "loss": 0.0979, "step": 13650 }, { "epoch": 3.67, "learning_rate": 3.2466666666666664e-07, "loss": 0.0945, "step": 13675 }, { "epoch": 3.68, "learning_rate": 3.2338461538461536e-07, "loss": 0.0869, "step": 13700 }, { "epoch": 3.68, "learning_rate": 3.221025641025641e-07, "loss": 0.1068, "step": 13725 }, { "epoch": 3.69, "learning_rate": 3.208205128205128e-07, "loss": 0.1098, "step": 13750 }, { "epoch": 3.7, "learning_rate": 3.1953846153846153e-07, "loss": 0.1014, "step": 13775 }, { "epoch": 3.7, "learning_rate": 3.1825641025641026e-07, "loss": 0.0995, "step": 13800 }, { "epoch": 3.71, "learning_rate": 3.16974358974359e-07, "loss": 0.1083, "step": 13825 }, { "epoch": 3.72, "learning_rate": 3.156923076923077e-07, "loss": 0.1026, "step": 13850 }, { "epoch": 3.72, "learning_rate": 3.144102564102564e-07, "loss": 0.0953, "step": 13875 }, { "epoch": 3.73, "learning_rate": 3.131282051282051e-07, "loss": 0.0977, "step": 13900 }, { "epoch": 3.74, "learning_rate": 3.1184615384615383e-07, "loss": 0.1029, "step": 13925 }, { "epoch": 3.74, "learning_rate": 3.1056410256410256e-07, "loss": 0.098, "step": 13950 }, { "epoch": 3.75, "learning_rate": 3.092820512820513e-07, "loss": 0.0999, "step": 13975 }, { "epoch": 3.76, "learning_rate": 3.08e-07, "loss": 0.0938, "step": 14000 }, { "epoch": 3.76, "eval_loss": 0.13395904004573822, "eval_runtime": 12554.5443, "eval_samples_per_second": 1.583, "eval_steps_per_second": 0.066, "eval_wer": 5.306011181919649, "step": 14000 }, { "epoch": 3.76, "learning_rate": 3.0671794871794873e-07, "loss": 0.0887, "step": 14025 }, { "epoch": 3.77, "learning_rate": 3.054358974358974e-07, "loss": 0.1075, "step": 14050 }, { "epoch": 3.78, "learning_rate": 3.0415384615384613e-07, "loss": 0.0979, "step": 14075 }, { "epoch": 3.78, "learning_rate": 3.0287179487179485e-07, "loss": 0.0931, "step": 14100 }, { "epoch": 3.79, "learning_rate": 3.015897435897436e-07, "loss": 0.1023, "step": 14125 }, { "epoch": 3.8, "learning_rate": 3.003076923076923e-07, "loss": 0.0986, "step": 14150 }, { "epoch": 3.8, "learning_rate": 2.9902564102564097e-07, "loss": 0.0949, "step": 14175 }, { "epoch": 3.81, "learning_rate": 2.977435897435897e-07, "loss": 0.0976, "step": 14200 }, { "epoch": 3.82, "learning_rate": 2.964615384615384e-07, "loss": 0.1003, "step": 14225 }, { "epoch": 3.82, "learning_rate": 2.9517948717948715e-07, "loss": 0.0934, "step": 14250 }, { "epoch": 3.83, "learning_rate": 2.9389743589743587e-07, "loss": 0.0967, "step": 14275 }, { "epoch": 3.84, "learning_rate": 2.926153846153846e-07, "loss": 0.0927, "step": 14300 }, { "epoch": 3.84, "learning_rate": 2.913333333333333e-07, "loss": 0.0981, "step": 14325 }, { "epoch": 3.85, "learning_rate": 2.9005128205128205e-07, "loss": 0.0986, "step": 14350 }, { "epoch": 3.86, "learning_rate": 2.8876923076923077e-07, "loss": 0.0946, "step": 14375 }, { "epoch": 3.86, "learning_rate": 2.874871794871795e-07, "loss": 0.0939, "step": 14400 }, { "epoch": 3.87, "learning_rate": 2.862051282051282e-07, "loss": 0.0978, "step": 14425 }, { "epoch": 3.88, "learning_rate": 2.8492307692307695e-07, "loss": 0.1037, "step": 14450 }, { "epoch": 3.88, "learning_rate": 2.8364102564102567e-07, "loss": 0.0908, "step": 14475 }, { "epoch": 3.89, "learning_rate": 2.8235897435897434e-07, "loss": 0.0967, "step": 14500 }, { "epoch": 3.89, "eval_loss": 0.13375158607959747, "eval_runtime": 12788.5636, "eval_samples_per_second": 1.554, "eval_steps_per_second": 0.065, "eval_wer": 5.267435102175177, "step": 14500 }, { "epoch": 3.9, "learning_rate": 2.8107692307692307e-07, "loss": 0.0957, "step": 14525 }, { "epoch": 3.9, "learning_rate": 2.797948717948718e-07, "loss": 0.0958, "step": 14550 }, { "epoch": 3.91, "learning_rate": 2.7851282051282046e-07, "loss": 0.0969, "step": 14575 }, { "epoch": 3.92, "learning_rate": 2.772307692307692e-07, "loss": 0.0992, "step": 14600 }, { "epoch": 3.92, "learning_rate": 2.759487179487179e-07, "loss": 0.1019, "step": 14625 }, { "epoch": 3.93, "learning_rate": 2.7466666666666664e-07, "loss": 0.0914, "step": 14650 }, { "epoch": 3.94, "learning_rate": 2.7338461538461536e-07, "loss": 0.1043, "step": 14675 }, { "epoch": 3.94, "learning_rate": 2.721025641025641e-07, "loss": 0.1023, "step": 14700 }, { "epoch": 3.95, "learning_rate": 2.708205128205128e-07, "loss": 0.0936, "step": 14725 }, { "epoch": 3.96, "learning_rate": 2.6953846153846154e-07, "loss": 0.1004, "step": 14750 }, { "epoch": 3.96, "learning_rate": 2.6825641025641026e-07, "loss": 0.0935, "step": 14775 }, { "epoch": 3.97, "learning_rate": 2.6697435897435893e-07, "loss": 0.1067, "step": 14800 }, { "epoch": 3.98, "learning_rate": 2.6569230769230766e-07, "loss": 0.1018, "step": 14825 }, { "epoch": 3.98, "learning_rate": 2.644102564102564e-07, "loss": 0.0898, "step": 14850 }, { "epoch": 3.99, "learning_rate": 2.631282051282051e-07, "loss": 0.0969, "step": 14875 }, { "epoch": 4.0, "learning_rate": 2.6184615384615383e-07, "loss": 0.106, "step": 14900 }, { "epoch": 4.0, "learning_rate": 2.6056410256410256e-07, "loss": 0.0915, "step": 14925 }, { "epoch": 4.01, "learning_rate": 2.592820512820513e-07, "loss": 0.0916, "step": 14950 }, { "epoch": 4.02, "learning_rate": 2.58e-07, "loss": 0.0877, "step": 14975 }, { "epoch": 4.02, "learning_rate": 2.5671794871794873e-07, "loss": 0.095, "step": 15000 }, { "epoch": 4.02, "eval_loss": 0.13309867680072784, "eval_runtime": 12777.662, "eval_samples_per_second": 1.556, "eval_steps_per_second": 0.065, "eval_wer": 5.259719886226282, "step": 15000 }, { "epoch": 4.03, "learning_rate": 2.5543589743589746e-07, "loss": 0.0897, "step": 15025 }, { "epoch": 4.04, "learning_rate": 2.542051282051282e-07, "loss": 0.0792, "step": 15050 }, { "epoch": 4.04, "learning_rate": 2.529230769230769e-07, "loss": 0.0902, "step": 15075 }, { "epoch": 4.05, "learning_rate": 2.5164102564102565e-07, "loss": 0.089, "step": 15100 }, { "epoch": 4.06, "learning_rate": 2.5035897435897437e-07, "loss": 0.0876, "step": 15125 }, { "epoch": 4.06, "learning_rate": 2.490769230769231e-07, "loss": 0.0968, "step": 15150 }, { "epoch": 4.07, "learning_rate": 2.4779487179487177e-07, "loss": 0.0925, "step": 15175 }, { "epoch": 4.08, "learning_rate": 2.465128205128205e-07, "loss": 0.0935, "step": 15200 }, { "epoch": 4.09, "learning_rate": 2.452307692307692e-07, "loss": 0.0835, "step": 15225 }, { "epoch": 4.09, "learning_rate": 2.4394871794871794e-07, "loss": 0.0904, "step": 15250 }, { "epoch": 4.1, "learning_rate": 2.4266666666666667e-07, "loss": 0.0828, "step": 15275 }, { "epoch": 4.11, "learning_rate": 2.413846153846154e-07, "loss": 0.0925, "step": 15300 }, { "epoch": 4.11, "learning_rate": 2.401025641025641e-07, "loss": 0.0864, "step": 15325 }, { "epoch": 4.12, "learning_rate": 2.388205128205128e-07, "loss": 0.0982, "step": 15350 }, { "epoch": 4.13, "learning_rate": 2.3753846153846152e-07, "loss": 0.088, "step": 15375 }, { "epoch": 4.13, "learning_rate": 2.3625641025641024e-07, "loss": 0.0829, "step": 15400 }, { "epoch": 4.14, "learning_rate": 2.3497435897435896e-07, "loss": 0.0819, "step": 15425 }, { "epoch": 4.15, "learning_rate": 2.336923076923077e-07, "loss": 0.0877, "step": 15450 }, { "epoch": 4.15, "learning_rate": 2.3241025641025641e-07, "loss": 0.0877, "step": 15475 }, { "epoch": 4.16, "learning_rate": 2.311282051282051e-07, "loss": 0.0843, "step": 15500 }, { "epoch": 4.16, "eval_loss": 0.13372430205345154, "eval_runtime": 12897.3, "eval_samples_per_second": 1.541, "eval_steps_per_second": 0.064, "eval_wer": 5.375962473189625, "step": 15500 }, { "epoch": 4.17, "learning_rate": 2.2984615384615384e-07, "loss": 0.0946, "step": 15525 }, { "epoch": 4.17, "learning_rate": 2.2856410256410256e-07, "loss": 0.0846, "step": 15550 }, { "epoch": 4.18, "learning_rate": 2.2728205128205126e-07, "loss": 0.0938, "step": 15575 }, { "epoch": 4.19, "learning_rate": 2.2599999999999999e-07, "loss": 0.0913, "step": 15600 }, { "epoch": 4.19, "learning_rate": 2.247179487179487e-07, "loss": 0.0953, "step": 15625 }, { "epoch": 4.2, "learning_rate": 2.234358974358974e-07, "loss": 0.0842, "step": 15650 }, { "epoch": 4.21, "learning_rate": 2.2215384615384613e-07, "loss": 0.084, "step": 15675 }, { "epoch": 4.21, "learning_rate": 2.2087179487179486e-07, "loss": 0.0884, "step": 15700 }, { "epoch": 4.22, "learning_rate": 2.1958974358974358e-07, "loss": 0.0868, "step": 15725 }, { "epoch": 4.23, "learning_rate": 2.183076923076923e-07, "loss": 0.0831, "step": 15750 }, { "epoch": 4.23, "learning_rate": 2.1702564102564103e-07, "loss": 0.0846, "step": 15775 }, { "epoch": 4.24, "learning_rate": 2.1574358974358973e-07, "loss": 0.0861, "step": 15800 }, { "epoch": 4.25, "learning_rate": 2.1446153846153846e-07, "loss": 0.0842, "step": 15825 }, { "epoch": 4.25, "learning_rate": 2.1317948717948715e-07, "loss": 0.0891, "step": 15850 }, { "epoch": 4.26, "learning_rate": 2.1189743589743588e-07, "loss": 0.0868, "step": 15875 }, { "epoch": 4.27, "learning_rate": 2.106153846153846e-07, "loss": 0.0781, "step": 15900 }, { "epoch": 4.27, "learning_rate": 2.0933333333333333e-07, "loss": 0.084, "step": 15925 }, { "epoch": 4.28, "learning_rate": 2.0805128205128203e-07, "loss": 0.0866, "step": 15950 }, { "epoch": 4.29, "learning_rate": 2.0676923076923075e-07, "loss": 0.0804, "step": 15975 }, { "epoch": 4.29, "learning_rate": 2.0548717948717948e-07, "loss": 0.0977, "step": 16000 }, { "epoch": 4.29, "eval_loss": 0.13362078368663788, "eval_runtime": 12651.7783, "eval_samples_per_second": 1.571, "eval_steps_per_second": 0.066, "eval_wer": 5.245318149788345, "step": 16000 }, { "epoch": 4.3, "learning_rate": 2.042051282051282e-07, "loss": 0.0833, "step": 16025 }, { "epoch": 4.31, "learning_rate": 2.0292307692307693e-07, "loss": 0.092, "step": 16050 }, { "epoch": 4.31, "learning_rate": 2.0164102564102565e-07, "loss": 0.0835, "step": 16075 }, { "epoch": 4.32, "learning_rate": 2.0035897435897435e-07, "loss": 0.0874, "step": 16100 }, { "epoch": 4.33, "learning_rate": 1.9907692307692305e-07, "loss": 0.096, "step": 16125 }, { "epoch": 4.33, "learning_rate": 1.9779487179487177e-07, "loss": 0.0864, "step": 16150 }, { "epoch": 4.34, "learning_rate": 1.965128205128205e-07, "loss": 0.0828, "step": 16175 }, { "epoch": 4.35, "learning_rate": 1.9523076923076922e-07, "loss": 0.0839, "step": 16200 }, { "epoch": 4.35, "learning_rate": 1.9394871794871795e-07, "loss": 0.0849, "step": 16225 }, { "epoch": 4.36, "learning_rate": 1.9266666666666667e-07, "loss": 0.0947, "step": 16250 }, { "epoch": 4.37, "learning_rate": 1.9138461538461537e-07, "loss": 0.0895, "step": 16275 }, { "epoch": 4.37, "learning_rate": 1.901025641025641e-07, "loss": 0.0841, "step": 16300 }, { "epoch": 4.38, "learning_rate": 1.8882051282051282e-07, "loss": 0.0826, "step": 16325 }, { "epoch": 4.39, "learning_rate": 1.8753846153846154e-07, "loss": 0.0877, "step": 16350 }, { "epoch": 4.39, "learning_rate": 1.8625641025641027e-07, "loss": 0.0914, "step": 16375 }, { "epoch": 4.4, "learning_rate": 1.8497435897435897e-07, "loss": 0.0849, "step": 16400 }, { "epoch": 4.41, "learning_rate": 1.8369230769230767e-07, "loss": 0.0913, "step": 16425 }, { "epoch": 4.41, "learning_rate": 1.824102564102564e-07, "loss": 0.0927, "step": 16450 }, { "epoch": 4.42, "learning_rate": 1.8112820512820512e-07, "loss": 0.0935, "step": 16475 }, { "epoch": 4.43, "learning_rate": 1.7984615384615384e-07, "loss": 0.1008, "step": 16500 }, { "epoch": 4.43, "eval_loss": 0.1329377144575119, "eval_runtime": 12583.6843, "eval_samples_per_second": 1.58, "eval_steps_per_second": 0.066, "eval_wer": 5.227315979240926, "step": 16500 }, { "epoch": 4.43, "learning_rate": 1.7856410256410256e-07, "loss": 0.0873, "step": 16525 }, { "epoch": 4.44, "learning_rate": 1.772820512820513e-07, "loss": 0.0849, "step": 16550 }, { "epoch": 4.45, "learning_rate": 1.76e-07, "loss": 0.0864, "step": 16575 }, { "epoch": 4.45, "learning_rate": 1.747179487179487e-07, "loss": 0.0844, "step": 16600 }, { "epoch": 4.46, "learning_rate": 1.7343589743589744e-07, "loss": 0.089, "step": 16625 }, { "epoch": 4.47, "learning_rate": 1.7215384615384616e-07, "loss": 0.0854, "step": 16650 }, { "epoch": 4.47, "learning_rate": 1.7087179487179486e-07, "loss": 0.0903, "step": 16675 }, { "epoch": 4.48, "learning_rate": 1.6958974358974359e-07, "loss": 0.0958, "step": 16700 }, { "epoch": 4.49, "learning_rate": 1.6830769230769228e-07, "loss": 0.0883, "step": 16725 }, { "epoch": 4.49, "learning_rate": 1.67025641025641e-07, "loss": 0.0842, "step": 16750 }, { "epoch": 4.5, "learning_rate": 1.6574358974358973e-07, "loss": 0.0817, "step": 16775 }, { "epoch": 4.51, "learning_rate": 1.6446153846153846e-07, "loss": 0.0889, "step": 16800 }, { "epoch": 4.51, "learning_rate": 1.6317948717948718e-07, "loss": 0.0807, "step": 16825 }, { "epoch": 4.52, "learning_rate": 1.618974358974359e-07, "loss": 0.0873, "step": 16850 }, { "epoch": 4.53, "learning_rate": 1.606153846153846e-07, "loss": 0.0895, "step": 16875 }, { "epoch": 4.53, "learning_rate": 1.5933333333333333e-07, "loss": 0.0842, "step": 16900 }, { "epoch": 4.54, "learning_rate": 1.5805128205128203e-07, "loss": 0.0915, "step": 16925 }, { "epoch": 4.55, "learning_rate": 1.5676923076923075e-07, "loss": 0.0882, "step": 16950 }, { "epoch": 4.55, "learning_rate": 1.5548717948717948e-07, "loss": 0.0833, "step": 16975 }, { "epoch": 4.56, "learning_rate": 1.542051282051282e-07, "loss": 0.0861, "step": 17000 }, { "epoch": 4.56, "eval_loss": 0.13320882618427277, "eval_runtime": 12674.6505, "eval_samples_per_second": 1.568, "eval_steps_per_second": 0.065, "eval_wer": 5.231945108810263, "step": 17000 }, { "epoch": 4.57, "learning_rate": 1.529230769230769e-07, "loss": 0.0832, "step": 17025 }, { "epoch": 4.57, "learning_rate": 1.5164102564102563e-07, "loss": 0.0876, "step": 17050 }, { "epoch": 4.58, "learning_rate": 1.5035897435897435e-07, "loss": 0.0839, "step": 17075 }, { "epoch": 4.59, "learning_rate": 1.4907692307692308e-07, "loss": 0.093, "step": 17100 }, { "epoch": 4.59, "learning_rate": 1.477948717948718e-07, "loss": 0.0922, "step": 17125 }, { "epoch": 4.6, "learning_rate": 1.4656410256410257e-07, "loss": 0.0906, "step": 17150 }, { "epoch": 4.61, "learning_rate": 1.452820512820513e-07, "loss": 0.0802, "step": 17175 }, { "epoch": 4.61, "learning_rate": 1.44e-07, "loss": 0.0806, "step": 17200 }, { "epoch": 4.62, "learning_rate": 1.4271794871794872e-07, "loss": 0.085, "step": 17225 }, { "epoch": 4.63, "learning_rate": 1.4143589743589742e-07, "loss": 0.0887, "step": 17250 }, { "epoch": 4.64, "learning_rate": 1.4015384615384614e-07, "loss": 0.086, "step": 17275 }, { "epoch": 4.64, "learning_rate": 1.3887179487179487e-07, "loss": 0.0871, "step": 17300 }, { "epoch": 4.65, "learning_rate": 1.375897435897436e-07, "loss": 0.0811, "step": 17325 }, { "epoch": 4.66, "learning_rate": 1.363076923076923e-07, "loss": 0.0947, "step": 17350 }, { "epoch": 4.66, "learning_rate": 1.35025641025641e-07, "loss": 0.0925, "step": 17375 }, { "epoch": 4.67, "learning_rate": 1.3374358974358974e-07, "loss": 0.0866, "step": 17400 }, { "epoch": 4.68, "learning_rate": 1.3246153846153846e-07, "loss": 0.0899, "step": 17425 }, { "epoch": 4.68, "learning_rate": 1.311794871794872e-07, "loss": 0.0816, "step": 17450 }, { "epoch": 4.69, "learning_rate": 1.298974358974359e-07, "loss": 0.0899, "step": 17475 }, { "epoch": 4.7, "learning_rate": 1.286153846153846e-07, "loss": 0.0922, "step": 17500 }, { "epoch": 4.7, "eval_loss": 0.13287433981895447, "eval_runtime": 12683.7713, "eval_samples_per_second": 1.567, "eval_steps_per_second": 0.065, "eval_wer": 5.220115111021958, "step": 17500 }, { "epoch": 4.7, "learning_rate": 1.273333333333333e-07, "loss": 0.0827, "step": 17525 }, { "epoch": 4.71, "learning_rate": 1.2605128205128203e-07, "loss": 0.0846, "step": 17550 }, { "epoch": 4.72, "learning_rate": 1.2476923076923076e-07, "loss": 0.0915, "step": 17575 }, { "epoch": 4.72, "learning_rate": 1.2348717948717948e-07, "loss": 0.1012, "step": 17600 }, { "epoch": 4.73, "learning_rate": 1.222051282051282e-07, "loss": 0.0916, "step": 17625 }, { "epoch": 4.74, "learning_rate": 1.209230769230769e-07, "loss": 0.0888, "step": 17650 }, { "epoch": 4.74, "learning_rate": 1.1964102564102563e-07, "loss": 0.0858, "step": 17675 }, { "epoch": 4.75, "learning_rate": 1.1835897435897436e-07, "loss": 0.0884, "step": 17700 }, { "epoch": 4.76, "learning_rate": 1.1707692307692307e-07, "loss": 0.0916, "step": 17725 }, { "epoch": 4.76, "learning_rate": 1.1579487179487179e-07, "loss": 0.0878, "step": 17750 }, { "epoch": 4.77, "learning_rate": 1.145128205128205e-07, "loss": 0.0825, "step": 17775 }, { "epoch": 4.78, "learning_rate": 1.1323076923076923e-07, "loss": 0.0918, "step": 17800 }, { "epoch": 4.78, "learning_rate": 1.1194871794871795e-07, "loss": 0.0821, "step": 17825 }, { "epoch": 4.79, "learning_rate": 1.1066666666666665e-07, "loss": 0.092, "step": 17850 }, { "epoch": 4.8, "learning_rate": 1.0938461538461538e-07, "loss": 0.0797, "step": 17875 }, { "epoch": 4.8, "learning_rate": 1.081025641025641e-07, "loss": 0.0837, "step": 17900 }, { "epoch": 4.81, "learning_rate": 1.0682051282051281e-07, "loss": 0.091, "step": 17925 }, { "epoch": 4.82, "learning_rate": 1.0553846153846154e-07, "loss": 0.0882, "step": 17950 }, { "epoch": 4.82, "learning_rate": 1.0425641025641026e-07, "loss": 0.0903, "step": 17975 }, { "epoch": 4.83, "learning_rate": 1.0297435897435896e-07, "loss": 0.0884, "step": 18000 }, { "epoch": 4.83, "eval_loss": 0.13286374509334564, "eval_runtime": 12680.9254, "eval_samples_per_second": 1.568, "eval_steps_per_second": 0.065, "eval_wer": 5.22114380648181, "step": 18000 }, { "epoch": 4.84, "learning_rate": 1.0169230769230769e-07, "loss": 0.0848, "step": 18025 }, { "epoch": 4.84, "learning_rate": 1.0041025641025641e-07, "loss": 0.0859, "step": 18050 }, { "epoch": 4.85, "learning_rate": 9.912820512820512e-08, "loss": 0.0826, "step": 18075 }, { "epoch": 4.86, "learning_rate": 9.784615384615385e-08, "loss": 0.0864, "step": 18100 }, { "epoch": 4.86, "learning_rate": 9.656410256410256e-08, "loss": 0.0788, "step": 18125 }, { "epoch": 4.87, "learning_rate": 9.528205128205127e-08, "loss": 0.0867, "step": 18150 }, { "epoch": 4.88, "learning_rate": 9.4e-08, "loss": 0.0828, "step": 18175 }, { "epoch": 4.88, "learning_rate": 9.271794871794872e-08, "loss": 0.0905, "step": 18200 }, { "epoch": 4.89, "learning_rate": 9.143589743589743e-08, "loss": 0.0895, "step": 18225 }, { "epoch": 4.9, "learning_rate": 9.015384615384616e-08, "loss": 0.0875, "step": 18250 }, { "epoch": 4.9, "learning_rate": 8.887179487179487e-08, "loss": 0.0877, "step": 18275 }, { "epoch": 4.91, "learning_rate": 8.758974358974358e-08, "loss": 0.0865, "step": 18300 }, { "epoch": 4.92, "learning_rate": 8.63076923076923e-08, "loss": 0.0961, "step": 18325 }, { "epoch": 4.92, "learning_rate": 8.502564102564103e-08, "loss": 0.0949, "step": 18350 }, { "epoch": 4.93, "learning_rate": 8.374358974358974e-08, "loss": 0.0974, "step": 18375 }, { "epoch": 4.94, "learning_rate": 8.246153846153845e-08, "loss": 0.0862, "step": 18400 }, { "epoch": 4.94, "learning_rate": 8.117948717948718e-08, "loss": 0.0882, "step": 18425 }, { "epoch": 4.95, "learning_rate": 7.989743589743589e-08, "loss": 0.0877, "step": 18450 }, { "epoch": 4.96, "learning_rate": 7.861538461538461e-08, "loss": 0.087, "step": 18475 }, { "epoch": 4.96, "learning_rate": 7.733333333333334e-08, "loss": 0.0897, "step": 18500 }, { "epoch": 4.96, "eval_loss": 0.13251684606075287, "eval_runtime": 12652.0252, "eval_samples_per_second": 1.571, "eval_steps_per_second": 0.066, "eval_wer": 5.207770765503726, "step": 18500 }, { "epoch": 4.97, "learning_rate": 7.605128205128204e-08, "loss": 0.0884, "step": 18525 }, { "epoch": 4.98, "learning_rate": 7.476923076923076e-08, "loss": 0.0919, "step": 18550 }, { "epoch": 4.98, "learning_rate": 7.348717948717949e-08, "loss": 0.0764, "step": 18575 }, { "epoch": 4.99, "learning_rate": 7.22051282051282e-08, "loss": 0.0947, "step": 18600 }, { "epoch": 5.0, "learning_rate": 7.092307692307692e-08, "loss": 0.0919, "step": 18625 }, { "epoch": 5.0, "learning_rate": 6.964102564102565e-08, "loss": 0.077, "step": 18650 }, { "epoch": 5.01, "learning_rate": 6.835897435897434e-08, "loss": 0.0793, "step": 18675 }, { "epoch": 5.02, "learning_rate": 6.707692307692307e-08, "loss": 0.0792, "step": 18700 }, { "epoch": 5.02, "learning_rate": 6.57948717948718e-08, "loss": 0.0856, "step": 18725 }, { "epoch": 5.03, "learning_rate": 6.45128205128205e-08, "loss": 0.0895, "step": 18750 }, { "epoch": 5.04, "learning_rate": 6.323076923076923e-08, "loss": 0.0786, "step": 18775 }, { "epoch": 5.04, "learning_rate": 6.194871794871794e-08, "loss": 0.0846, "step": 18800 }, { "epoch": 5.05, "learning_rate": 6.066666666666667e-08, "loss": 0.0781, "step": 18825 }, { "epoch": 5.06, "learning_rate": 5.938461538461538e-08, "loss": 0.0761, "step": 18850 }, { "epoch": 5.06, "learning_rate": 5.8102564102564104e-08, "loss": 0.086, "step": 18875 }, { "epoch": 5.07, "learning_rate": 5.6820512820512815e-08, "loss": 0.0718, "step": 18900 }, { "epoch": 5.08, "learning_rate": 5.5538461538461533e-08, "loss": 0.0816, "step": 18925 }, { "epoch": 5.08, "learning_rate": 5.425641025641026e-08, "loss": 0.0854, "step": 18950 }, { "epoch": 5.09, "learning_rate": 5.297435897435897e-08, "loss": 0.0889, "step": 18975 }, { "epoch": 5.1, "learning_rate": 5.169230769230769e-08, "loss": 0.0851, "step": 19000 }, { "epoch": 5.1, "eval_loss": 0.13268667459487915, "eval_runtime": 12638.2717, "eval_samples_per_second": 1.573, "eval_steps_per_second": 0.066, "eval_wer": 5.195940767715421, "step": 19000 } ], "max_steps": 20000, "num_train_epochs": 6, "total_flos": 9.308529119526912e+20, "trial_name": null, "trial_params": null }