{ "best_metric": null, "best_model_checkpoint": null, "epoch": 19.216436126932464, "eval_steps": 100, "global_step": 11800, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.02, "learning_rate": 1.4999999999999999e-05, "loss": 11.8163, "step": 10 }, { "epoch": 0.03, "learning_rate": 2.8499999999999998e-05, "loss": 11.8929, "step": 20 }, { "epoch": 0.05, "learning_rate": 4.3499999999999993e-05, "loss": 12.0064, "step": 30 }, { "epoch": 0.07, "learning_rate": 5.6999999999999996e-05, "loss": 10.5132, "step": 40 }, { "epoch": 0.08, "learning_rate": 7.199999999999999e-05, "loss": 7.9699, "step": 50 }, { "epoch": 0.1, "learning_rate": 8.699999999999999e-05, "loss": 4.7703, "step": 60 }, { "epoch": 0.11, "learning_rate": 0.000102, "loss": 3.7982, "step": 70 }, { "epoch": 0.13, "learning_rate": 0.000117, "loss": 3.436, "step": 80 }, { "epoch": 0.15, "learning_rate": 0.00013199999999999998, "loss": 3.2573, "step": 90 }, { "epoch": 0.16, "learning_rate": 0.000147, "loss": 3.1412, "step": 100 }, { "epoch": 0.16, "eval_loss": 3.0855324268341064, "eval_runtime": 1038.5545, "eval_samples_per_second": 11.178, "eval_steps_per_second": 0.699, "eval_wer": 1.0, "step": 100 }, { "epoch": 0.18, "learning_rate": 0.000162, "loss": 3.0577, "step": 110 }, { "epoch": 0.2, "learning_rate": 0.00017699999999999997, "loss": 3.0443, "step": 120 }, { "epoch": 0.21, "learning_rate": 0.00019199999999999998, "loss": 3.0597, "step": 130 }, { "epoch": 0.23, "learning_rate": 0.00020699999999999996, "loss": 3.061, "step": 140 }, { "epoch": 0.24, "learning_rate": 0.00022199999999999998, "loss": 3.0717, "step": 150 }, { "epoch": 0.26, "learning_rate": 0.000237, "loss": 3.0299, "step": 160 }, { "epoch": 0.28, "learning_rate": 0.00025199999999999995, "loss": 3.0295, "step": 170 }, { "epoch": 0.29, "learning_rate": 0.000267, "loss": 3.0423, "step": 180 }, { "epoch": 0.31, "learning_rate": 0.00028199999999999997, "loss": 3.0437, "step": 190 }, { "epoch": 0.33, "learning_rate": 0.00029699999999999996, "loss": 3.0569, "step": 200 }, { "epoch": 0.33, "eval_loss": 3.036912679672241, "eval_runtime": 1030.0275, "eval_samples_per_second": 11.271, "eval_steps_per_second": 0.705, "eval_wer": 1.0, "step": 200 }, { "epoch": 0.34, "learning_rate": 0.00029973362930077685, "loss": 3.0243, "step": 210 }, { "epoch": 0.36, "learning_rate": 0.000299400665926748, "loss": 3.0236, "step": 220 }, { "epoch": 0.37, "learning_rate": 0.0002990677025527192, "loss": 3.0316, "step": 230 }, { "epoch": 0.39, "learning_rate": 0.0002987347391786903, "loss": 3.0363, "step": 240 }, { "epoch": 0.41, "learning_rate": 0.00029840177580466146, "loss": 3.0582, "step": 250 }, { "epoch": 0.42, "learning_rate": 0.0002980688124306326, "loss": 3.0144, "step": 260 }, { "epoch": 0.44, "learning_rate": 0.00029773584905660376, "loss": 3.0126, "step": 270 }, { "epoch": 0.46, "learning_rate": 0.0002974028856825749, "loss": 3.0035, "step": 280 }, { "epoch": 0.47, "learning_rate": 0.000297069922308546, "loss": 2.9913, "step": 290 }, { "epoch": 0.49, "learning_rate": 0.00029673695893451716, "loss": 2.9625, "step": 300 }, { "epoch": 0.49, "eval_loss": 2.977771520614624, "eval_runtime": 1042.3792, "eval_samples_per_second": 11.137, "eval_steps_per_second": 0.696, "eval_wer": 1.0, "step": 300 }, { "epoch": 0.5, "learning_rate": 0.0002964039955604883, "loss": 2.9686, "step": 310 }, { "epoch": 0.52, "learning_rate": 0.00029607103218645946, "loss": 2.8773, "step": 320 }, { "epoch": 0.54, "learning_rate": 0.0002957380688124306, "loss": 2.726, "step": 330 }, { "epoch": 0.55, "learning_rate": 0.00029540510543840177, "loss": 2.3671, "step": 340 }, { "epoch": 0.57, "learning_rate": 0.0002950721420643729, "loss": 1.9479, "step": 350 }, { "epoch": 0.59, "learning_rate": 0.000294739178690344, "loss": 1.5549, "step": 360 }, { "epoch": 0.6, "learning_rate": 0.00029440621531631517, "loss": 1.1765, "step": 370 }, { "epoch": 0.62, "learning_rate": 0.0002940732519422863, "loss": 0.9155, "step": 380 }, { "epoch": 0.63, "learning_rate": 0.00029374028856825747, "loss": 0.7823, "step": 390 }, { "epoch": 0.65, "learning_rate": 0.0002934073251942286, "loss": 0.7715, "step": 400 }, { "epoch": 0.65, "eval_loss": 0.5112669467926025, "eval_runtime": 1048.8544, "eval_samples_per_second": 11.068, "eval_steps_per_second": 0.692, "eval_wer": 0.718480591159969, "step": 400 }, { "epoch": 0.67, "learning_rate": 0.00029307436182019977, "loss": 0.6635, "step": 410 }, { "epoch": 0.68, "learning_rate": 0.0002927413984461709, "loss": 0.6032, "step": 420 }, { "epoch": 0.7, "learning_rate": 0.000292408435072142, "loss": 0.5401, "step": 430 }, { "epoch": 0.72, "learning_rate": 0.00029207547169811317, "loss": 0.5276, "step": 440 }, { "epoch": 0.73, "learning_rate": 0.0002917425083240843, "loss": 0.5713, "step": 450 }, { "epoch": 0.75, "learning_rate": 0.0002914095449500555, "loss": 0.46, "step": 460 }, { "epoch": 0.76, "learning_rate": 0.0002910765815760266, "loss": 0.4508, "step": 470 }, { "epoch": 0.78, "learning_rate": 0.0002907436182019977, "loss": 0.447, "step": 480 }, { "epoch": 0.8, "learning_rate": 0.00029041065482796893, "loss": 0.4454, "step": 490 }, { "epoch": 0.81, "learning_rate": 0.0002900776914539401, "loss": 0.4725, "step": 500 }, { "epoch": 0.81, "eval_loss": 0.3072386682033539, "eval_runtime": 1051.5656, "eval_samples_per_second": 11.04, "eval_steps_per_second": 0.69, "eval_wer": 0.5137640833827487, "step": 500 }, { "epoch": 0.83, "learning_rate": 0.0002897447280799112, "loss": 0.3822, "step": 510 }, { "epoch": 0.85, "learning_rate": 0.00028941176470588233, "loss": 0.3807, "step": 520 }, { "epoch": 0.86, "learning_rate": 0.0002890788013318535, "loss": 0.3827, "step": 530 }, { "epoch": 0.88, "learning_rate": 0.00028874583795782463, "loss": 0.3945, "step": 540 }, { "epoch": 0.9, "learning_rate": 0.00028841287458379573, "loss": 0.4519, "step": 550 }, { "epoch": 0.91, "learning_rate": 0.0002880799112097669, "loss": 0.3534, "step": 560 }, { "epoch": 0.93, "learning_rate": 0.00028774694783573803, "loss": 0.3731, "step": 570 }, { "epoch": 0.94, "learning_rate": 0.0002874139844617092, "loss": 0.3433, "step": 580 }, { "epoch": 0.96, "learning_rate": 0.00028708102108768033, "loss": 0.3547, "step": 590 }, { "epoch": 0.98, "learning_rate": 0.0002867480577136515, "loss": 0.4103, "step": 600 }, { "epoch": 0.98, "eval_loss": 0.24468904733657837, "eval_runtime": 1043.2136, "eval_samples_per_second": 11.128, "eval_steps_per_second": 0.696, "eval_wer": 0.4336997673676048, "step": 600 }, { "epoch": 0.99, "learning_rate": 0.00028641509433962264, "loss": 0.3484, "step": 610 }, { "epoch": 1.01, "learning_rate": 0.00028608213096559373, "loss": 0.3775, "step": 620 }, { "epoch": 1.03, "learning_rate": 0.0002857491675915649, "loss": 0.3207, "step": 630 }, { "epoch": 1.04, "learning_rate": 0.00028541620421753604, "loss": 0.2971, "step": 640 }, { "epoch": 1.06, "learning_rate": 0.0002850832408435072, "loss": 0.3129, "step": 650 }, { "epoch": 1.07, "learning_rate": 0.00028475027746947834, "loss": 0.3143, "step": 660 }, { "epoch": 1.09, "learning_rate": 0.0002844173140954495, "loss": 0.3179, "step": 670 }, { "epoch": 1.11, "learning_rate": 0.00028408435072142064, "loss": 0.299, "step": 680 }, { "epoch": 1.12, "learning_rate": 0.0002837513873473918, "loss": 0.2734, "step": 690 }, { "epoch": 1.14, "learning_rate": 0.0002834184239733629, "loss": 0.2775, "step": 700 }, { "epoch": 1.14, "eval_loss": 0.2055242359638214, "eval_runtime": 1053.2634, "eval_samples_per_second": 11.022, "eval_steps_per_second": 0.689, "eval_wer": 0.3768644802262464, "step": 700 }, { "epoch": 1.16, "learning_rate": 0.00028308546059933404, "loss": 0.2878, "step": 710 }, { "epoch": 1.17, "learning_rate": 0.0002827524972253052, "loss": 0.307, "step": 720 }, { "epoch": 1.19, "learning_rate": 0.00028241953385127634, "loss": 0.2848, "step": 730 }, { "epoch": 1.2, "learning_rate": 0.00028208657047724744, "loss": 0.2744, "step": 740 }, { "epoch": 1.22, "learning_rate": 0.00028175360710321865, "loss": 0.2839, "step": 750 }, { "epoch": 1.24, "learning_rate": 0.0002814206437291898, "loss": 0.2888, "step": 760 }, { "epoch": 1.25, "learning_rate": 0.0002810876803551609, "loss": 0.288, "step": 770 }, { "epoch": 1.27, "learning_rate": 0.00028075471698113205, "loss": 0.2559, "step": 780 }, { "epoch": 1.29, "learning_rate": 0.0002804217536071032, "loss": 0.2498, "step": 790 }, { "epoch": 1.3, "learning_rate": 0.00028008879023307435, "loss": 0.2554, "step": 800 }, { "epoch": 1.3, "eval_loss": 0.19498416781425476, "eval_runtime": 1045.8615, "eval_samples_per_second": 11.1, "eval_steps_per_second": 0.694, "eval_wer": 0.360272316744971, "step": 800 }, { "epoch": 1.32, "learning_rate": 0.00027975582685904545, "loss": 0.3059, "step": 810 }, { "epoch": 1.33, "learning_rate": 0.0002794228634850166, "loss": 0.281, "step": 820 }, { "epoch": 1.35, "learning_rate": 0.00027908990011098775, "loss": 0.2642, "step": 830 }, { "epoch": 1.37, "learning_rate": 0.0002787569367369589, "loss": 0.242, "step": 840 }, { "epoch": 1.38, "learning_rate": 0.00027842397336293005, "loss": 0.2541, "step": 850 }, { "epoch": 1.4, "learning_rate": 0.0002780910099889012, "loss": 0.2614, "step": 860 }, { "epoch": 1.42, "learning_rate": 0.00027775804661487236, "loss": 0.3001, "step": 870 }, { "epoch": 1.43, "learning_rate": 0.0002774250832408435, "loss": 0.2365, "step": 880 }, { "epoch": 1.45, "learning_rate": 0.0002770921198668146, "loss": 0.2373, "step": 890 }, { "epoch": 1.46, "learning_rate": 0.00027675915649278575, "loss": 0.263, "step": 900 }, { "epoch": 1.46, "eval_loss": 0.1812964379787445, "eval_runtime": 1053.0871, "eval_samples_per_second": 11.024, "eval_steps_per_second": 0.689, "eval_wer": 0.3371687269078137, "step": 900 }, { "epoch": 1.48, "learning_rate": 0.0002764261931187569, "loss": 0.289, "step": 910 }, { "epoch": 1.5, "learning_rate": 0.00027609322974472806, "loss": 0.2578, "step": 920 }, { "epoch": 1.51, "learning_rate": 0.0002757602663706992, "loss": 0.2565, "step": 930 }, { "epoch": 1.53, "learning_rate": 0.00027542730299667036, "loss": 0.2543, "step": 940 }, { "epoch": 1.55, "learning_rate": 0.0002750943396226415, "loss": 0.2404, "step": 950 }, { "epoch": 1.56, "learning_rate": 0.0002747613762486126, "loss": 0.2291, "step": 960 }, { "epoch": 1.58, "learning_rate": 0.00027442841287458376, "loss": 0.2636, "step": 970 }, { "epoch": 1.59, "learning_rate": 0.0002740954495005549, "loss": 0.2222, "step": 980 }, { "epoch": 1.61, "learning_rate": 0.00027376248612652606, "loss": 0.2322, "step": 990 }, { "epoch": 1.63, "learning_rate": 0.0002734295227524972, "loss": 0.2294, "step": 1000 }, { "epoch": 1.63, "eval_loss": 0.1664419174194336, "eval_runtime": 1047.59, "eval_samples_per_second": 11.082, "eval_steps_per_second": 0.693, "eval_wer": 0.3131756602654746, "step": 1000 }, { "epoch": 1.64, "learning_rate": 0.00027309655937846837, "loss": 0.249, "step": 1010 }, { "epoch": 1.66, "learning_rate": 0.0002727635960044395, "loss": 0.2682, "step": 1020 }, { "epoch": 1.68, "learning_rate": 0.0002724306326304106, "loss": 0.2242, "step": 1030 }, { "epoch": 1.69, "learning_rate": 0.00027209766925638177, "loss": 0.2333, "step": 1040 }, { "epoch": 1.71, "learning_rate": 0.0002717647058823529, "loss": 0.2188, "step": 1050 }, { "epoch": 1.72, "learning_rate": 0.00027143174250832407, "loss": 0.2592, "step": 1060 }, { "epoch": 1.74, "learning_rate": 0.0002710987791342952, "loss": 0.268, "step": 1070 }, { "epoch": 1.76, "learning_rate": 0.0002707658157602663, "loss": 0.2091, "step": 1080 }, { "epoch": 1.77, "learning_rate": 0.00027043285238623747, "loss": 0.2069, "step": 1090 }, { "epoch": 1.79, "learning_rate": 0.0002700998890122087, "loss": 0.2296, "step": 1100 }, { "epoch": 1.79, "eval_loss": 0.15650752186775208, "eval_runtime": 1054.82, "eval_samples_per_second": 11.006, "eval_steps_per_second": 0.688, "eval_wer": 0.2962413903206678, "step": 1100 }, { "epoch": 1.81, "learning_rate": 0.00026976692563817977, "loss": 0.2402, "step": 1110 }, { "epoch": 1.82, "learning_rate": 0.0002694339622641509, "loss": 0.2258, "step": 1120 }, { "epoch": 1.84, "learning_rate": 0.0002691009988901221, "loss": 0.2215, "step": 1130 }, { "epoch": 1.86, "learning_rate": 0.0002687680355160932, "loss": 0.2059, "step": 1140 }, { "epoch": 1.87, "learning_rate": 0.0002684350721420643, "loss": 0.2022, "step": 1150 }, { "epoch": 1.89, "learning_rate": 0.0002681021087680355, "loss": 0.2261, "step": 1160 }, { "epoch": 1.9, "learning_rate": 0.0002677691453940066, "loss": 0.2132, "step": 1170 }, { "epoch": 1.92, "learning_rate": 0.0002674361820199778, "loss": 0.2245, "step": 1180 }, { "epoch": 1.94, "learning_rate": 0.00026710321864594893, "loss": 0.2133, "step": 1190 }, { "epoch": 1.95, "learning_rate": 0.0002667702552719201, "loss": 0.2183, "step": 1200 }, { "epoch": 1.95, "eval_loss": 0.14735129475593567, "eval_runtime": 1059.1579, "eval_samples_per_second": 10.961, "eval_steps_per_second": 0.685, "eval_wer": 0.29856771427268164, "step": 1200 }, { "epoch": 1.97, "learning_rate": 0.00026643729189789123, "loss": 0.2204, "step": 1210 }, { "epoch": 1.99, "learning_rate": 0.00026610432852386233, "loss": 0.2482, "step": 1220 }, { "epoch": 2.0, "learning_rate": 0.0002657713651498335, "loss": 0.2214, "step": 1230 }, { "epoch": 2.02, "learning_rate": 0.00026543840177580463, "loss": 0.1779, "step": 1240 }, { "epoch": 2.03, "learning_rate": 0.0002651054384017758, "loss": 0.1688, "step": 1250 }, { "epoch": 2.05, "learning_rate": 0.00026477247502774693, "loss": 0.1915, "step": 1260 }, { "epoch": 2.07, "learning_rate": 0.0002644395116537181, "loss": 0.1942, "step": 1270 }, { "epoch": 2.08, "learning_rate": 0.00026410654827968924, "loss": 0.2266, "step": 1280 }, { "epoch": 2.1, "learning_rate": 0.0002637735849056604, "loss": 0.1931, "step": 1290 }, { "epoch": 2.12, "learning_rate": 0.0002634406215316315, "loss": 0.1822, "step": 1300 }, { "epoch": 2.12, "eval_loss": 0.1546158790588379, "eval_runtime": 1062.5837, "eval_samples_per_second": 10.925, "eval_steps_per_second": 0.683, "eval_wer": 0.2810974775350089, "step": 1300 }, { "epoch": 2.13, "learning_rate": 0.00026310765815760264, "loss": 0.1735, "step": 1310 }, { "epoch": 2.15, "learning_rate": 0.0002627746947835738, "loss": 0.2042, "step": 1320 }, { "epoch": 2.16, "learning_rate": 0.00026244173140954494, "loss": 0.2347, "step": 1330 }, { "epoch": 2.18, "learning_rate": 0.00026210876803551604, "loss": 0.1816, "step": 1340 }, { "epoch": 2.2, "learning_rate": 0.0002617758046614872, "loss": 0.1838, "step": 1350 }, { "epoch": 2.21, "learning_rate": 0.0002614428412874584, "loss": 0.1856, "step": 1360 }, { "epoch": 2.23, "learning_rate": 0.0002611098779134295, "loss": 0.1866, "step": 1370 }, { "epoch": 2.25, "learning_rate": 0.00026077691453940064, "loss": 0.2016, "step": 1380 }, { "epoch": 2.26, "learning_rate": 0.0002604439511653718, "loss": 0.1813, "step": 1390 }, { "epoch": 2.28, "learning_rate": 0.00026011098779134294, "loss": 0.1798, "step": 1400 }, { "epoch": 2.28, "eval_loss": 0.14420565962791443, "eval_runtime": 1057.3911, "eval_samples_per_second": 10.979, "eval_steps_per_second": 0.687, "eval_wer": 0.28108607398622454, "step": 1400 }, { "epoch": 2.29, "learning_rate": 0.00025977802441731404, "loss": 0.184, "step": 1410 }, { "epoch": 2.31, "learning_rate": 0.0002594450610432852, "loss": 0.19, "step": 1420 }, { "epoch": 2.33, "learning_rate": 0.00025911209766925634, "loss": 0.1982, "step": 1430 }, { "epoch": 2.34, "learning_rate": 0.0002587791342952275, "loss": 0.1627, "step": 1440 }, { "epoch": 2.36, "learning_rate": 0.00025844617092119865, "loss": 0.1914, "step": 1450 }, { "epoch": 2.38, "learning_rate": 0.0002581132075471698, "loss": 0.171, "step": 1460 }, { "epoch": 2.39, "learning_rate": 0.00025778024417314095, "loss": 0.1653, "step": 1470 }, { "epoch": 2.41, "learning_rate": 0.0002574472807991121, "loss": 0.2263, "step": 1480 }, { "epoch": 2.42, "learning_rate": 0.0002571143174250832, "loss": 0.1801, "step": 1490 }, { "epoch": 2.44, "learning_rate": 0.00025678135405105435, "loss": 0.179, "step": 1500 }, { "epoch": 2.44, "eval_loss": 0.14106744527816772, "eval_runtime": 1059.0138, "eval_samples_per_second": 10.962, "eval_steps_per_second": 0.686, "eval_wer": 0.26864480226246407, "step": 1500 }, { "epoch": 2.46, "learning_rate": 0.0002564483906770255, "loss": 0.1718, "step": 1510 }, { "epoch": 2.47, "learning_rate": 0.00025611542730299665, "loss": 0.1659, "step": 1520 }, { "epoch": 2.49, "learning_rate": 0.0002557824639289678, "loss": 0.2146, "step": 1530 }, { "epoch": 2.51, "learning_rate": 0.00025544950055493896, "loss": 0.1789, "step": 1540 }, { "epoch": 2.52, "learning_rate": 0.0002551165371809101, "loss": 0.18, "step": 1550 }, { "epoch": 2.54, "learning_rate": 0.0002547835738068812, "loss": 0.176, "step": 1560 }, { "epoch": 2.55, "learning_rate": 0.00025445061043285236, "loss": 0.1789, "step": 1570 }, { "epoch": 2.57, "learning_rate": 0.0002541176470588235, "loss": 0.2318, "step": 1580 }, { "epoch": 2.59, "learning_rate": 0.00025378468368479466, "loss": 0.161, "step": 1590 }, { "epoch": 2.6, "learning_rate": 0.00025345172031076576, "loss": 0.1593, "step": 1600 }, { "epoch": 2.6, "eval_loss": 0.14078468084335327, "eval_runtime": 1052.8874, "eval_samples_per_second": 11.026, "eval_steps_per_second": 0.69, "eval_wer": 0.2739132418008484, "step": 1600 }, { "epoch": 2.62, "learning_rate": 0.0002531187569367369, "loss": 5.5922, "step": 1610 }, { "epoch": 2.64, "learning_rate": 0.0002527857935627081, "loss": 3.2059, "step": 1620 }, { "epoch": 2.65, "learning_rate": 0.0002524528301886792, "loss": 3.0426, "step": 1630 }, { "epoch": 2.67, "learning_rate": 0.00025211986681465036, "loss": 2.9869, "step": 1640 }, { "epoch": 2.69, "learning_rate": 0.0002517869034406215, "loss": 2.936, "step": 1650 }, { "epoch": 2.7, "learning_rate": 0.00025145394006659266, "loss": 2.5037, "step": 1660 }, { "epoch": 2.72, "learning_rate": 0.0002511209766925638, "loss": 1.297, "step": 1670 }, { "epoch": 2.74, "learning_rate": 0.0002507880133185349, "loss": 0.56, "step": 1680 }, { "epoch": 2.75, "learning_rate": 0.00025045504994450606, "loss": 0.3262, "step": 1690 }, { "epoch": 2.77, "learning_rate": 0.0002501220865704772, "loss": 0.2652, "step": 1700 }, { "epoch": 2.77, "eval_loss": 0.20742562413215637, "eval_runtime": 1056.4173, "eval_samples_per_second": 10.989, "eval_steps_per_second": 0.687, "eval_wer": 0.4499384208365643, "step": 1700 }, { "epoch": 2.78, "learning_rate": 0.00024978912319644837, "loss": 0.2585, "step": 1710 }, { "epoch": 2.8, "learning_rate": 0.0002494561598224195, "loss": 0.2722, "step": 1720 }, { "epoch": 2.82, "learning_rate": 0.00024912319644839067, "loss": 0.2571, "step": 1730 }, { "epoch": 2.83, "learning_rate": 0.0002487902330743618, "loss": 0.2256, "step": 1740 }, { "epoch": 2.85, "learning_rate": 0.0002484572697003329, "loss": 0.2176, "step": 1750 }, { "epoch": 2.87, "learning_rate": 0.00024812430632630407, "loss": 0.1954, "step": 1760 }, { "epoch": 2.88, "learning_rate": 0.0002477913429522752, "loss": 0.2099, "step": 1770 }, { "epoch": 2.9, "learning_rate": 0.00024745837957824637, "loss": 0.2316, "step": 1780 }, { "epoch": 2.91, "learning_rate": 0.0002471254162042175, "loss": 0.1814, "step": 1790 }, { "epoch": 2.93, "learning_rate": 0.0002467924528301887, "loss": 0.1834, "step": 1800 }, { "epoch": 2.93, "eval_loss": 0.15701305866241455, "eval_runtime": 1047.7925, "eval_samples_per_second": 11.079, "eval_steps_per_second": 0.693, "eval_wer": 0.39416366373215345, "step": 1800 }, { "epoch": 2.95, "learning_rate": 0.0002464594894561598, "loss": 0.2047, "step": 1810 }, { "epoch": 2.96, "learning_rate": 0.0002461265260821309, "loss": 0.2051, "step": 1820 }, { "epoch": 2.98, "learning_rate": 0.0002457935627081021, "loss": 0.2246, "step": 1830 }, { "epoch": 3.0, "learning_rate": 0.0002454605993340732, "loss": 0.1797, "step": 1840 }, { "epoch": 3.01, "learning_rate": 0.0002451276359600444, "loss": 0.2112, "step": 1850 }, { "epoch": 3.03, "learning_rate": 0.00024479467258601553, "loss": 0.1914, "step": 1860 }, { "epoch": 3.04, "learning_rate": 0.0002444617092119866, "loss": 0.1761, "step": 1870 }, { "epoch": 3.06, "learning_rate": 0.0002441287458379578, "loss": 0.1917, "step": 1880 }, { "epoch": 3.08, "learning_rate": 0.00024379578246392896, "loss": 0.2106, "step": 1890 }, { "epoch": 3.09, "learning_rate": 0.0002434628190899001, "loss": 0.2015, "step": 1900 }, { "epoch": 3.09, "eval_loss": 0.15158045291900635, "eval_runtime": 1057.6306, "eval_samples_per_second": 10.976, "eval_steps_per_second": 0.686, "eval_wer": 0.38587328376590796, "step": 1900 }, { "epoch": 3.11, "learning_rate": 0.00024312985571587123, "loss": 0.1756, "step": 1910 }, { "epoch": 3.13, "learning_rate": 0.00024279689234184238, "loss": 0.1733, "step": 1920 }, { "epoch": 3.14, "learning_rate": 0.0002424639289678135, "loss": 0.1805, "step": 1930 }, { "epoch": 3.16, "learning_rate": 0.00024213096559378466, "loss": 0.181, "step": 1940 }, { "epoch": 3.17, "learning_rate": 0.00024179800221975578, "loss": 0.1701, "step": 1950 }, { "epoch": 3.19, "learning_rate": 0.00024146503884572693, "loss": 0.1894, "step": 1960 }, { "epoch": 3.21, "learning_rate": 0.0002411320754716981, "loss": 0.1749, "step": 1970 }, { "epoch": 3.22, "learning_rate": 0.00024079911209766924, "loss": 0.1559, "step": 1980 }, { "epoch": 3.24, "learning_rate": 0.0002404661487236404, "loss": 0.2001, "step": 1990 }, { "epoch": 3.26, "learning_rate": 0.0002401331853496115, "loss": 0.1696, "step": 2000 }, { "epoch": 3.26, "eval_loss": 0.14519302546977997, "eval_runtime": 1046.4919, "eval_samples_per_second": 11.093, "eval_steps_per_second": 0.694, "eval_wer": 0.3826004652647904, "step": 2000 }, { "epoch": 3.27, "learning_rate": 0.00023980022197558266, "loss": 0.1714, "step": 2010 }, { "epoch": 3.29, "learning_rate": 0.0002394672586015538, "loss": 0.1527, "step": 2020 }, { "epoch": 3.31, "learning_rate": 0.00023913429522752494, "loss": 0.1589, "step": 2030 }, { "epoch": 3.32, "learning_rate": 0.0002388013318534961, "loss": 0.1786, "step": 2040 }, { "epoch": 3.34, "learning_rate": 0.00023846836847946724, "loss": 0.1958, "step": 2050 }, { "epoch": 3.35, "learning_rate": 0.0002381354051054384, "loss": 0.1671, "step": 2060 }, { "epoch": 3.37, "learning_rate": 0.00023780244173140955, "loss": 0.177, "step": 2070 }, { "epoch": 3.39, "learning_rate": 0.00023746947835738067, "loss": 0.1618, "step": 2080 }, { "epoch": 3.4, "learning_rate": 0.00023713651498335182, "loss": 0.1859, "step": 2090 }, { "epoch": 3.42, "learning_rate": 0.00023680355160932295, "loss": 0.1782, "step": 2100 }, { "epoch": 3.42, "eval_loss": 0.14127187430858612, "eval_runtime": 1047.734, "eval_samples_per_second": 11.08, "eval_steps_per_second": 0.693, "eval_wer": 0.3763285134333805, "step": 2100 }, { "epoch": 3.44, "learning_rate": 0.0002364705882352941, "loss": 0.1563, "step": 2110 }, { "epoch": 3.45, "learning_rate": 0.00023613762486126522, "loss": 0.159, "step": 2120 }, { "epoch": 3.47, "learning_rate": 0.00023580466148723637, "loss": 0.1781, "step": 2130 }, { "epoch": 3.48, "learning_rate": 0.00023547169811320755, "loss": 0.1841, "step": 2140 }, { "epoch": 3.5, "learning_rate": 0.00023513873473917868, "loss": 0.1755, "step": 2150 }, { "epoch": 3.52, "learning_rate": 0.00023480577136514983, "loss": 0.153, "step": 2160 }, { "epoch": 3.53, "learning_rate": 0.00023447280799112095, "loss": 0.151, "step": 2170 }, { "epoch": 3.55, "learning_rate": 0.0002341398446170921, "loss": 0.1691, "step": 2180 }, { "epoch": 3.57, "learning_rate": 0.00023380688124306323, "loss": 0.1685, "step": 2190 }, { "epoch": 3.58, "learning_rate": 0.00023347391786903438, "loss": 0.1636, "step": 2200 }, { "epoch": 3.58, "eval_loss": 0.13500617444515228, "eval_runtime": 1049.8895, "eval_samples_per_second": 11.057, "eval_steps_per_second": 0.692, "eval_wer": 0.3761004424576928, "step": 2200 }, { "epoch": 3.6, "learning_rate": 0.00023314095449500553, "loss": 0.1803, "step": 2210 }, { "epoch": 3.61, "learning_rate": 0.00023280799112097665, "loss": 0.15, "step": 2220 }, { "epoch": 3.63, "learning_rate": 0.00023247502774694783, "loss": 0.1721, "step": 2230 }, { "epoch": 3.65, "learning_rate": 0.00023214206437291896, "loss": 0.1691, "step": 2240 }, { "epoch": 3.66, "learning_rate": 0.0002318091009988901, "loss": 0.1548, "step": 2250 }, { "epoch": 3.68, "learning_rate": 0.00023147613762486126, "loss": 0.1616, "step": 2260 }, { "epoch": 3.7, "learning_rate": 0.00023114317425083238, "loss": 0.159, "step": 2270 }, { "epoch": 3.71, "learning_rate": 0.00023081021087680353, "loss": 0.1483, "step": 2280 }, { "epoch": 3.73, "learning_rate": 0.00023047724750277466, "loss": 0.1693, "step": 2290 }, { "epoch": 3.74, "learning_rate": 0.0002301442841287458, "loss": 0.173, "step": 2300 }, { "epoch": 3.74, "eval_loss": 0.13234160840511322, "eval_runtime": 1044.2826, "eval_samples_per_second": 11.117, "eval_steps_per_second": 0.695, "eval_wer": 0.3621653058431784, "step": 2300 }, { "epoch": 3.76, "learning_rate": 0.000229811320754717, "loss": 0.1488, "step": 2310 }, { "epoch": 3.78, "learning_rate": 0.0002294783573806881, "loss": 0.1491, "step": 2320 }, { "epoch": 3.79, "learning_rate": 0.00022914539400665926, "loss": 0.1632, "step": 2330 }, { "epoch": 3.81, "learning_rate": 0.0002288124306326304, "loss": 0.1733, "step": 2340 }, { "epoch": 3.83, "learning_rate": 0.00022847946725860154, "loss": 0.1648, "step": 2350 }, { "epoch": 3.84, "learning_rate": 0.00022814650388457266, "loss": 0.1513, "step": 2360 }, { "epoch": 3.86, "learning_rate": 0.00022781354051054382, "loss": 0.1602, "step": 2370 }, { "epoch": 3.87, "learning_rate": 0.00022748057713651494, "loss": 0.1528, "step": 2380 }, { "epoch": 3.89, "learning_rate": 0.0002271476137624861, "loss": 0.1688, "step": 2390 }, { "epoch": 3.91, "learning_rate": 0.00022681465038845727, "loss": 0.1704, "step": 2400 }, { "epoch": 3.91, "eval_loss": 0.12890243530273438, "eval_runtime": 1049.8266, "eval_samples_per_second": 11.058, "eval_steps_per_second": 0.692, "eval_wer": 0.364423208502486, "step": 2400 }, { "epoch": 3.92, "learning_rate": 0.0002264816870144284, "loss": 0.1427, "step": 2410 }, { "epoch": 3.94, "learning_rate": 0.00022614872364039955, "loss": 0.1585, "step": 2420 }, { "epoch": 3.96, "learning_rate": 0.00022581576026637067, "loss": 0.1504, "step": 2430 }, { "epoch": 3.97, "learning_rate": 0.00022548279689234182, "loss": 0.1819, "step": 2440 }, { "epoch": 3.99, "learning_rate": 0.00022514983351831297, "loss": 0.1679, "step": 2450 }, { "epoch": 4.0, "learning_rate": 0.0002248168701442841, "loss": 0.1619, "step": 2460 }, { "epoch": 4.02, "learning_rate": 0.00022448390677025525, "loss": 0.1288, "step": 2470 }, { "epoch": 4.04, "learning_rate": 0.00022415094339622637, "loss": 0.1288, "step": 2480 }, { "epoch": 4.05, "learning_rate": 0.00022381798002219755, "loss": 0.1465, "step": 2490 }, { "epoch": 4.07, "learning_rate": 0.0002234850166481687, "loss": 0.1418, "step": 2500 }, { "epoch": 4.07, "eval_loss": 0.12660406529903412, "eval_runtime": 1055.0446, "eval_samples_per_second": 11.003, "eval_steps_per_second": 0.688, "eval_wer": 0.34805911599689826, "step": 2500 }, { "epoch": 4.09, "learning_rate": 0.00022315205327413983, "loss": 0.1618, "step": 2510 }, { "epoch": 4.1, "learning_rate": 0.00022281908990011098, "loss": 0.1337, "step": 2520 }, { "epoch": 4.12, "learning_rate": 0.0002224861265260821, "loss": 0.1545, "step": 2530 }, { "epoch": 4.14, "learning_rate": 0.00022215316315205325, "loss": 0.1343, "step": 2540 }, { "epoch": 4.15, "learning_rate": 0.00022182019977802438, "loss": 0.1309, "step": 2550 }, { "epoch": 4.17, "learning_rate": 0.00022148723640399553, "loss": 0.1449, "step": 2560 }, { "epoch": 4.18, "learning_rate": 0.0002211542730299667, "loss": 0.1214, "step": 2570 }, { "epoch": 4.2, "learning_rate": 0.00022082130965593783, "loss": 0.1397, "step": 2580 }, { "epoch": 4.22, "learning_rate": 0.00022048834628190898, "loss": 0.124, "step": 2590 }, { "epoch": 4.23, "learning_rate": 0.0002201553829078801, "loss": 0.1403, "step": 2600 }, { "epoch": 4.23, "eval_loss": 0.1273965835571289, "eval_runtime": 1056.7903, "eval_samples_per_second": 10.985, "eval_steps_per_second": 0.687, "eval_wer": 0.34823016922866395, "step": 2600 }, { "epoch": 4.25, "learning_rate": 0.00021982241953385126, "loss": 0.1616, "step": 2610 }, { "epoch": 4.27, "learning_rate": 0.00021948945615982238, "loss": 0.1368, "step": 2620 }, { "epoch": 4.28, "learning_rate": 0.00021915649278579354, "loss": 0.1288, "step": 2630 }, { "epoch": 4.3, "learning_rate": 0.0002188235294117647, "loss": 0.1299, "step": 2640 }, { "epoch": 4.31, "learning_rate": 0.0002184905660377358, "loss": 0.1349, "step": 2650 }, { "epoch": 4.33, "learning_rate": 0.000218157602663707, "loss": 0.166, "step": 2660 }, { "epoch": 4.35, "learning_rate": 0.00021782463928967811, "loss": 0.1347, "step": 2670 }, { "epoch": 4.36, "learning_rate": 0.00021749167591564927, "loss": 0.1315, "step": 2680 }, { "epoch": 4.38, "learning_rate": 0.00021715871254162042, "loss": 0.1311, "step": 2690 }, { "epoch": 4.4, "learning_rate": 0.00021682574916759154, "loss": 0.1353, "step": 2700 }, { "epoch": 4.4, "eval_loss": 0.1287122368812561, "eval_runtime": 1067.8021, "eval_samples_per_second": 10.872, "eval_steps_per_second": 0.68, "eval_wer": 0.3488915750581581, "step": 2700 }, { "epoch": 4.41, "learning_rate": 0.0002164927857935627, "loss": 0.1741, "step": 2710 }, { "epoch": 4.43, "learning_rate": 0.00021615982241953382, "loss": 0.1331, "step": 2720 }, { "epoch": 4.44, "learning_rate": 0.00021582685904550497, "loss": 0.1422, "step": 2730 }, { "epoch": 4.46, "learning_rate": 0.0002154938956714761, "loss": 0.1313, "step": 2740 }, { "epoch": 4.48, "learning_rate": 0.00021516093229744727, "loss": 0.1508, "step": 2750 }, { "epoch": 4.49, "learning_rate": 0.00021482796892341842, "loss": 0.1604, "step": 2760 }, { "epoch": 4.51, "learning_rate": 0.00021449500554938955, "loss": 0.1301, "step": 2770 }, { "epoch": 4.53, "learning_rate": 0.0002141620421753607, "loss": 0.1432, "step": 2780 }, { "epoch": 4.54, "learning_rate": 0.00021382907880133182, "loss": 0.1274, "step": 2790 }, { "epoch": 4.56, "learning_rate": 0.00021349611542730297, "loss": 0.1432, "step": 2800 }, { "epoch": 4.56, "eval_loss": 0.12929297983646393, "eval_runtime": 1060.4858, "eval_samples_per_second": 10.947, "eval_steps_per_second": 0.685, "eval_wer": 0.35316790585230123, "step": 2800 }, { "epoch": 4.57, "learning_rate": 0.0002131631520532741, "loss": 0.1652, "step": 2810 }, { "epoch": 4.59, "learning_rate": 0.00021283018867924525, "loss": 0.1304, "step": 2820 }, { "epoch": 4.61, "learning_rate": 0.00021249722530521643, "loss": 0.1461, "step": 2830 }, { "epoch": 4.62, "learning_rate": 0.00021216426193118755, "loss": 0.1342, "step": 2840 }, { "epoch": 4.64, "learning_rate": 0.0002118312985571587, "loss": 0.1506, "step": 2850 }, { "epoch": 4.66, "learning_rate": 0.00021149833518312985, "loss": 0.1487, "step": 2860 }, { "epoch": 4.67, "learning_rate": 0.00021116537180910098, "loss": 0.1291, "step": 2870 }, { "epoch": 4.69, "learning_rate": 0.00021083240843507213, "loss": 0.1326, "step": 2880 }, { "epoch": 4.7, "learning_rate": 0.00021049944506104325, "loss": 0.1352, "step": 2890 }, { "epoch": 4.72, "learning_rate": 0.0002101664816870144, "loss": 0.1283, "step": 2900 }, { "epoch": 4.72, "eval_loss": 0.12256743013858795, "eval_runtime": 1069.9155, "eval_samples_per_second": 10.85, "eval_steps_per_second": 0.679, "eval_wer": 0.34155909318980066, "step": 2900 }, { "epoch": 4.74, "learning_rate": 0.00020983351831298553, "loss": 0.1592, "step": 2910 }, { "epoch": 4.75, "learning_rate": 0.0002095005549389567, "loss": 0.137, "step": 2920 }, { "epoch": 4.77, "learning_rate": 0.00020916759156492786, "loss": 0.1447, "step": 2930 }, { "epoch": 4.79, "learning_rate": 0.00020883462819089898, "loss": 0.13, "step": 2940 }, { "epoch": 4.8, "learning_rate": 0.00020850166481687014, "loss": 0.1336, "step": 2950 }, { "epoch": 4.82, "learning_rate": 0.00020816870144284126, "loss": 0.1451, "step": 2960 }, { "epoch": 4.83, "learning_rate": 0.0002078357380688124, "loss": 0.1191, "step": 2970 }, { "epoch": 4.85, "learning_rate": 0.00020750277469478354, "loss": 0.1266, "step": 2980 }, { "epoch": 4.87, "learning_rate": 0.0002071698113207547, "loss": 0.1349, "step": 2990 }, { "epoch": 4.88, "learning_rate": 0.00020683684794672584, "loss": 0.1367, "step": 3000 }, { "epoch": 4.88, "eval_loss": 0.12064128369092941, "eval_runtime": 1063.0232, "eval_samples_per_second": 10.921, "eval_steps_per_second": 0.683, "eval_wer": 0.3425740090316106, "step": 3000 }, { "epoch": 4.9, "learning_rate": 0.000206503884572697, "loss": 0.1426, "step": 3010 }, { "epoch": 4.92, "learning_rate": 0.00020617092119866814, "loss": 0.1221, "step": 3020 }, { "epoch": 4.93, "learning_rate": 0.00020583795782463927, "loss": 0.1268, "step": 3030 }, { "epoch": 4.95, "learning_rate": 0.00020550499445061042, "loss": 0.1421, "step": 3040 }, { "epoch": 4.97, "learning_rate": 0.00020517203107658157, "loss": 0.142, "step": 3050 }, { "epoch": 4.98, "learning_rate": 0.0002048390677025527, "loss": 0.1659, "step": 3060 }, { "epoch": 5.0, "learning_rate": 0.00020450610432852384, "loss": 0.1353, "step": 3070 }, { "epoch": 5.01, "learning_rate": 0.00020417314095449497, "loss": 0.1325, "step": 3080 }, { "epoch": 5.03, "learning_rate": 0.00020384017758046615, "loss": 0.1191, "step": 3090 }, { "epoch": 5.05, "learning_rate": 0.0002035072142064373, "loss": 0.1162, "step": 3100 }, { "epoch": 5.05, "eval_loss": 0.12216832488775253, "eval_runtime": 1074.6528, "eval_samples_per_second": 10.803, "eval_steps_per_second": 0.676, "eval_wer": 0.33939241892076816, "step": 3100 }, { "epoch": 5.06, "learning_rate": 0.00020317425083240842, "loss": 0.1213, "step": 3110 }, { "epoch": 5.08, "learning_rate": 0.00020284128745837957, "loss": 0.1631, "step": 3120 }, { "epoch": 5.1, "learning_rate": 0.0002025083240843507, "loss": 0.1285, "step": 3130 }, { "epoch": 5.11, "learning_rate": 0.00020217536071032185, "loss": 0.1122, "step": 3140 }, { "epoch": 5.13, "learning_rate": 0.00020184239733629297, "loss": 0.1202, "step": 3150 }, { "epoch": 5.14, "learning_rate": 0.00020150943396226413, "loss": 0.1222, "step": 3160 }, { "epoch": 5.16, "learning_rate": 0.00020117647058823525, "loss": 0.1442, "step": 3170 }, { "epoch": 5.18, "learning_rate": 0.00020084350721420643, "loss": 0.1197, "step": 3180 }, { "epoch": 5.19, "learning_rate": 0.00020051054384017758, "loss": 0.1124, "step": 3190 }, { "epoch": 5.21, "learning_rate": 0.0002001775804661487, "loss": 0.1267, "step": 3200 }, { "epoch": 5.21, "eval_loss": 0.11833450943231583, "eval_runtime": 1070.4879, "eval_samples_per_second": 10.845, "eval_steps_per_second": 0.678, "eval_wer": 0.33128449573507274, "step": 3200 }, { "epoch": 5.23, "learning_rate": 0.00019984461709211985, "loss": 0.1266, "step": 3210 }, { "epoch": 5.24, "learning_rate": 0.00019951165371809098, "loss": 0.1213, "step": 3220 }, { "epoch": 5.26, "learning_rate": 0.00019917869034406213, "loss": 0.1246, "step": 3230 }, { "epoch": 5.27, "learning_rate": 0.00019884572697003328, "loss": 0.1117, "step": 3240 }, { "epoch": 5.29, "learning_rate": 0.0001985127635960044, "loss": 0.1127, "step": 3250 }, { "epoch": 5.31, "learning_rate": 0.00019817980022197556, "loss": 0.1134, "step": 3260 }, { "epoch": 5.32, "learning_rate": 0.0001978468368479467, "loss": 0.1505, "step": 3270 }, { "epoch": 5.34, "learning_rate": 0.00019751387347391786, "loss": 0.1252, "step": 3280 }, { "epoch": 5.36, "learning_rate": 0.000197180910099889, "loss": 0.1227, "step": 3290 }, { "epoch": 5.37, "learning_rate": 0.00019684794672586014, "loss": 0.1126, "step": 3300 }, { "epoch": 5.37, "eval_loss": 0.11798416078090668, "eval_runtime": 1068.0787, "eval_samples_per_second": 10.869, "eval_steps_per_second": 0.68, "eval_wer": 0.32992747342973133, "step": 3300 }, { "epoch": 5.39, "learning_rate": 0.0001965149833518313, "loss": 0.1238, "step": 3310 }, { "epoch": 5.4, "learning_rate": 0.0001961820199778024, "loss": 0.1361, "step": 3320 }, { "epoch": 5.42, "learning_rate": 0.00019584905660377356, "loss": 0.1354, "step": 3330 }, { "epoch": 5.44, "learning_rate": 0.0001955160932297447, "loss": 0.1225, "step": 3340 }, { "epoch": 5.45, "learning_rate": 0.00019518312985571587, "loss": 0.1127, "step": 3350 }, { "epoch": 5.47, "learning_rate": 0.00019485016648168702, "loss": 0.1162, "step": 3360 }, { "epoch": 5.49, "learning_rate": 0.00019451720310765814, "loss": 0.142, "step": 3370 }, { "epoch": 5.5, "learning_rate": 0.0001941842397336293, "loss": 0.1239, "step": 3380 }, { "epoch": 5.52, "learning_rate": 0.00019385127635960042, "loss": 0.1173, "step": 3390 }, { "epoch": 5.53, "learning_rate": 0.00019351831298557157, "loss": 0.1127, "step": 3400 }, { "epoch": 5.53, "eval_loss": 0.11773715913295746, "eval_runtime": 1066.7509, "eval_samples_per_second": 10.883, "eval_steps_per_second": 0.681, "eval_wer": 0.330486247320166, "step": 3400 }, { "epoch": 5.55, "learning_rate": 0.0001931853496115427, "loss": 0.1243, "step": 3410 }, { "epoch": 5.57, "learning_rate": 0.00019285238623751384, "loss": 0.1427, "step": 3420 }, { "epoch": 5.58, "learning_rate": 0.000192519422863485, "loss": 0.1182, "step": 3430 }, { "epoch": 5.6, "learning_rate": 0.00019218645948945615, "loss": 0.1189, "step": 3440 }, { "epoch": 5.62, "learning_rate": 0.0001918534961154273, "loss": 0.1062, "step": 3450 }, { "epoch": 5.63, "learning_rate": 0.00019152053274139842, "loss": 0.1192, "step": 3460 }, { "epoch": 5.65, "learning_rate": 0.00019118756936736957, "loss": 0.1308, "step": 3470 }, { "epoch": 5.66, "learning_rate": 0.00019085460599334073, "loss": 0.1137, "step": 3480 }, { "epoch": 5.68, "learning_rate": 0.00019052164261931185, "loss": 0.1087, "step": 3490 }, { "epoch": 5.7, "learning_rate": 0.000190188679245283, "loss": 0.1155, "step": 3500 }, { "epoch": 5.7, "eval_loss": 0.11851849406957626, "eval_runtime": 1077.6014, "eval_samples_per_second": 10.773, "eval_steps_per_second": 0.674, "eval_wer": 0.33171783058887927, "step": 3500 }, { "epoch": 5.71, "learning_rate": 0.00018985571587125413, "loss": 0.1336, "step": 3510 }, { "epoch": 5.73, "learning_rate": 0.00018952275249722528, "loss": 0.1248, "step": 3520 }, { "epoch": 5.75, "learning_rate": 0.00018918978912319646, "loss": 0.1247, "step": 3530 }, { "epoch": 5.76, "learning_rate": 0.00018885682574916758, "loss": 0.1057, "step": 3540 }, { "epoch": 5.78, "learning_rate": 0.00018852386237513873, "loss": 0.1213, "step": 3550 }, { "epoch": 5.79, "learning_rate": 0.00018819089900110986, "loss": 0.1122, "step": 3560 }, { "epoch": 5.81, "learning_rate": 0.000187857935627081, "loss": 0.1251, "step": 3570 }, { "epoch": 5.83, "learning_rate": 0.00018752497225305213, "loss": 0.1164, "step": 3580 }, { "epoch": 5.84, "learning_rate": 0.00018719200887902328, "loss": 0.1161, "step": 3590 }, { "epoch": 5.86, "learning_rate": 0.00018685904550499443, "loss": 0.1086, "step": 3600 }, { "epoch": 5.86, "eval_loss": 0.11293730139732361, "eval_runtime": 1066.1285, "eval_samples_per_second": 10.889, "eval_steps_per_second": 0.681, "eval_wer": 0.322686219951649, "step": 3600 }, { "epoch": 5.88, "learning_rate": 0.00018652608213096559, "loss": 0.1214, "step": 3610 }, { "epoch": 5.89, "learning_rate": 0.00018619311875693674, "loss": 0.1279, "step": 3620 }, { "epoch": 5.91, "learning_rate": 0.00018586015538290786, "loss": 0.1201, "step": 3630 }, { "epoch": 5.93, "learning_rate": 0.000185527192008879, "loss": 0.1016, "step": 3640 }, { "epoch": 5.94, "learning_rate": 0.00018519422863485016, "loss": 0.1116, "step": 3650 }, { "epoch": 5.96, "learning_rate": 0.0001848612652608213, "loss": 0.1134, "step": 3660 }, { "epoch": 5.97, "learning_rate": 0.00018452830188679244, "loss": 0.1374, "step": 3670 }, { "epoch": 5.99, "learning_rate": 0.00018419533851276356, "loss": 0.1045, "step": 3680 }, { "epoch": 6.01, "learning_rate": 0.00018386237513873471, "loss": 0.1122, "step": 3690 }, { "epoch": 6.02, "learning_rate": 0.0001835294117647059, "loss": 0.1135, "step": 3700 }, { "epoch": 6.02, "eval_loss": 0.11180546879768372, "eval_runtime": 1077.7461, "eval_samples_per_second": 10.772, "eval_steps_per_second": 0.674, "eval_wer": 0.32663184783104504, "step": 3700 }, { "epoch": 6.04, "learning_rate": 0.00018319644839067702, "loss": 0.0972, "step": 3710 }, { "epoch": 6.06, "learning_rate": 0.00018286348501664817, "loss": 0.1149, "step": 3720 }, { "epoch": 6.07, "learning_rate": 0.0001825305216426193, "loss": 0.1192, "step": 3730 }, { "epoch": 6.09, "learning_rate": 0.00018219755826859044, "loss": 0.1311, "step": 3740 }, { "epoch": 6.1, "learning_rate": 0.00018186459489456157, "loss": 0.1149, "step": 3750 }, { "epoch": 6.12, "learning_rate": 0.00018153163152053272, "loss": 0.0989, "step": 3760 }, { "epoch": 6.14, "learning_rate": 0.00018119866814650384, "loss": 0.1192, "step": 3770 }, { "epoch": 6.15, "learning_rate": 0.000180865704772475, "loss": 0.1011, "step": 3780 }, { "epoch": 6.17, "learning_rate": 0.00018053274139844617, "loss": 0.1315, "step": 3790 }, { "epoch": 6.19, "learning_rate": 0.0001801997780244173, "loss": 0.1112, "step": 3800 }, { "epoch": 6.19, "eval_loss": 0.1141621470451355, "eval_runtime": 1071.2114, "eval_samples_per_second": 10.837, "eval_steps_per_second": 0.678, "eval_wer": 0.32282306253706156, "step": 3800 }, { "epoch": 6.2, "learning_rate": 0.00017986681465038845, "loss": 0.12, "step": 3810 }, { "epoch": 6.22, "learning_rate": 0.00017953385127635957, "loss": 0.1038, "step": 3820 }, { "epoch": 6.23, "learning_rate": 0.00017920088790233073, "loss": 0.1138, "step": 3830 }, { "epoch": 6.25, "learning_rate": 0.00017886792452830188, "loss": 0.1376, "step": 3840 }, { "epoch": 6.27, "learning_rate": 0.000178534961154273, "loss": 0.1072, "step": 3850 }, { "epoch": 6.28, "learning_rate": 0.00017820199778024415, "loss": 0.1063, "step": 3860 }, { "epoch": 6.3, "learning_rate": 0.0001778690344062153, "loss": 0.1071, "step": 3870 }, { "epoch": 6.32, "learning_rate": 0.00017753607103218646, "loss": 0.1016, "step": 3880 }, { "epoch": 6.33, "learning_rate": 0.0001772031076581576, "loss": 0.1246, "step": 3890 }, { "epoch": 6.35, "learning_rate": 0.00017687014428412873, "loss": 0.0866, "step": 3900 }, { "epoch": 6.35, "eval_loss": 0.11722169816493988, "eval_runtime": 1077.3899, "eval_samples_per_second": 10.775, "eval_steps_per_second": 0.674, "eval_wer": 0.32837659079505543, "step": 3900 }, { "epoch": 6.36, "learning_rate": 0.00017653718091009988, "loss": 0.1062, "step": 3910 }, { "epoch": 6.38, "learning_rate": 0.000176204217536071, "loss": 0.1214, "step": 3920 }, { "epoch": 6.4, "learning_rate": 0.00017587125416204216, "loss": 0.1138, "step": 3930 }, { "epoch": 6.41, "learning_rate": 0.00017553829078801328, "loss": 0.1216, "step": 3940 }, { "epoch": 6.43, "learning_rate": 0.00017520532741398443, "loss": 0.1111, "step": 3950 }, { "epoch": 6.45, "learning_rate": 0.0001748723640399556, "loss": 0.1044, "step": 3960 }, { "epoch": 6.46, "learning_rate": 0.00017453940066592674, "loss": 0.1108, "step": 3970 }, { "epoch": 6.48, "learning_rate": 0.0001742064372918979, "loss": 0.1083, "step": 3980 }, { "epoch": 6.49, "learning_rate": 0.000173873473917869, "loss": 0.136, "step": 3990 }, { "epoch": 6.51, "learning_rate": 0.00017354051054384016, "loss": 0.1003, "step": 4000 }, { "epoch": 6.51, "eval_loss": 0.11332289129495621, "eval_runtime": 1068.9642, "eval_samples_per_second": 10.86, "eval_steps_per_second": 0.679, "eval_wer": 0.32444236646444374, "step": 4000 }, { "epoch": 6.53, "learning_rate": 0.0001732075471698113, "loss": 6.155, "step": 4010 }, { "epoch": 6.55, "learning_rate": 0.00017287458379578244, "loss": 3.1349, "step": 4020 }, { "epoch": 6.56, "learning_rate": 0.0001725416204217536, "loss": 3.0542, "step": 4030 }, { "epoch": 6.58, "learning_rate": 0.00017220865704772474, "loss": 3.0084, "step": 4040 }, { "epoch": 6.6, "learning_rate": 0.0001718756936736959, "loss": 2.9937, "step": 4050 }, { "epoch": 6.61, "learning_rate": 0.00017154273029966702, "loss": 2.9765, "step": 4060 }, { "epoch": 6.63, "learning_rate": 0.00017120976692563817, "loss": 2.9408, "step": 4070 }, { "epoch": 6.64, "learning_rate": 0.00017087680355160932, "loss": 2.6233, "step": 4080 }, { "epoch": 6.66, "learning_rate": 0.00017054384017758045, "loss": 1.4399, "step": 4090 }, { "epoch": 6.68, "learning_rate": 0.0001702108768035516, "loss": 0.4366, "step": 4100 }, { "epoch": 6.68, "eval_loss": 0.24362534284591675, "eval_runtime": 1064.9477, "eval_samples_per_second": 10.901, "eval_steps_per_second": 0.682, "eval_wer": 0.4587077498517539, "step": 4100 }, { "epoch": 6.69, "learning_rate": 0.00016987791342952272, "loss": 0.2313, "step": 4110 }, { "epoch": 6.71, "learning_rate": 0.00016954495005549387, "loss": 0.1763, "step": 4120 }, { "epoch": 6.73, "learning_rate": 0.00016921198668146505, "loss": 0.1956, "step": 4130 }, { "epoch": 6.74, "learning_rate": 0.00016887902330743617, "loss": 0.1825, "step": 4140 }, { "epoch": 6.76, "learning_rate": 0.00016854605993340733, "loss": 0.1533, "step": 4150 }, { "epoch": 6.77, "learning_rate": 0.00016821309655937845, "loss": 0.1459, "step": 4160 }, { "epoch": 6.79, "learning_rate": 0.0001678801331853496, "loss": 0.1547, "step": 4170 }, { "epoch": 6.81, "learning_rate": 0.00016754716981132073, "loss": 0.1347, "step": 4180 }, { "epoch": 6.82, "learning_rate": 0.00016721420643729188, "loss": 0.1387, "step": 4190 }, { "epoch": 6.84, "learning_rate": 0.000166881243063263, "loss": 0.1216, "step": 4200 }, { "epoch": 6.84, "eval_loss": 0.1344495564699173, "eval_runtime": 1042.0044, "eval_samples_per_second": 11.141, "eval_steps_per_second": 0.697, "eval_wer": 0.33859417050586144, "step": 4200 }, { "epoch": 6.86, "learning_rate": 0.00016654827968923415, "loss": 0.1096, "step": 4210 }, { "epoch": 6.87, "learning_rate": 0.00016621531631520533, "loss": 0.1429, "step": 4220 }, { "epoch": 6.89, "learning_rate": 0.00016588235294117646, "loss": 0.139, "step": 4230 }, { "epoch": 6.9, "learning_rate": 0.0001655493895671476, "loss": 0.1343, "step": 4240 }, { "epoch": 6.92, "learning_rate": 0.00016521642619311873, "loss": 0.1223, "step": 4250 }, { "epoch": 6.94, "learning_rate": 0.00016488346281908988, "loss": 0.1298, "step": 4260 }, { "epoch": 6.95, "learning_rate": 0.00016455049944506103, "loss": 0.1549, "step": 4270 }, { "epoch": 6.97, "learning_rate": 0.00016421753607103216, "loss": 0.1378, "step": 4280 }, { "epoch": 6.99, "learning_rate": 0.0001638845726970033, "loss": 0.1492, "step": 4290 }, { "epoch": 7.0, "learning_rate": 0.00016355160932297446, "loss": 0.1165, "step": 4300 }, { "epoch": 7.0, "eval_loss": 0.12804068624973297, "eval_runtime": 1059.2138, "eval_samples_per_second": 10.96, "eval_steps_per_second": 0.685, "eval_wer": 0.3323564293208046, "step": 4300 }, { "epoch": 7.02, "learning_rate": 0.0001632186459489456, "loss": 0.1136, "step": 4310 }, { "epoch": 7.03, "learning_rate": 0.00016288568257491676, "loss": 0.1041, "step": 4320 }, { "epoch": 7.05, "learning_rate": 0.0001625527192008879, "loss": 0.1216, "step": 4330 }, { "epoch": 7.07, "learning_rate": 0.00016221975582685904, "loss": 0.1156, "step": 4340 }, { "epoch": 7.08, "learning_rate": 0.00016188679245283016, "loss": 0.1433, "step": 4350 }, { "epoch": 7.1, "learning_rate": 0.00016155382907880132, "loss": 0.0961, "step": 4360 }, { "epoch": 7.12, "learning_rate": 0.00016122086570477244, "loss": 0.1292, "step": 4370 }, { "epoch": 7.13, "learning_rate": 0.0001608879023307436, "loss": 0.1118, "step": 4380 }, { "epoch": 7.15, "learning_rate": 0.00016055493895671477, "loss": 0.1076, "step": 4390 }, { "epoch": 7.17, "learning_rate": 0.0001602219755826859, "loss": 0.131, "step": 4400 }, { "epoch": 7.17, "eval_loss": 0.12520039081573486, "eval_runtime": 1045.0008, "eval_samples_per_second": 11.109, "eval_steps_per_second": 0.695, "eval_wer": 0.3244879806595813, "step": 4400 }, { "epoch": 7.18, "learning_rate": 0.00015988901220865705, "loss": 0.1108, "step": 4410 }, { "epoch": 7.2, "learning_rate": 0.00015955604883462817, "loss": 0.1011, "step": 4420 }, { "epoch": 7.21, "learning_rate": 0.00015922308546059932, "loss": 0.1123, "step": 4430 }, { "epoch": 7.23, "learning_rate": 0.00015889012208657047, "loss": 0.119, "step": 4440 }, { "epoch": 7.25, "learning_rate": 0.0001585571587125416, "loss": 0.1318, "step": 4450 }, { "epoch": 7.26, "learning_rate": 0.00015822419533851275, "loss": 0.1212, "step": 4460 }, { "epoch": 7.28, "learning_rate": 0.00015789123196448387, "loss": 0.1036, "step": 4470 }, { "epoch": 7.3, "learning_rate": 0.00015755826859045505, "loss": 0.1113, "step": 4480 }, { "epoch": 7.31, "learning_rate": 0.0001572253052164262, "loss": 0.1028, "step": 4490 }, { "epoch": 7.33, "learning_rate": 0.00015689234184239733, "loss": 0.1407, "step": 4500 }, { "epoch": 7.33, "eval_loss": 0.12339464575052261, "eval_runtime": 1067.1403, "eval_samples_per_second": 10.879, "eval_steps_per_second": 0.68, "eval_wer": 0.32521780778178166, "step": 4500 }, { "epoch": 7.34, "learning_rate": 0.00015655937846836848, "loss": 0.1006, "step": 4510 }, { "epoch": 7.36, "learning_rate": 0.0001562264150943396, "loss": 0.099, "step": 4520 }, { "epoch": 7.38, "learning_rate": 0.00015589345172031075, "loss": 0.1091, "step": 4530 }, { "epoch": 7.39, "learning_rate": 0.00015556048834628188, "loss": 0.1087, "step": 4540 }, { "epoch": 7.41, "learning_rate": 0.00015522752497225303, "loss": 0.1429, "step": 4550 }, { "epoch": 7.43, "learning_rate": 0.0001548945615982242, "loss": 0.1038, "step": 4560 }, { "epoch": 7.44, "learning_rate": 0.00015456159822419533, "loss": 0.1086, "step": 4570 }, { "epoch": 7.46, "learning_rate": 0.00015422863485016648, "loss": 0.1144, "step": 4580 }, { "epoch": 7.47, "learning_rate": 0.0001538956714761376, "loss": 0.1135, "step": 4590 }, { "epoch": 7.49, "learning_rate": 0.00015356270810210876, "loss": 0.1394, "step": 4600 }, { "epoch": 7.49, "eval_loss": 0.12083031237125397, "eval_runtime": 1054.1862, "eval_samples_per_second": 11.012, "eval_steps_per_second": 0.689, "eval_wer": 0.3176800620353054, "step": 4600 }, { "epoch": 7.51, "learning_rate": 0.00015322974472807988, "loss": 0.0979, "step": 4610 }, { "epoch": 7.52, "learning_rate": 0.00015289678135405103, "loss": 0.1036, "step": 4620 }, { "epoch": 7.54, "learning_rate": 0.00015256381798002219, "loss": 0.1145, "step": 4630 }, { "epoch": 7.56, "learning_rate": 0.0001522308546059933, "loss": 0.1173, "step": 4640 }, { "epoch": 7.57, "learning_rate": 0.0001518978912319645, "loss": 0.1222, "step": 4650 }, { "epoch": 7.59, "learning_rate": 0.0001515649278579356, "loss": 0.0967, "step": 4660 }, { "epoch": 7.6, "learning_rate": 0.00015123196448390676, "loss": 0.1034, "step": 4670 }, { "epoch": 7.62, "learning_rate": 0.00015089900110987792, "loss": 0.1085, "step": 4680 }, { "epoch": 7.64, "learning_rate": 0.00015056603773584904, "loss": 0.1081, "step": 4690 }, { "epoch": 7.65, "learning_rate": 0.0001502330743618202, "loss": 0.1449, "step": 4700 }, { "epoch": 7.65, "eval_loss": 0.11798720061779022, "eval_runtime": 1055.6152, "eval_samples_per_second": 10.997, "eval_steps_per_second": 0.688, "eval_wer": 0.31648268941294533, "step": 4700 }, { "epoch": 7.67, "learning_rate": 0.00014990011098779132, "loss": 0.0958, "step": 4710 }, { "epoch": 7.69, "learning_rate": 0.00014956714761376247, "loss": 0.0984, "step": 4720 }, { "epoch": 7.7, "learning_rate": 0.00014923418423973362, "loss": 0.11, "step": 4730 }, { "epoch": 7.72, "learning_rate": 0.00014890122086570477, "loss": 0.1064, "step": 4740 }, { "epoch": 7.73, "learning_rate": 0.0001485682574916759, "loss": 0.1332, "step": 4750 }, { "epoch": 7.75, "learning_rate": 0.00014823529411764705, "loss": 0.1068, "step": 4760 }, { "epoch": 7.77, "learning_rate": 0.0001479023307436182, "loss": 0.097, "step": 4770 }, { "epoch": 7.78, "learning_rate": 0.00014756936736958932, "loss": 0.1058, "step": 4780 }, { "epoch": 7.8, "learning_rate": 0.00014723640399556047, "loss": 0.1074, "step": 4790 }, { "epoch": 7.82, "learning_rate": 0.00014690344062153162, "loss": 0.1295, "step": 4800 }, { "epoch": 7.82, "eval_loss": 0.11703231930732727, "eval_runtime": 1061.0638, "eval_samples_per_second": 10.941, "eval_steps_per_second": 0.684, "eval_wer": 0.3151712813027414, "step": 4800 }, { "epoch": 7.83, "learning_rate": 0.00014657047724750278, "loss": 0.104, "step": 4810 }, { "epoch": 7.85, "learning_rate": 0.0001462375138734739, "loss": 0.1086, "step": 4820 }, { "epoch": 7.86, "learning_rate": 0.00014590455049944505, "loss": 0.106, "step": 4830 }, { "epoch": 7.88, "learning_rate": 0.00014557158712541618, "loss": 0.1115, "step": 4840 }, { "epoch": 7.9, "learning_rate": 0.00014523862375138733, "loss": 0.1274, "step": 4850 }, { "epoch": 7.91, "learning_rate": 0.00014490566037735848, "loss": 0.0944, "step": 4860 }, { "epoch": 7.93, "learning_rate": 0.00014457269700332963, "loss": 0.1054, "step": 4870 }, { "epoch": 7.95, "learning_rate": 0.00014423973362930075, "loss": 0.1, "step": 4880 }, { "epoch": 7.96, "learning_rate": 0.0001439067702552719, "loss": 0.098, "step": 4890 }, { "epoch": 7.98, "learning_rate": 0.00014357380688124306, "loss": 0.1228, "step": 4900 }, { "epoch": 7.98, "eval_loss": 0.11823362857103348, "eval_runtime": 1053.6032, "eval_samples_per_second": 11.018, "eval_steps_per_second": 0.689, "eval_wer": 0.31600374036400125, "step": 4900 }, { "epoch": 8.0, "learning_rate": 0.00014324084350721418, "loss": 0.0909, "step": 4910 }, { "epoch": 8.01, "learning_rate": 0.00014290788013318533, "loss": 0.1087, "step": 4920 }, { "epoch": 8.03, "learning_rate": 0.00014257491675915648, "loss": 0.0955, "step": 4930 }, { "epoch": 8.04, "learning_rate": 0.00014224195338512764, "loss": 0.0875, "step": 4940 }, { "epoch": 8.06, "learning_rate": 0.00014190899001109876, "loss": 0.0944, "step": 4950 }, { "epoch": 8.08, "learning_rate": 0.0001415760266370699, "loss": 0.1136, "step": 4960 }, { "epoch": 8.09, "learning_rate": 0.00014124306326304104, "loss": 0.0962, "step": 4970 }, { "epoch": 8.11, "learning_rate": 0.00014091009988901221, "loss": 0.0968, "step": 4980 }, { "epoch": 8.13, "learning_rate": 0.00014057713651498334, "loss": 0.0902, "step": 4990 }, { "epoch": 8.14, "learning_rate": 0.0001402441731409545, "loss": 0.0913, "step": 5000 }, { "epoch": 8.14, "eval_loss": 0.11221442371606827, "eval_runtime": 1054.3263, "eval_samples_per_second": 11.011, "eval_steps_per_second": 0.689, "eval_wer": 0.30864845139807506, "step": 5000 }, { "epoch": 8.16, "learning_rate": 0.00013991120976692561, "loss": 0.0965, "step": 5010 }, { "epoch": 8.17, "learning_rate": 0.00013957824639289677, "loss": 0.1048, "step": 5020 }, { "epoch": 8.19, "learning_rate": 0.00013924528301886792, "loss": 0.0965, "step": 5030 }, { "epoch": 8.21, "learning_rate": 0.00013891231964483904, "loss": 0.0891, "step": 5040 }, { "epoch": 8.22, "learning_rate": 0.0001385793562708102, "loss": 0.0898, "step": 5050 }, { "epoch": 8.24, "learning_rate": 0.00013824639289678134, "loss": 0.1044, "step": 5060 }, { "epoch": 8.26, "learning_rate": 0.0001379134295227525, "loss": 0.1056, "step": 5070 }, { "epoch": 8.27, "learning_rate": 0.00013758046614872362, "loss": 0.0937, "step": 5080 }, { "epoch": 8.29, "learning_rate": 0.00013724750277469477, "loss": 0.0898, "step": 5090 }, { "epoch": 8.3, "learning_rate": 0.0001369145394006659, "loss": 0.1014, "step": 5100 }, { "epoch": 8.3, "eval_loss": 0.11184623092412949, "eval_runtime": 1060.664, "eval_samples_per_second": 10.945, "eval_steps_per_second": 0.684, "eval_wer": 0.3099826666058477, "step": 5100 }, { "epoch": 8.32, "learning_rate": 0.00013658157602663707, "loss": 0.0992, "step": 5110 }, { "epoch": 8.34, "learning_rate": 0.0001362486126526082, "loss": 0.0992, "step": 5120 }, { "epoch": 8.35, "learning_rate": 0.00013591564927857935, "loss": 0.0922, "step": 5130 }, { "epoch": 8.37, "learning_rate": 0.00013558268590455047, "loss": 0.0861, "step": 5140 }, { "epoch": 8.39, "learning_rate": 0.00013524972253052162, "loss": 0.0847, "step": 5150 }, { "epoch": 8.4, "learning_rate": 0.00013491675915649278, "loss": 0.1016, "step": 5160 }, { "epoch": 8.42, "learning_rate": 0.00013458379578246393, "loss": 0.0985, "step": 5170 }, { "epoch": 8.43, "learning_rate": 0.00013425083240843505, "loss": 0.0808, "step": 5180 }, { "epoch": 8.45, "learning_rate": 0.0001339178690344062, "loss": 0.1095, "step": 5190 }, { "epoch": 8.47, "learning_rate": 0.00013358490566037735, "loss": 0.0861, "step": 5200 }, { "epoch": 8.47, "eval_loss": 0.11255118250846863, "eval_runtime": 1046.9073, "eval_samples_per_second": 11.089, "eval_steps_per_second": 0.693, "eval_wer": 0.3073940610317931, "step": 5200 }, { "epoch": 8.48, "learning_rate": 0.00013325194228634848, "loss": 6.5883, "step": 5210 }, { "epoch": 8.5, "learning_rate": 0.00013291897891231963, "loss": 3.2032, "step": 5220 }, { "epoch": 8.52, "learning_rate": 0.00013258601553829078, "loss": 3.043, "step": 5230 }, { "epoch": 8.53, "learning_rate": 0.00013225305216426193, "loss": 2.9766, "step": 5240 }, { "epoch": 8.55, "learning_rate": 0.00013192008879023306, "loss": 2.652, "step": 5250 }, { "epoch": 8.57, "learning_rate": 0.0001315871254162042, "loss": 1.4113, "step": 5260 }, { "epoch": 8.58, "learning_rate": 0.00013125416204217533, "loss": 0.4183, "step": 5270 }, { "epoch": 8.6, "learning_rate": 0.0001309211986681465, "loss": 0.2112, "step": 5280 }, { "epoch": 8.62, "learning_rate": 0.00013058823529411764, "loss": 0.1702, "step": 5290 }, { "epoch": 8.63, "learning_rate": 0.0001302552719200888, "loss": 0.1442, "step": 5300 }, { "epoch": 8.63, "eval_loss": 0.13734713196754456, "eval_runtime": 1026.0546, "eval_samples_per_second": 11.314, "eval_steps_per_second": 0.708, "eval_wer": 0.3311362496008758, "step": 5300 }, { "epoch": 8.65, "learning_rate": 0.0001299223085460599, "loss": 0.1728, "step": 5310 }, { "epoch": 8.66, "learning_rate": 0.00012958934517203106, "loss": 0.1354, "step": 5320 }, { "epoch": 8.68, "learning_rate": 0.00012925638179800221, "loss": 0.1177, "step": 5330 }, { "epoch": 8.7, "learning_rate": 0.00012892341842397334, "loss": 0.1181, "step": 5340 }, { "epoch": 8.71, "learning_rate": 0.0001285904550499445, "loss": 0.1264, "step": 5350 }, { "epoch": 8.73, "learning_rate": 0.00012825749167591564, "loss": 0.1512, "step": 5360 }, { "epoch": 8.75, "learning_rate": 0.0001279245283018868, "loss": 0.13, "step": 5370 }, { "epoch": 8.76, "learning_rate": 0.00012759156492785792, "loss": 0.1147, "step": 5380 }, { "epoch": 8.78, "learning_rate": 0.00012725860155382907, "loss": 0.1161, "step": 5390 }, { "epoch": 8.79, "learning_rate": 0.0001269256381798002, "loss": 0.1054, "step": 5400 }, { "epoch": 8.79, "eval_loss": 0.12251464277505875, "eval_runtime": 1019.1954, "eval_samples_per_second": 11.39, "eval_steps_per_second": 0.712, "eval_wer": 0.31429320804634403, "step": 5400 }, { "epoch": 8.81, "learning_rate": 0.00012659267480577137, "loss": 0.1209, "step": 5410 }, { "epoch": 8.83, "learning_rate": 0.0001262597114317425, "loss": 0.1074, "step": 5420 }, { "epoch": 8.84, "learning_rate": 0.00012592674805771365, "loss": 0.1034, "step": 5430 }, { "epoch": 8.86, "learning_rate": 0.00012559378468368477, "loss": 0.1071, "step": 5440 }, { "epoch": 8.88, "learning_rate": 0.00012526082130965592, "loss": 0.1046, "step": 5450 }, { "epoch": 8.89, "learning_rate": 0.00012492785793562707, "loss": 0.1367, "step": 5460 }, { "epoch": 8.91, "learning_rate": 0.00012459489456159823, "loss": 0.1182, "step": 5470 }, { "epoch": 8.92, "learning_rate": 0.00012426193118756935, "loss": 0.1057, "step": 5480 }, { "epoch": 8.94, "learning_rate": 0.0001239289678135405, "loss": 0.1024, "step": 5490 }, { "epoch": 8.96, "learning_rate": 0.00012359600443951165, "loss": 0.104, "step": 5500 }, { "epoch": 8.96, "eval_loss": 0.11897258460521698, "eval_runtime": 1026.6362, "eval_samples_per_second": 11.308, "eval_steps_per_second": 0.707, "eval_wer": 0.315695844546823, "step": 5500 }, { "epoch": 8.97, "learning_rate": 0.00012326304106548278, "loss": 0.1141, "step": 5510 }, { "epoch": 8.99, "learning_rate": 0.00012293007769145393, "loss": 0.1075, "step": 5520 }, { "epoch": 9.01, "learning_rate": 0.00012259711431742508, "loss": 0.1168, "step": 5530 }, { "epoch": 9.02, "learning_rate": 0.00012226415094339623, "loss": 0.0969, "step": 5540 }, { "epoch": 9.04, "learning_rate": 0.00012193118756936735, "loss": 0.0875, "step": 5550 }, { "epoch": 9.05, "learning_rate": 0.0001215982241953385, "loss": 0.1025, "step": 5560 }, { "epoch": 9.07, "learning_rate": 0.00012126526082130964, "loss": 0.1077, "step": 5570 }, { "epoch": 9.09, "learning_rate": 0.0001209322974472808, "loss": 0.121, "step": 5580 }, { "epoch": 9.1, "learning_rate": 0.00012059933407325193, "loss": 0.0857, "step": 5590 }, { "epoch": 9.12, "learning_rate": 0.00012026637069922307, "loss": 0.0972, "step": 5600 }, { "epoch": 9.12, "eval_loss": 0.11396780610084534, "eval_runtime": 1020.1909, "eval_samples_per_second": 11.379, "eval_steps_per_second": 0.712, "eval_wer": 0.3075651142635588, "step": 5600 }, { "epoch": 9.14, "learning_rate": 0.00011993340732519421, "loss": 0.092, "step": 5610 }, { "epoch": 9.15, "learning_rate": 0.00011960044395116535, "loss": 0.0977, "step": 5620 }, { "epoch": 9.17, "learning_rate": 0.00011926748057713651, "loss": 0.105, "step": 5630 }, { "epoch": 9.18, "learning_rate": 0.00011893451720310765, "loss": 0.0866, "step": 5640 }, { "epoch": 9.2, "learning_rate": 0.00011860155382907879, "loss": 0.102, "step": 5650 }, { "epoch": 9.22, "learning_rate": 0.00011826859045504993, "loss": 0.0974, "step": 5660 }, { "epoch": 9.23, "learning_rate": 0.00011793562708102109, "loss": 0.0957, "step": 5670 }, { "epoch": 9.25, "learning_rate": 0.00011760266370699223, "loss": 0.1096, "step": 5680 }, { "epoch": 9.27, "learning_rate": 0.00011726970033296337, "loss": 0.092, "step": 5690 }, { "epoch": 9.28, "learning_rate": 0.0001169367369589345, "loss": 0.0948, "step": 5700 }, { "epoch": 9.28, "eval_loss": 0.1089872196316719, "eval_runtime": 1021.071, "eval_samples_per_second": 11.369, "eval_steps_per_second": 0.711, "eval_wer": 0.306732655202299, "step": 5700 }, { "epoch": 9.3, "learning_rate": 0.00011660377358490566, "loss": 0.1061, "step": 5710 }, { "epoch": 9.31, "learning_rate": 0.00011627081021087679, "loss": 0.1185, "step": 5720 }, { "epoch": 9.33, "learning_rate": 0.00011593784683684793, "loss": 0.1225, "step": 5730 }, { "epoch": 9.35, "learning_rate": 0.00011560488346281908, "loss": 0.0911, "step": 5740 }, { "epoch": 9.36, "learning_rate": 0.00011527192008879022, "loss": 0.0929, "step": 5750 }, { "epoch": 9.38, "learning_rate": 0.00011493895671476137, "loss": 0.0932, "step": 5760 }, { "epoch": 9.4, "learning_rate": 0.00011460599334073251, "loss": 0.0949, "step": 5770 }, { "epoch": 9.41, "learning_rate": 0.00011427302996670365, "loss": 0.1152, "step": 5780 }, { "epoch": 9.43, "learning_rate": 0.00011394006659267478, "loss": 0.0827, "step": 5790 }, { "epoch": 9.45, "learning_rate": 0.00011360710321864595, "loss": 0.1067, "step": 5800 }, { "epoch": 9.45, "eval_loss": 0.11170622706413269, "eval_runtime": 1017.821, "eval_samples_per_second": 11.406, "eval_steps_per_second": 0.713, "eval_wer": 0.3074396752269306, "step": 5800 }, { "epoch": 9.46, "learning_rate": 0.00011327413984461709, "loss": 0.083, "step": 5810 }, { "epoch": 9.48, "learning_rate": 0.00011294117647058823, "loss": 0.0941, "step": 5820 }, { "epoch": 9.49, "learning_rate": 0.00011260821309655936, "loss": 0.1089, "step": 5830 }, { "epoch": 9.51, "learning_rate": 0.00011227524972253051, "loss": 0.0827, "step": 5840 }, { "epoch": 9.53, "learning_rate": 0.00011194228634850165, "loss": 0.0809, "step": 5850 }, { "epoch": 9.54, "learning_rate": 0.0001116093229744728, "loss": 0.0882, "step": 5860 }, { "epoch": 9.56, "learning_rate": 0.00011127635960044394, "loss": 0.1088, "step": 5870 }, { "epoch": 9.58, "learning_rate": 0.00011094339622641508, "loss": 0.1089, "step": 5880 }, { "epoch": 9.59, "learning_rate": 0.00011061043285238623, "loss": 0.089, "step": 5890 }, { "epoch": 9.61, "learning_rate": 0.00011027746947835737, "loss": 0.0798, "step": 5900 }, { "epoch": 9.61, "eval_loss": 0.10973190516233444, "eval_runtime": 1019.7896, "eval_samples_per_second": 11.384, "eval_steps_per_second": 0.712, "eval_wer": 0.3039501892989098, "step": 5900 }, { "epoch": 9.62, "learning_rate": 0.0001099445061043285, "loss": 0.0906, "step": 5910 }, { "epoch": 9.64, "learning_rate": 0.00010961154273029964, "loss": 0.1002, "step": 5920 }, { "epoch": 9.66, "learning_rate": 0.00010927857935627081, "loss": 0.1105, "step": 5930 }, { "epoch": 9.67, "learning_rate": 0.00010894561598224195, "loss": 0.0924, "step": 5940 }, { "epoch": 9.69, "learning_rate": 0.00010861265260821308, "loss": 0.0896, "step": 5950 }, { "epoch": 9.71, "learning_rate": 0.00010827968923418422, "loss": 0.0874, "step": 5960 }, { "epoch": 9.72, "learning_rate": 0.00010794672586015537, "loss": 0.1014, "step": 5970 }, { "epoch": 9.74, "learning_rate": 0.00010761376248612653, "loss": 0.0963, "step": 5980 }, { "epoch": 9.75, "learning_rate": 0.00010728079911209766, "loss": 0.1001, "step": 5990 }, { "epoch": 9.77, "learning_rate": 0.0001069478357380688, "loss": 0.089, "step": 6000 }, { "epoch": 9.77, "eval_loss": 0.10492939502000809, "eval_runtime": 1025.2992, "eval_samples_per_second": 11.323, "eval_steps_per_second": 0.708, "eval_wer": 0.3004721069196734, "step": 6000 }, { "epoch": 9.79, "learning_rate": 0.00010661487236403994, "loss": 0.0929, "step": 6010 }, { "epoch": 9.8, "learning_rate": 0.00010628190899001109, "loss": 0.098, "step": 6020 }, { "epoch": 9.82, "learning_rate": 0.00010594894561598223, "loss": 0.103, "step": 6030 }, { "epoch": 9.84, "learning_rate": 0.00010561598224195338, "loss": 0.0882, "step": 6040 }, { "epoch": 9.85, "learning_rate": 0.00010528301886792452, "loss": 0.0824, "step": 6050 }, { "epoch": 9.87, "learning_rate": 0.00010495005549389567, "loss": 0.0839, "step": 6060 }, { "epoch": 9.88, "learning_rate": 0.0001046170921198668, "loss": 0.0947, "step": 6070 }, { "epoch": 9.9, "learning_rate": 0.00010428412874583794, "loss": 0.1102, "step": 6080 }, { "epoch": 9.92, "learning_rate": 0.00010395116537180908, "loss": 0.0848, "step": 6090 }, { "epoch": 9.93, "learning_rate": 0.00010361820199778025, "loss": 0.0829, "step": 6100 }, { "epoch": 9.93, "eval_loss": 0.10562794655561447, "eval_runtime": 1019.3478, "eval_samples_per_second": 11.389, "eval_steps_per_second": 0.712, "eval_wer": 0.3005975459563016, "step": 6100 }, { "epoch": 9.95, "learning_rate": 0.00010328523862375139, "loss": 0.0847, "step": 6110 }, { "epoch": 9.97, "learning_rate": 0.00010295227524972252, "loss": 0.0887, "step": 6120 }, { "epoch": 9.98, "learning_rate": 0.00010261931187569366, "loss": 0.1017, "step": 6130 }, { "epoch": 10.0, "learning_rate": 0.0001022863485016648, "loss": 0.0982, "step": 6140 }, { "epoch": 10.01, "learning_rate": 0.00010195338512763595, "loss": 0.0813, "step": 6150 }, { "epoch": 10.03, "learning_rate": 0.0001016204217536071, "loss": 0.0799, "step": 6160 }, { "epoch": 10.05, "learning_rate": 0.00010128745837957824, "loss": 0.083, "step": 6170 }, { "epoch": 10.06, "learning_rate": 0.00010095449500554938, "loss": 0.0856, "step": 6180 }, { "epoch": 10.08, "learning_rate": 0.00010062153163152053, "loss": 0.1095, "step": 6190 }, { "epoch": 10.1, "learning_rate": 0.00010028856825749167, "loss": 0.0687, "step": 6200 }, { "epoch": 10.1, "eval_loss": 0.11015890538692474, "eval_runtime": 1019.1234, "eval_samples_per_second": 11.391, "eval_steps_per_second": 0.712, "eval_wer": 0.3017721114810929, "step": 6200 }, { "epoch": 10.11, "learning_rate": 9.99556048834628e-05, "loss": 0.0762, "step": 6210 }, { "epoch": 10.13, "learning_rate": 9.962264150943394e-05, "loss": 0.084, "step": 6220 }, { "epoch": 10.14, "learning_rate": 9.928967813540511e-05, "loss": 0.086, "step": 6230 }, { "epoch": 10.16, "learning_rate": 9.895671476137624e-05, "loss": 0.1075, "step": 6240 }, { "epoch": 10.18, "learning_rate": 9.862375138734738e-05, "loss": 0.0899, "step": 6250 }, { "epoch": 10.19, "learning_rate": 9.829078801331852e-05, "loss": 0.0807, "step": 6260 }, { "epoch": 10.21, "learning_rate": 9.795782463928966e-05, "loss": 0.0864, "step": 6270 }, { "epoch": 10.23, "learning_rate": 9.762486126526082e-05, "loss": 0.0886, "step": 6280 }, { "epoch": 10.24, "learning_rate": 9.729189789123196e-05, "loss": 0.1098, "step": 6290 }, { "epoch": 10.26, "learning_rate": 9.69589345172031e-05, "loss": 0.0844, "step": 6300 }, { "epoch": 10.26, "eval_loss": 0.1055637001991272, "eval_runtime": 1048.4219, "eval_samples_per_second": 11.073, "eval_steps_per_second": 0.692, "eval_wer": 0.2984764858824066, "step": 6300 }, { "epoch": 10.28, "learning_rate": 9.662597114317424e-05, "loss": 0.0835, "step": 6310 }, { "epoch": 10.29, "learning_rate": 9.629300776914539e-05, "loss": 0.0915, "step": 6320 }, { "epoch": 10.31, "learning_rate": 9.596004439511653e-05, "loss": 0.0948, "step": 6330 }, { "epoch": 10.32, "learning_rate": 9.562708102108766e-05, "loss": 0.1001, "step": 6340 }, { "epoch": 10.34, "learning_rate": 9.529411764705882e-05, "loss": 0.0963, "step": 6350 }, { "epoch": 10.36, "learning_rate": 9.496115427302997e-05, "loss": 0.0793, "step": 6360 }, { "epoch": 10.37, "learning_rate": 9.46281908990011e-05, "loss": 0.0777, "step": 6370 }, { "epoch": 10.39, "learning_rate": 9.429522752497224e-05, "loss": 0.0838, "step": 6380 }, { "epoch": 10.41, "learning_rate": 9.396226415094338e-05, "loss": 0.0973, "step": 6390 }, { "epoch": 10.42, "learning_rate": 9.362930077691452e-05, "loss": 0.0862, "step": 6400 }, { "epoch": 10.42, "eval_loss": 0.10725821554660797, "eval_runtime": 1020.0641, "eval_samples_per_second": 11.381, "eval_steps_per_second": 0.712, "eval_wer": 0.29897824202891937, "step": 6400 }, { "epoch": 10.44, "learning_rate": 9.329633740288568e-05, "loss": 0.0768, "step": 6410 }, { "epoch": 10.45, "learning_rate": 9.296337402885682e-05, "loss": 0.0807, "step": 6420 }, { "epoch": 10.47, "learning_rate": 9.263041065482796e-05, "loss": 0.0908, "step": 6430 }, { "epoch": 10.49, "learning_rate": 9.22974472807991e-05, "loss": 0.092, "step": 6440 }, { "epoch": 10.5, "learning_rate": 9.196448390677025e-05, "loss": 0.082, "step": 6450 }, { "epoch": 10.52, "learning_rate": 9.16315205327414e-05, "loss": 0.0787, "step": 6460 }, { "epoch": 10.54, "learning_rate": 9.129855715871254e-05, "loss": 0.077, "step": 6470 }, { "epoch": 10.55, "learning_rate": 9.096559378468367e-05, "loss": 0.0939, "step": 6480 }, { "epoch": 10.57, "learning_rate": 9.063263041065483e-05, "loss": 0.0936, "step": 6490 }, { "epoch": 10.58, "learning_rate": 9.029966703662596e-05, "loss": 0.0936, "step": 6500 }, { "epoch": 10.58, "eval_loss": 0.10486315935850143, "eval_runtime": 1031.8306, "eval_samples_per_second": 11.251, "eval_steps_per_second": 0.704, "eval_wer": 0.29490717511289516, "step": 6500 }, { "epoch": 10.6, "learning_rate": 8.99667036625971e-05, "loss": 0.0771, "step": 6510 }, { "epoch": 10.62, "learning_rate": 8.963374028856824e-05, "loss": 0.086, "step": 6520 }, { "epoch": 10.63, "learning_rate": 8.930077691453939e-05, "loss": 0.0941, "step": 6530 }, { "epoch": 10.65, "learning_rate": 8.896781354051054e-05, "loss": 0.1008, "step": 6540 }, { "epoch": 10.67, "learning_rate": 8.863485016648168e-05, "loss": 0.0782, "step": 6550 }, { "epoch": 10.68, "learning_rate": 8.830188679245282e-05, "loss": 0.0774, "step": 6560 }, { "epoch": 10.7, "learning_rate": 8.796892341842396e-05, "loss": 0.086, "step": 6570 }, { "epoch": 10.71, "learning_rate": 8.763596004439512e-05, "loss": 0.0861, "step": 6580 }, { "epoch": 10.73, "learning_rate": 8.730299667036626e-05, "loss": 0.1058, "step": 6590 }, { "epoch": 10.75, "learning_rate": 8.69700332963374e-05, "loss": 0.0821, "step": 6600 }, { "epoch": 10.75, "eval_loss": 0.10533939301967621, "eval_runtime": 1025.8233, "eval_samples_per_second": 11.317, "eval_steps_per_second": 0.708, "eval_wer": 0.29659490033298364, "step": 6600 }, { "epoch": 10.76, "learning_rate": 8.663706992230853e-05, "loss": 0.0753, "step": 6610 }, { "epoch": 10.78, "learning_rate": 8.630410654827969e-05, "loss": 0.072, "step": 6620 }, { "epoch": 10.8, "learning_rate": 8.597114317425082e-05, "loss": 0.0777, "step": 6630 }, { "epoch": 10.81, "learning_rate": 8.563817980022196e-05, "loss": 0.0837, "step": 6640 }, { "epoch": 10.83, "learning_rate": 8.530521642619311e-05, "loss": 0.0848, "step": 6650 }, { "epoch": 10.84, "learning_rate": 8.497225305216425e-05, "loss": 0.0769, "step": 6660 }, { "epoch": 10.86, "learning_rate": 8.46392896781354e-05, "loss": 0.0796, "step": 6670 }, { "epoch": 10.88, "learning_rate": 8.430632630410654e-05, "loss": 0.0936, "step": 6680 }, { "epoch": 10.89, "learning_rate": 8.397336293007768e-05, "loss": 0.1119, "step": 6690 }, { "epoch": 10.91, "learning_rate": 8.364039955604882e-05, "loss": 0.0797, "step": 6700 }, { "epoch": 10.91, "eval_loss": 0.1042635440826416, "eval_runtime": 1027.5008, "eval_samples_per_second": 11.298, "eval_steps_per_second": 0.707, "eval_wer": 0.2939264699174383, "step": 6700 }, { "epoch": 10.93, "learning_rate": 8.330743618201998e-05, "loss": 0.0779, "step": 6710 }, { "epoch": 10.94, "learning_rate": 8.297447280799112e-05, "loss": 0.0772, "step": 6720 }, { "epoch": 10.96, "learning_rate": 8.264150943396226e-05, "loss": 0.0822, "step": 6730 }, { "epoch": 10.97, "learning_rate": 8.23085460599334e-05, "loss": 0.096, "step": 6740 }, { "epoch": 10.99, "learning_rate": 8.197558268590455e-05, "loss": 0.0825, "step": 6750 }, { "epoch": 11.01, "learning_rate": 8.164261931187568e-05, "loss": 0.0863, "step": 6760 }, { "epoch": 11.02, "learning_rate": 8.130965593784683e-05, "loss": 0.0702, "step": 6770 }, { "epoch": 11.04, "learning_rate": 8.097669256381797e-05, "loss": 0.0787, "step": 6780 }, { "epoch": 11.06, "learning_rate": 8.064372918978911e-05, "loss": 0.0807, "step": 6790 }, { "epoch": 11.07, "learning_rate": 8.031076581576026e-05, "loss": 0.0802, "step": 6800 }, { "epoch": 11.07, "eval_loss": 0.10567396134138107, "eval_runtime": 1030.4189, "eval_samples_per_second": 11.266, "eval_steps_per_second": 0.705, "eval_wer": 0.2911211969164804, "step": 6800 }, { "epoch": 11.09, "learning_rate": 7.99778024417314e-05, "loss": 0.1039, "step": 6810 }, { "epoch": 11.1, "learning_rate": 7.964483906770254e-05, "loss": 0.0698, "step": 6820 }, { "epoch": 11.12, "learning_rate": 7.931187569367369e-05, "loss": 0.0843, "step": 6830 }, { "epoch": 11.14, "learning_rate": 7.897891231964484e-05, "loss": 0.0758, "step": 6840 }, { "epoch": 11.15, "learning_rate": 7.864594894561598e-05, "loss": 0.0751, "step": 6850 }, { "epoch": 11.17, "learning_rate": 7.831298557158712e-05, "loss": 0.1093, "step": 6860 }, { "epoch": 11.19, "learning_rate": 7.798002219755825e-05, "loss": 0.0745, "step": 6870 }, { "epoch": 11.2, "learning_rate": 7.764705882352942e-05, "loss": 0.0738, "step": 6880 }, { "epoch": 11.22, "learning_rate": 7.731409544950056e-05, "loss": 0.0773, "step": 6890 }, { "epoch": 11.24, "learning_rate": 7.69811320754717e-05, "loss": 0.0895, "step": 6900 }, { "epoch": 11.24, "eval_loss": 0.10290031135082245, "eval_runtime": 1030.0756, "eval_samples_per_second": 11.27, "eval_steps_per_second": 0.705, "eval_wer": 0.293379099575788, "step": 6900 }, { "epoch": 11.25, "learning_rate": 7.664816870144283e-05, "loss": 0.0891, "step": 6910 }, { "epoch": 11.27, "learning_rate": 7.631520532741397e-05, "loss": 0.0718, "step": 6920 }, { "epoch": 11.28, "learning_rate": 7.598224195338512e-05, "loss": 0.0702, "step": 6930 }, { "epoch": 11.3, "learning_rate": 7.564927857935626e-05, "loss": 0.0843, "step": 6940 }, { "epoch": 11.32, "learning_rate": 7.531631520532741e-05, "loss": 0.0793, "step": 6950 }, { "epoch": 11.33, "learning_rate": 7.498335183129855e-05, "loss": 0.0874, "step": 6960 }, { "epoch": 11.35, "learning_rate": 7.465038845726969e-05, "loss": 0.0693, "step": 6970 }, { "epoch": 11.37, "learning_rate": 7.431742508324084e-05, "loss": 0.0772, "step": 6980 }, { "epoch": 11.38, "learning_rate": 7.398446170921198e-05, "loss": 0.0803, "step": 6990 }, { "epoch": 11.4, "learning_rate": 7.365149833518313e-05, "loss": 0.073, "step": 7000 }, { "epoch": 11.4, "eval_loss": 0.10418305546045303, "eval_runtime": 1029.0269, "eval_samples_per_second": 11.282, "eval_steps_per_second": 0.706, "eval_wer": 0.28971856041600147, "step": 7000 }, { "epoch": 11.41, "learning_rate": 7.331853496115426e-05, "loss": 0.0925, "step": 7010 }, { "epoch": 11.43, "learning_rate": 7.298557158712542e-05, "loss": 0.0698, "step": 7020 }, { "epoch": 11.45, "learning_rate": 7.265260821309655e-05, "loss": 0.0807, "step": 7030 }, { "epoch": 11.46, "learning_rate": 7.23196448390677e-05, "loss": 0.0722, "step": 7040 }, { "epoch": 11.48, "learning_rate": 7.198668146503884e-05, "loss": 0.0799, "step": 7050 }, { "epoch": 11.5, "learning_rate": 7.165371809100998e-05, "loss": 0.0923, "step": 7060 }, { "epoch": 11.51, "learning_rate": 7.132075471698113e-05, "loss": 0.074, "step": 7070 }, { "epoch": 11.53, "learning_rate": 7.098779134295227e-05, "loss": 0.0999, "step": 7080 }, { "epoch": 11.54, "learning_rate": 7.065482796892341e-05, "loss": 0.0845, "step": 7090 }, { "epoch": 11.56, "learning_rate": 7.032186459489456e-05, "loss": 0.0842, "step": 7100 }, { "epoch": 11.56, "eval_loss": 0.10231784731149673, "eval_runtime": 1027.7817, "eval_samples_per_second": 11.295, "eval_steps_per_second": 0.706, "eval_wer": 0.29024312366008304, "step": 7100 }, { "epoch": 11.58, "learning_rate": 6.99889012208657e-05, "loss": 0.1014, "step": 7110 }, { "epoch": 11.59, "learning_rate": 6.965593784683683e-05, "loss": 0.0694, "step": 7120 }, { "epoch": 11.61, "learning_rate": 6.932297447280799e-05, "loss": 0.0845, "step": 7130 }, { "epoch": 11.63, "learning_rate": 6.899001109877912e-05, "loss": 0.0703, "step": 7140 }, { "epoch": 11.64, "learning_rate": 6.865704772475028e-05, "loss": 0.0744, "step": 7150 }, { "epoch": 11.66, "learning_rate": 6.832408435072141e-05, "loss": 0.0957, "step": 7160 }, { "epoch": 11.67, "learning_rate": 6.799112097669256e-05, "loss": 0.0696, "step": 7170 }, { "epoch": 11.69, "learning_rate": 6.76581576026637e-05, "loss": 0.075, "step": 7180 }, { "epoch": 11.71, "learning_rate": 6.732519422863484e-05, "loss": 0.0769, "step": 7190 }, { "epoch": 11.72, "learning_rate": 6.699223085460599e-05, "loss": 0.0825, "step": 7200 }, { "epoch": 11.72, "eval_loss": 0.10243339836597443, "eval_runtime": 1032.3954, "eval_samples_per_second": 11.245, "eval_steps_per_second": 0.703, "eval_wer": 0.29114400401404916, "step": 7200 }, { "epoch": 11.74, "learning_rate": 6.665926748057713e-05, "loss": 0.0785, "step": 7210 }, { "epoch": 11.76, "learning_rate": 6.632630410654828e-05, "loss": 0.0677, "step": 7220 }, { "epoch": 11.77, "learning_rate": 6.599334073251942e-05, "loss": 0.077, "step": 7230 }, { "epoch": 11.79, "learning_rate": 6.566037735849056e-05, "loss": 0.0703, "step": 7240 }, { "epoch": 11.8, "learning_rate": 6.532741398446171e-05, "loss": 0.0805, "step": 7250 }, { "epoch": 11.82, "learning_rate": 6.499445061043285e-05, "loss": 0.0803, "step": 7260 }, { "epoch": 11.84, "learning_rate": 6.466148723640398e-05, "loss": 0.0746, "step": 7270 }, { "epoch": 11.85, "learning_rate": 6.432852386237514e-05, "loss": 0.0772, "step": 7280 }, { "epoch": 11.87, "learning_rate": 6.399556048834627e-05, "loss": 0.0702, "step": 7290 }, { "epoch": 11.89, "learning_rate": 6.366259711431742e-05, "loss": 0.0958, "step": 7300 }, { "epoch": 11.89, "eval_loss": 0.10183099657297134, "eval_runtime": 1032.979, "eval_samples_per_second": 11.238, "eval_steps_per_second": 0.703, "eval_wer": 0.28876066231811337, "step": 7300 }, { "epoch": 11.9, "learning_rate": 6.332963374028856e-05, "loss": 0.0867, "step": 7310 }, { "epoch": 11.92, "learning_rate": 6.29966703662597e-05, "loss": 0.0707, "step": 7320 }, { "epoch": 11.93, "learning_rate": 6.266370699223085e-05, "loss": 0.0745, "step": 7330 }, { "epoch": 11.95, "learning_rate": 6.233074361820199e-05, "loss": 0.0775, "step": 7340 }, { "epoch": 11.97, "learning_rate": 6.199778024417314e-05, "loss": 0.0833, "step": 7350 }, { "epoch": 11.98, "learning_rate": 6.166481687014428e-05, "loss": 0.0928, "step": 7360 }, { "epoch": 12.0, "learning_rate": 6.133185349611543e-05, "loss": 0.0838, "step": 7370 }, { "epoch": 12.02, "learning_rate": 6.099889012208656e-05, "loss": 0.0649, "step": 7380 }, { "epoch": 12.03, "learning_rate": 6.066592674805771e-05, "loss": 0.0677, "step": 7390 }, { "epoch": 12.05, "learning_rate": 6.033296337402885e-05, "loss": 0.0698, "step": 7400 }, { "epoch": 12.05, "eval_loss": 0.10296176373958588, "eval_runtime": 1030.5952, "eval_samples_per_second": 11.264, "eval_steps_per_second": 0.704, "eval_wer": 0.2883159239155225, "step": 7400 }, { "epoch": 12.07, "learning_rate": 5.9999999999999995e-05, "loss": 0.0685, "step": 7410 }, { "epoch": 12.08, "learning_rate": 5.966703662597114e-05, "loss": 0.1015, "step": 7420 }, { "epoch": 12.1, "learning_rate": 5.9334073251942284e-05, "loss": 0.0678, "step": 7430 }, { "epoch": 12.11, "learning_rate": 5.900110987791342e-05, "loss": 0.0654, "step": 7440 }, { "epoch": 12.13, "learning_rate": 5.8668146503884566e-05, "loss": 0.0662, "step": 7450 }, { "epoch": 12.15, "learning_rate": 5.833518312985571e-05, "loss": 0.068, "step": 7460 }, { "epoch": 12.16, "learning_rate": 5.800221975582685e-05, "loss": 0.0943, "step": 7470 }, { "epoch": 12.18, "learning_rate": 5.7669256381798e-05, "loss": 0.0603, "step": 7480 }, { "epoch": 12.2, "learning_rate": 5.733629300776914e-05, "loss": 0.0646, "step": 7490 }, { "epoch": 12.21, "learning_rate": 5.700332963374028e-05, "loss": 0.0693, "step": 7500 }, { "epoch": 12.21, "eval_loss": 0.1018766313791275, "eval_runtime": 1033.6211, "eval_samples_per_second": 11.231, "eval_steps_per_second": 0.702, "eval_wer": 0.28722118323222184, "step": 7500 }, { "epoch": 12.23, "learning_rate": 5.667036625971143e-05, "loss": 0.0735, "step": 7510 }, { "epoch": 12.24, "learning_rate": 5.633740288568257e-05, "loss": 0.0899, "step": 7520 }, { "epoch": 12.26, "learning_rate": 5.600443951165371e-05, "loss": 0.0655, "step": 7530 }, { "epoch": 12.28, "learning_rate": 5.567147613762486e-05, "loss": 0.065, "step": 7540 }, { "epoch": 12.29, "learning_rate": 5.5338512763596e-05, "loss": 0.0631, "step": 7550 }, { "epoch": 12.31, "learning_rate": 5.5005549389567144e-05, "loss": 0.0735, "step": 7560 }, { "epoch": 12.33, "learning_rate": 5.467258601553829e-05, "loss": 0.0943, "step": 7570 }, { "epoch": 12.34, "learning_rate": 5.4339622641509426e-05, "loss": 0.0765, "step": 7580 }, { "epoch": 12.36, "learning_rate": 5.400665926748057e-05, "loss": 0.0704, "step": 7590 }, { "epoch": 12.37, "learning_rate": 5.3673695893451715e-05, "loss": 0.0736, "step": 7600 }, { "epoch": 12.37, "eval_loss": 0.10030846297740936, "eval_runtime": 1031.8808, "eval_samples_per_second": 11.25, "eval_steps_per_second": 0.704, "eval_wer": 0.28710714774437807, "step": 7600 }, { "epoch": 12.39, "learning_rate": 5.334073251942286e-05, "loss": 0.0654, "step": 7610 }, { "epoch": 12.41, "learning_rate": 5.3007769145394e-05, "loss": 0.1007, "step": 7620 }, { "epoch": 12.42, "learning_rate": 5.267480577136515e-05, "loss": 0.0671, "step": 7630 }, { "epoch": 12.44, "learning_rate": 5.234184239733629e-05, "loss": 0.0656, "step": 7640 }, { "epoch": 12.46, "learning_rate": 5.200887902330743e-05, "loss": 0.0681, "step": 7650 }, { "epoch": 12.47, "learning_rate": 5.1675915649278576e-05, "loss": 0.0827, "step": 7660 }, { "epoch": 12.49, "learning_rate": 5.134295227524972e-05, "loss": 0.0873, "step": 7670 }, { "epoch": 12.5, "learning_rate": 5.100998890122086e-05, "loss": 0.0725, "step": 7680 }, { "epoch": 12.52, "learning_rate": 5.067702552719201e-05, "loss": 0.0748, "step": 7690 }, { "epoch": 12.54, "learning_rate": 5.034406215316315e-05, "loss": 0.0683, "step": 7700 }, { "epoch": 12.54, "eval_loss": 0.10039916634559631, "eval_runtime": 1032.5995, "eval_samples_per_second": 11.242, "eval_steps_per_second": 0.703, "eval_wer": 0.2865255667563746, "step": 7700 }, { "epoch": 12.55, "learning_rate": 5.0011098779134286e-05, "loss": 0.0906, "step": 7710 }, { "epoch": 12.57, "learning_rate": 4.967813540510544e-05, "loss": 0.1002, "step": 7720 }, { "epoch": 12.59, "learning_rate": 4.9345172031076575e-05, "loss": 0.0671, "step": 7730 }, { "epoch": 12.6, "learning_rate": 4.901220865704772e-05, "loss": 0.0581, "step": 7740 }, { "epoch": 12.62, "learning_rate": 4.8679245283018864e-05, "loss": 0.074, "step": 7750 }, { "epoch": 12.63, "learning_rate": 4.834628190899001e-05, "loss": 0.0765, "step": 7760 }, { "epoch": 12.65, "learning_rate": 4.8013318534961146e-05, "loss": 0.0875, "step": 7770 }, { "epoch": 12.67, "learning_rate": 4.76803551609323e-05, "loss": 0.0695, "step": 7780 }, { "epoch": 12.68, "learning_rate": 4.7347391786903436e-05, "loss": 0.0726, "step": 7790 }, { "epoch": 12.7, "learning_rate": 4.701442841287458e-05, "loss": 0.0649, "step": 7800 }, { "epoch": 12.7, "eval_loss": 0.10054486244916916, "eval_runtime": 1029.5202, "eval_samples_per_second": 11.276, "eval_steps_per_second": 0.705, "eval_wer": 0.2835264334260822, "step": 7800 }, { "epoch": 12.72, "learning_rate": 4.6681465038845725e-05, "loss": 0.0696, "step": 7810 }, { "epoch": 12.73, "learning_rate": 4.634850166481687e-05, "loss": 0.0952, "step": 7820 }, { "epoch": 12.75, "learning_rate": 4.601553829078801e-05, "loss": 0.0658, "step": 7830 }, { "epoch": 12.76, "learning_rate": 4.568257491675915e-05, "loss": 0.087, "step": 7840 }, { "epoch": 12.78, "learning_rate": 4.5349611542730297e-05, "loss": 0.0678, "step": 7850 }, { "epoch": 12.8, "learning_rate": 4.5016648168701434e-05, "loss": 0.0851, "step": 7860 }, { "epoch": 12.81, "learning_rate": 4.4683684794672586e-05, "loss": 0.0991, "step": 7870 }, { "epoch": 12.83, "learning_rate": 4.4350721420643724e-05, "loss": 0.0622, "step": 7880 }, { "epoch": 12.85, "learning_rate": 4.401775804661487e-05, "loss": 0.0663, "step": 7890 }, { "epoch": 12.86, "learning_rate": 4.368479467258601e-05, "loss": 0.0669, "step": 7900 }, { "epoch": 12.86, "eval_loss": 0.0984959602355957, "eval_runtime": 1029.9234, "eval_samples_per_second": 11.272, "eval_steps_per_second": 0.705, "eval_wer": 0.28456415636546095, "step": 7900 }, { "epoch": 12.88, "learning_rate": 4.335183129855716e-05, "loss": 0.0772, "step": 7910 }, { "epoch": 12.9, "learning_rate": 4.3018867924528295e-05, "loss": 0.091, "step": 7920 }, { "epoch": 12.91, "learning_rate": 4.268590455049945e-05, "loss": 0.0564, "step": 7930 }, { "epoch": 12.93, "learning_rate": 4.2352941176470585e-05, "loss": 0.0644, "step": 7940 }, { "epoch": 12.94, "learning_rate": 4.201997780244173e-05, "loss": 0.0738, "step": 7950 }, { "epoch": 12.96, "learning_rate": 4.1687014428412874e-05, "loss": 0.0769, "step": 7960 }, { "epoch": 12.98, "learning_rate": 4.135405105438401e-05, "loss": 0.081, "step": 7970 }, { "epoch": 12.99, "learning_rate": 4.1021087680355156e-05, "loss": 0.0693, "step": 7980 }, { "epoch": 13.01, "learning_rate": 4.0688124306326294e-05, "loss": 0.073, "step": 7990 }, { "epoch": 13.03, "learning_rate": 4.0355160932297445e-05, "loss": 0.069, "step": 8000 }, { "epoch": 13.03, "eval_loss": 0.09989926964044571, "eval_runtime": 1024.9117, "eval_samples_per_second": 11.327, "eval_steps_per_second": 0.708, "eval_wer": 0.2844045066824796, "step": 8000 }, { "epoch": 13.04, "learning_rate": 4.002219755826858e-05, "loss": 0.0654, "step": 8010 }, { "epoch": 13.06, "learning_rate": 3.9689234184239735e-05, "loss": 0.0579, "step": 8020 }, { "epoch": 13.07, "learning_rate": 3.935627081021087e-05, "loss": 0.0734, "step": 8030 }, { "epoch": 13.09, "learning_rate": 3.902330743618202e-05, "loss": 0.0823, "step": 8040 }, { "epoch": 13.11, "learning_rate": 3.869034406215316e-05, "loss": 0.0712, "step": 8050 }, { "epoch": 13.12, "learning_rate": 3.8357380688124306e-05, "loss": 0.0659, "step": 8060 }, { "epoch": 13.14, "learning_rate": 3.8024417314095444e-05, "loss": 0.0703, "step": 8070 }, { "epoch": 13.16, "learning_rate": 3.7691453940066596e-05, "loss": 0.0837, "step": 8080 }, { "epoch": 13.17, "learning_rate": 3.735849056603773e-05, "loss": 0.0738, "step": 8090 }, { "epoch": 13.19, "learning_rate": 3.702552719200888e-05, "loss": 0.0674, "step": 8100 }, { "epoch": 13.19, "eval_loss": 0.1002257913351059, "eval_runtime": 1037.9605, "eval_samples_per_second": 11.184, "eval_steps_per_second": 0.699, "eval_wer": 0.2834694156821603, "step": 8100 }, { "epoch": 13.2, "learning_rate": 3.669256381798002e-05, "loss": 0.0637, "step": 8110 }, { "epoch": 13.22, "learning_rate": 3.635960044395116e-05, "loss": 0.0759, "step": 8120 }, { "epoch": 13.24, "learning_rate": 3.6026637069922305e-05, "loss": 0.0771, "step": 8130 }, { "epoch": 13.25, "learning_rate": 3.569367369589345e-05, "loss": 0.0668, "step": 8140 }, { "epoch": 13.27, "learning_rate": 3.5360710321864594e-05, "loss": 0.0605, "step": 8150 }, { "epoch": 13.29, "learning_rate": 3.502774694783573e-05, "loss": 0.0723, "step": 8160 }, { "epoch": 13.3, "learning_rate": 3.469478357380688e-05, "loss": 0.0683, "step": 8170 }, { "epoch": 13.32, "learning_rate": 3.436182019977802e-05, "loss": 0.0753, "step": 8180 }, { "epoch": 13.33, "learning_rate": 3.4028856825749166e-05, "loss": 0.0793, "step": 8190 }, { "epoch": 13.35, "learning_rate": 3.369589345172031e-05, "loss": 0.0695, "step": 8200 }, { "epoch": 13.35, "eval_loss": 0.10133621096611023, "eval_runtime": 1029.4384, "eval_samples_per_second": 11.277, "eval_steps_per_second": 0.705, "eval_wer": 0.28288783469415685, "step": 8200 }, { "epoch": 13.37, "learning_rate": 3.336293007769145e-05, "loss": 0.0636, "step": 8210 }, { "epoch": 13.38, "learning_rate": 3.302996670366259e-05, "loss": 0.0656, "step": 8220 }, { "epoch": 13.4, "learning_rate": 3.269700332963374e-05, "loss": 0.0717, "step": 8230 }, { "epoch": 13.42, "learning_rate": 3.236403995560488e-05, "loss": 0.0753, "step": 8240 }, { "epoch": 13.43, "learning_rate": 3.203107658157603e-05, "loss": 0.0604, "step": 8250 }, { "epoch": 13.45, "learning_rate": 3.169811320754717e-05, "loss": 0.0677, "step": 8260 }, { "epoch": 13.46, "learning_rate": 3.136514983351831e-05, "loss": 0.0664, "step": 8270 }, { "epoch": 13.48, "learning_rate": 3.1032186459489454e-05, "loss": 0.0711, "step": 8280 }, { "epoch": 13.5, "learning_rate": 3.069922308546059e-05, "loss": 0.0827, "step": 8290 }, { "epoch": 13.51, "learning_rate": 3.036625971143174e-05, "loss": 0.0578, "step": 8300 }, { "epoch": 13.51, "eval_loss": 0.10188211500644684, "eval_runtime": 1037.4819, "eval_samples_per_second": 11.19, "eval_steps_per_second": 0.7, "eval_wer": 0.28205537563289695, "step": 8300 }, { "epoch": 13.53, "learning_rate": 3.003329633740288e-05, "loss": 0.0681, "step": 8310 }, { "epoch": 13.55, "learning_rate": 2.9700332963374025e-05, "loss": 0.06, "step": 8320 }, { "epoch": 13.56, "learning_rate": 2.936736958934517e-05, "loss": 0.0731, "step": 8330 }, { "epoch": 13.58, "learning_rate": 2.903440621531631e-05, "loss": 0.0768, "step": 8340 }, { "epoch": 13.59, "learning_rate": 2.8701442841287456e-05, "loss": 0.0673, "step": 8350 }, { "epoch": 13.61, "learning_rate": 2.83684794672586e-05, "loss": 0.0693, "step": 8360 }, { "epoch": 13.63, "learning_rate": 2.8035516093229742e-05, "loss": 0.069, "step": 8370 }, { "epoch": 13.64, "learning_rate": 2.7702552719200886e-05, "loss": 0.0879, "step": 8380 }, { "epoch": 13.66, "learning_rate": 2.736958934517203e-05, "loss": 0.0831, "step": 8390 }, { "epoch": 13.68, "learning_rate": 2.7036625971143172e-05, "loss": 0.0614, "step": 8400 }, { "epoch": 13.68, "eval_loss": 0.09782098233699799, "eval_runtime": 1036.2973, "eval_samples_per_second": 11.202, "eval_steps_per_second": 0.701, "eval_wer": 0.2814509875473247, "step": 8400 }, { "epoch": 13.69, "learning_rate": 2.6703662597114317e-05, "loss": 0.0581, "step": 8410 }, { "epoch": 13.71, "learning_rate": 2.6370699223085458e-05, "loss": 0.0722, "step": 8420 }, { "epoch": 13.72, "learning_rate": 2.60377358490566e-05, "loss": 0.0744, "step": 8430 }, { "epoch": 13.74, "learning_rate": 2.5704772475027744e-05, "loss": 0.0752, "step": 8440 }, { "epoch": 13.76, "learning_rate": 2.537180910099889e-05, "loss": 0.0626, "step": 8450 }, { "epoch": 13.77, "learning_rate": 2.503884572697003e-05, "loss": 0.0632, "step": 8460 }, { "epoch": 13.79, "learning_rate": 2.4705882352941174e-05, "loss": 0.079, "step": 8470 }, { "epoch": 13.81, "learning_rate": 2.437291897891232e-05, "loss": 0.0706, "step": 8480 }, { "epoch": 13.82, "learning_rate": 2.403995560488346e-05, "loss": 0.0754, "step": 8490 }, { "epoch": 13.84, "learning_rate": 2.3706992230854605e-05, "loss": 0.0554, "step": 8500 }, { "epoch": 13.84, "eval_loss": 0.09835359454154968, "eval_runtime": 1033.9977, "eval_samples_per_second": 11.227, "eval_steps_per_second": 0.702, "eval_wer": 0.28125712721799023, "step": 8500 }, { "epoch": 13.86, "learning_rate": 2.337402885682575e-05, "loss": 0.0716, "step": 8510 }, { "epoch": 13.87, "learning_rate": 2.304106548279689e-05, "loss": 0.0684, "step": 8520 }, { "epoch": 13.89, "learning_rate": 2.2708102108768035e-05, "loss": 0.0753, "step": 8530 }, { "epoch": 13.9, "learning_rate": 2.237513873473918e-05, "loss": 0.0738, "step": 8540 }, { "epoch": 13.92, "learning_rate": 2.2042175360710318e-05, "loss": 0.0703, "step": 8550 }, { "epoch": 13.94, "learning_rate": 2.1709211986681462e-05, "loss": 0.069, "step": 8560 }, { "epoch": 13.95, "learning_rate": 2.1376248612652603e-05, "loss": 0.0696, "step": 8570 }, { "epoch": 13.97, "learning_rate": 2.1043285238623748e-05, "loss": 0.0794, "step": 8580 }, { "epoch": 13.99, "learning_rate": 2.0710321864594893e-05, "loss": 0.0776, "step": 8590 }, { "epoch": 14.0, "learning_rate": 2.0377358490566034e-05, "loss": 0.0763, "step": 8600 }, { "epoch": 14.0, "eval_loss": 0.10010204464197159, "eval_runtime": 1039.8833, "eval_samples_per_second": 11.164, "eval_steps_per_second": 0.698, "eval_wer": 0.2812685307667746, "step": 8600 }, { "epoch": 14.02, "learning_rate": 2.004439511653718e-05, "loss": 0.0573, "step": 8610 }, { "epoch": 14.03, "learning_rate": 1.9711431742508323e-05, "loss": 0.0604, "step": 8620 }, { "epoch": 14.05, "learning_rate": 1.9378468368479468e-05, "loss": 0.066, "step": 8630 }, { "epoch": 14.07, "learning_rate": 1.904550499445061e-05, "loss": 0.0607, "step": 8640 }, { "epoch": 14.08, "learning_rate": 1.871254162042175e-05, "loss": 0.0858, "step": 8650 }, { "epoch": 14.1, "learning_rate": 1.8379578246392895e-05, "loss": 0.073, "step": 8660 }, { "epoch": 14.12, "learning_rate": 1.804661487236404e-05, "loss": 0.0602, "step": 8670 }, { "epoch": 14.13, "learning_rate": 1.771365149833518e-05, "loss": 0.077, "step": 8680 }, { "epoch": 14.15, "learning_rate": 1.7380688124306325e-05, "loss": 0.0663, "step": 8690 }, { "epoch": 14.16, "learning_rate": 1.704772475027747e-05, "loss": 0.0877, "step": 8700 }, { "epoch": 14.16, "eval_loss": 0.10002926737070084, "eval_runtime": 1035.7641, "eval_samples_per_second": 11.208, "eval_steps_per_second": 0.701, "eval_wer": 0.2808351959129681, "step": 8700 }, { "epoch": 14.18, "learning_rate": 1.671476137624861e-05, "loss": 0.0557, "step": 8710 }, { "epoch": 14.2, "learning_rate": 1.6381798002219752e-05, "loss": 0.0642, "step": 8720 }, { "epoch": 14.21, "learning_rate": 1.6048834628190897e-05, "loss": 0.0689, "step": 8730 }, { "epoch": 14.23, "learning_rate": 1.571587125416204e-05, "loss": 0.0753, "step": 8740 }, { "epoch": 14.25, "learning_rate": 1.5382907880133183e-05, "loss": 0.0872, "step": 8750 }, { "epoch": 14.26, "learning_rate": 1.5049944506104327e-05, "loss": 0.0642, "step": 8760 }, { "epoch": 14.28, "learning_rate": 1.4716981132075472e-05, "loss": 0.0618, "step": 8770 }, { "epoch": 14.29, "learning_rate": 1.4384017758046613e-05, "loss": 0.0666, "step": 8780 }, { "epoch": 14.31, "learning_rate": 1.4051054384017756e-05, "loss": 0.0757, "step": 8790 }, { "epoch": 14.33, "learning_rate": 1.3718091009988899e-05, "loss": 0.0882, "step": 8800 }, { "epoch": 14.33, "eval_loss": 0.09793306142091751, "eval_runtime": 1034.3628, "eval_samples_per_second": 11.223, "eval_steps_per_second": 0.702, "eval_wer": 0.2802764220225334, "step": 8800 }, { "epoch": 14.34, "learning_rate": 1.3385127635960044e-05, "loss": 0.0671, "step": 8810 }, { "epoch": 14.36, "learning_rate": 1.3052164261931187e-05, "loss": 0.0598, "step": 8820 }, { "epoch": 14.38, "learning_rate": 1.2719200887902331e-05, "loss": 0.0653, "step": 8830 }, { "epoch": 14.39, "learning_rate": 1.2386237513873472e-05, "loss": 0.0745, "step": 8840 }, { "epoch": 14.41, "learning_rate": 1.2053274139844615e-05, "loss": 0.0852, "step": 8850 }, { "epoch": 14.42, "learning_rate": 1.1720310765815758e-05, "loss": 0.0534, "step": 8860 }, { "epoch": 14.44, "learning_rate": 1.1387347391786903e-05, "loss": 0.0621, "step": 8870 }, { "epoch": 14.46, "learning_rate": 1.1054384017758046e-05, "loss": 0.0651, "step": 8880 }, { "epoch": 14.47, "learning_rate": 1.0721420643729189e-05, "loss": 0.0624, "step": 8890 }, { "epoch": 14.49, "learning_rate": 1.0388457269700333e-05, "loss": 0.0864, "step": 8900 }, { "epoch": 14.49, "eval_loss": 0.09810327738523483, "eval_runtime": 1033.3407, "eval_samples_per_second": 11.234, "eval_steps_per_second": 0.703, "eval_wer": 0.27882817132691695, "step": 8900 }, { "epoch": 14.51, "learning_rate": 1.0055493895671474e-05, "loss": 0.067, "step": 8910 }, { "epoch": 14.52, "learning_rate": 9.722530521642617e-06, "loss": 0.0628, "step": 8920 }, { "epoch": 14.54, "learning_rate": 9.389567147613762e-06, "loss": 0.0686, "step": 8930 }, { "epoch": 14.55, "learning_rate": 9.056603773584905e-06, "loss": 0.0704, "step": 8940 }, { "epoch": 14.57, "learning_rate": 8.723640399556048e-06, "loss": 0.0845, "step": 8950 }, { "epoch": 14.59, "learning_rate": 8.39067702552719e-06, "loss": 0.0525, "step": 8960 }, { "epoch": 14.6, "learning_rate": 8.057713651498334e-06, "loss": 0.0591, "step": 8970 }, { "epoch": 14.62, "learning_rate": 7.724750277469478e-06, "loss": 0.069, "step": 8980 }, { "epoch": 14.64, "learning_rate": 7.39178690344062e-06, "loss": 0.0759, "step": 8990 }, { "epoch": 14.65, "learning_rate": 7.058823529411764e-06, "loss": 0.0828, "step": 9000 }, { "epoch": 14.65, "eval_loss": 0.09748378396034241, "eval_runtime": 1034.9341, "eval_samples_per_second": 11.217, "eval_steps_per_second": 0.701, "eval_wer": 0.279010628107467, "step": 9000 }, { "epoch": 14.67, "learning_rate": 2.543640897755611e-05, "loss": 10.9581, "step": 9010 }, { "epoch": 14.69, "learning_rate": 2.5124688279301746e-05, "loss": 5.0219, "step": 9020 }, { "epoch": 14.71, "learning_rate": 2.481296758104738e-05, "loss": 3.3068, "step": 9030 }, { "epoch": 14.72, "learning_rate": 2.4501246882793016e-05, "loss": 2.8649, "step": 9040 }, { "epoch": 14.74, "learning_rate": 2.4189526184538654e-05, "loss": 2.4786, "step": 9050 }, { "epoch": 14.76, "learning_rate": 2.387780548628429e-05, "loss": 1.759, "step": 9060 }, { "epoch": 14.77, "learning_rate": 2.3566084788029924e-05, "loss": 0.995, "step": 9070 }, { "epoch": 14.79, "learning_rate": 2.3254364089775562e-05, "loss": 0.5639, "step": 9080 }, { "epoch": 14.8, "learning_rate": 2.2942643391521197e-05, "loss": 0.3693, "step": 9090 }, { "epoch": 14.82, "learning_rate": 2.2630922693266832e-05, "loss": 0.3052, "step": 9100 }, { "epoch": 14.82, "eval_loss": 0.21501106023788452, "eval_runtime": 1054.7107, "eval_samples_per_second": 11.007, "eval_steps_per_second": 0.688, "eval_wer": 0.41747251744742964, "step": 9100 }, { "epoch": 14.84, "learning_rate": 2.231920199501247e-05, "loss": 0.2253, "step": 9110 }, { "epoch": 14.85, "learning_rate": 2.2007481296758105e-05, "loss": 0.1596, "step": 9120 }, { "epoch": 14.87, "learning_rate": 2.169576059850374e-05, "loss": 0.1664, "step": 9130 }, { "epoch": 14.89, "learning_rate": 2.1384039900249375e-05, "loss": 0.1484, "step": 9140 }, { "epoch": 14.9, "learning_rate": 2.1072319201995013e-05, "loss": 0.1696, "step": 9150 }, { "epoch": 14.92, "learning_rate": 2.0760598503740648e-05, "loss": 0.1394, "step": 9160 }, { "epoch": 14.93, "learning_rate": 2.0448877805486283e-05, "loss": 0.1212, "step": 9170 }, { "epoch": 14.95, "learning_rate": 2.013715710723192e-05, "loss": 0.121, "step": 9180 }, { "epoch": 14.97, "learning_rate": 1.9825436408977556e-05, "loss": 0.1263, "step": 9190 }, { "epoch": 14.98, "learning_rate": 1.951371571072319e-05, "loss": 0.1478, "step": 9200 }, { "epoch": 14.98, "eval_loss": 0.13250307738780975, "eval_runtime": 1044.4894, "eval_samples_per_second": 11.115, "eval_steps_per_second": 0.695, "eval_wer": 0.3026615882862747, "step": 9200 }, { "epoch": 15.0, "learning_rate": 7.629139072847681e-05, "loss": 6.7067, "step": 9210 }, { "epoch": 15.02, "learning_rate": 7.60430463576159e-05, "loss": 3.5775, "step": 9220 }, { "epoch": 15.03, "learning_rate": 7.579470198675495e-05, "loss": 3.1504, "step": 9230 }, { "epoch": 15.05, "learning_rate": 7.554635761589404e-05, "loss": 3.0605, "step": 9240 }, { "epoch": 15.06, "learning_rate": 7.52980132450331e-05, "loss": 3.0235, "step": 9250 }, { "epoch": 15.08, "learning_rate": 7.504966887417217e-05, "loss": 2.9939, "step": 9260 }, { "epoch": 15.1, "learning_rate": 7.480132450331126e-05, "loss": 2.9837, "step": 9270 }, { "epoch": 15.11, "learning_rate": 7.455298013245033e-05, "loss": 2.8137, "step": 9280 }, { "epoch": 15.13, "learning_rate": 7.43046357615894e-05, "loss": 1.9888, "step": 9290 }, { "epoch": 15.15, "learning_rate": 7.405629139072846e-05, "loss": 1.0386, "step": 9300 }, { "epoch": 15.15, "eval_loss": 0.43749570846557617, "eval_runtime": 1037.1435, "eval_samples_per_second": 11.193, "eval_steps_per_second": 0.7, "eval_wer": 0.679286593988049, "step": 9300 }, { "epoch": 15.16, "learning_rate": 7.380794701986755e-05, "loss": 0.4016, "step": 9310 }, { "epoch": 15.18, "learning_rate": 7.355960264900662e-05, "loss": 0.23, "step": 9320 }, { "epoch": 15.19, "learning_rate": 7.331125827814569e-05, "loss": 0.1834, "step": 9330 }, { "epoch": 15.21, "learning_rate": 7.306291390728476e-05, "loss": 0.149, "step": 9340 }, { "epoch": 15.23, "learning_rate": 7.281456953642384e-05, "loss": 0.1275, "step": 9350 }, { "epoch": 15.24, "learning_rate": 7.256622516556291e-05, "loss": 0.1483, "step": 9360 }, { "epoch": 15.26, "learning_rate": 7.231788079470198e-05, "loss": 0.1038, "step": 9370 }, { "epoch": 15.28, "learning_rate": 7.206953642384106e-05, "loss": 0.1185, "step": 9380 }, { "epoch": 15.29, "learning_rate": 7.182119205298013e-05, "loss": 0.1099, "step": 9390 }, { "epoch": 15.31, "learning_rate": 7.15728476821192e-05, "loss": 0.116, "step": 9400 }, { "epoch": 15.31, "eval_loss": 0.12657363712787628, "eval_runtime": 1045.2349, "eval_samples_per_second": 11.107, "eval_steps_per_second": 0.695, "eval_wer": 0.30423527801851935, "step": 9400 }, { "epoch": 15.32, "learning_rate": 7.132450331125827e-05, "loss": 0.1548, "step": 9410 }, { "epoch": 15.34, "learning_rate": 7.107615894039734e-05, "loss": 0.0953, "step": 9420 }, { "epoch": 15.36, "learning_rate": 7.082781456953642e-05, "loss": 0.0977, "step": 9430 }, { "epoch": 15.37, "learning_rate": 7.057947019867549e-05, "loss": 0.1106, "step": 9440 }, { "epoch": 15.39, "learning_rate": 7.033112582781457e-05, "loss": 0.093, "step": 9450 }, { "epoch": 15.41, "learning_rate": 7.008278145695364e-05, "loss": 0.114, "step": 9460 }, { "epoch": 15.42, "learning_rate": 6.983443708609271e-05, "loss": 0.1031, "step": 9470 }, { "epoch": 15.44, "learning_rate": 6.958609271523178e-05, "loss": 0.0946, "step": 9480 }, { "epoch": 15.45, "learning_rate": 6.933774834437085e-05, "loss": 0.0945, "step": 9490 }, { "epoch": 15.47, "learning_rate": 6.908940397350992e-05, "loss": 0.1226, "step": 9500 }, { "epoch": 15.47, "eval_loss": 0.1205865815281868, "eval_runtime": 1053.9637, "eval_samples_per_second": 11.015, "eval_steps_per_second": 0.689, "eval_wer": 0.2999703507731606, "step": 9500 }, { "epoch": 15.49, "learning_rate": 6.8841059602649e-05, "loss": 0.1336, "step": 9510 }, { "epoch": 15.5, "learning_rate": 6.859271523178807e-05, "loss": 0.086, "step": 9520 }, { "epoch": 15.52, "learning_rate": 6.834437086092714e-05, "loss": 0.1, "step": 9530 }, { "epoch": 15.54, "learning_rate": 6.809602649006622e-05, "loss": 0.094, "step": 9540 }, { "epoch": 15.55, "learning_rate": 6.784768211920529e-05, "loss": 0.0925, "step": 9550 }, { "epoch": 15.57, "learning_rate": 6.759933774834436e-05, "loss": 0.1214, "step": 9560 }, { "epoch": 15.59, "learning_rate": 6.735099337748343e-05, "loss": 0.095, "step": 9570 }, { "epoch": 15.6, "learning_rate": 6.710264900662251e-05, "loss": 0.0789, "step": 9580 }, { "epoch": 15.62, "learning_rate": 6.685430463576158e-05, "loss": 0.0932, "step": 9590 }, { "epoch": 15.63, "learning_rate": 6.660596026490065e-05, "loss": 0.0885, "step": 9600 }, { "epoch": 15.63, "eval_loss": 0.11733737587928772, "eval_runtime": 1040.6981, "eval_samples_per_second": 11.155, "eval_steps_per_second": 0.698, "eval_wer": 0.29583086256443003, "step": 9600 }, { "epoch": 15.65, "learning_rate": 6.635761589403972e-05, "loss": 0.1138, "step": 9610 }, { "epoch": 15.67, "learning_rate": 6.61092715231788e-05, "loss": 0.087, "step": 9620 }, { "epoch": 15.68, "learning_rate": 6.586092715231787e-05, "loss": 0.0932, "step": 9630 }, { "epoch": 15.7, "learning_rate": 6.561258278145696e-05, "loss": 0.0834, "step": 9640 }, { "epoch": 15.72, "learning_rate": 6.536423841059603e-05, "loss": 0.089, "step": 9650 }, { "epoch": 15.73, "learning_rate": 6.51158940397351e-05, "loss": 0.1027, "step": 9660 }, { "epoch": 15.75, "learning_rate": 6.486754966887417e-05, "loss": 0.0795, "step": 9670 }, { "epoch": 15.76, "learning_rate": 6.461920529801323e-05, "loss": 0.0748, "step": 9680 }, { "epoch": 15.78, "learning_rate": 6.43708609271523e-05, "loss": 0.0812, "step": 9690 }, { "epoch": 15.8, "learning_rate": 6.412251655629139e-05, "loss": 0.091, "step": 9700 }, { "epoch": 15.8, "eval_loss": 0.1144639402627945, "eval_runtime": 1044.6721, "eval_samples_per_second": 11.113, "eval_steps_per_second": 0.695, "eval_wer": 0.2928659398804908, "step": 9700 }, { "epoch": 15.81, "learning_rate": 6.387417218543046e-05, "loss": 0.1188, "step": 9710 }, { "epoch": 15.83, "learning_rate": 6.362582781456954e-05, "loss": 0.0814, "step": 9720 }, { "epoch": 15.85, "learning_rate": 6.337748344370861e-05, "loss": 0.0777, "step": 9730 }, { "epoch": 15.86, "learning_rate": 6.312913907284768e-05, "loss": 0.0798, "step": 9740 }, { "epoch": 15.88, "learning_rate": 6.288079470198675e-05, "loss": 0.0851, "step": 9750 }, { "epoch": 15.89, "learning_rate": 6.263245033112582e-05, "loss": 0.1165, "step": 9760 }, { "epoch": 15.91, "learning_rate": 6.23841059602649e-05, "loss": 0.0809, "step": 9770 }, { "epoch": 15.93, "learning_rate": 6.213576158940397e-05, "loss": 0.0719, "step": 9780 }, { "epoch": 15.94, "learning_rate": 6.188741721854304e-05, "loss": 0.0865, "step": 9790 }, { "epoch": 15.96, "learning_rate": 6.163907284768211e-05, "loss": 0.0886, "step": 9800 }, { "epoch": 15.96, "eval_loss": 0.11122256517410278, "eval_runtime": 1040.8284, "eval_samples_per_second": 11.154, "eval_steps_per_second": 0.698, "eval_wer": 0.29079049400173335, "step": 9800 }, { "epoch": 15.98, "learning_rate": 6.139072847682119e-05, "loss": 0.0927, "step": 9810 }, { "epoch": 15.99, "learning_rate": 6.114238410596026e-05, "loss": 0.0796, "step": 9820 }, { "epoch": 16.01, "learning_rate": 6.089403973509933e-05, "loss": 0.0906, "step": 9830 }, { "epoch": 16.02, "learning_rate": 6.064569536423841e-05, "loss": 0.0745, "step": 9840 }, { "epoch": 16.04, "learning_rate": 6.039735099337748e-05, "loss": 0.0839, "step": 9850 }, { "epoch": 16.06, "learning_rate": 6.014900662251655e-05, "loss": 0.0741, "step": 9860 }, { "epoch": 16.07, "learning_rate": 5.9900662251655627e-05, "loss": 0.0915, "step": 9870 }, { "epoch": 16.09, "learning_rate": 5.9652317880794696e-05, "loss": 0.0863, "step": 9880 }, { "epoch": 16.11, "learning_rate": 5.9403973509933765e-05, "loss": 0.074, "step": 9890 }, { "epoch": 16.12, "learning_rate": 5.915562913907285e-05, "loss": 0.0783, "step": 9900 }, { "epoch": 16.12, "eval_loss": 0.1075427234172821, "eval_runtime": 1040.0379, "eval_samples_per_second": 11.162, "eval_steps_per_second": 0.698, "eval_wer": 0.28727820097614376, "step": 9900 }, { "epoch": 16.14, "learning_rate": 5.890728476821192e-05, "loss": 0.0828, "step": 9910 }, { "epoch": 16.15, "learning_rate": 5.865894039735099e-05, "loss": 0.0841, "step": 9920 }, { "epoch": 16.17, "learning_rate": 5.841059602649006e-05, "loss": 0.0894, "step": 9930 }, { "epoch": 16.19, "learning_rate": 5.816225165562913e-05, "loss": 0.0777, "step": 9940 }, { "epoch": 16.2, "learning_rate": 5.79139072847682e-05, "loss": 0.0773, "step": 9950 }, { "epoch": 16.22, "learning_rate": 5.766556291390728e-05, "loss": 0.083, "step": 9960 }, { "epoch": 16.24, "learning_rate": 5.7417218543046354e-05, "loss": 0.0864, "step": 9970 }, { "epoch": 16.25, "learning_rate": 5.716887417218543e-05, "loss": 0.0916, "step": 9980 }, { "epoch": 16.27, "learning_rate": 5.69205298013245e-05, "loss": 0.0703, "step": 9990 }, { "epoch": 16.28, "learning_rate": 5.667218543046357e-05, "loss": 0.069, "step": 10000 }, { "epoch": 16.28, "eval_loss": 0.10723693668842316, "eval_runtime": 1040.1539, "eval_samples_per_second": 11.161, "eval_steps_per_second": 0.698, "eval_wer": 0.2876203074396752, "step": 10000 }, { "epoch": 16.3, "learning_rate": 5.6423841059602645e-05, "loss": 0.078, "step": 10010 }, { "epoch": 16.32, "learning_rate": 5.6175496688741714e-05, "loss": 0.0851, "step": 10020 }, { "epoch": 16.33, "learning_rate": 5.592715231788078e-05, "loss": 0.0898, "step": 10030 }, { "epoch": 16.35, "learning_rate": 5.5678807947019866e-05, "loss": 0.0649, "step": 10040 }, { "epoch": 16.37, "learning_rate": 5.5430463576158935e-05, "loss": 0.0717, "step": 10050 }, { "epoch": 16.38, "learning_rate": 5.518211920529801e-05, "loss": 0.0893, "step": 10060 }, { "epoch": 16.4, "learning_rate": 5.493377483443708e-05, "loss": 0.0755, "step": 10070 }, { "epoch": 16.41, "learning_rate": 5.468543046357615e-05, "loss": 0.0921, "step": 10080 }, { "epoch": 16.43, "learning_rate": 5.4437086092715226e-05, "loss": 0.0788, "step": 10090 }, { "epoch": 16.45, "learning_rate": 5.41887417218543e-05, "loss": 0.0783, "step": 10100 }, { "epoch": 16.45, "eval_loss": 0.10703727602958679, "eval_runtime": 1062.4638, "eval_samples_per_second": 10.926, "eval_steps_per_second": 0.683, "eval_wer": 0.28756328969575334, "step": 10100 }, { "epoch": 16.46, "learning_rate": 5.394039735099338e-05, "loss": 0.0742, "step": 10110 }, { "epoch": 16.48, "learning_rate": 5.369205298013245e-05, "loss": 0.0877, "step": 10120 }, { "epoch": 16.5, "learning_rate": 5.344370860927152e-05, "loss": 0.0882, "step": 10130 }, { "epoch": 16.51, "learning_rate": 5.319536423841059e-05, "loss": 0.0695, "step": 10140 }, { "epoch": 16.53, "learning_rate": 5.294701986754966e-05, "loss": 0.0738, "step": 10150 }, { "epoch": 16.55, "learning_rate": 5.269867549668873e-05, "loss": 0.0712, "step": 10160 }, { "epoch": 16.56, "learning_rate": 5.2450331125827815e-05, "loss": 0.0732, "step": 10170 }, { "epoch": 16.58, "learning_rate": 5.2201986754966884e-05, "loss": 0.0907, "step": 10180 }, { "epoch": 16.59, "learning_rate": 5.195364238410595e-05, "loss": 0.063, "step": 10190 }, { "epoch": 16.61, "learning_rate": 5.170529801324503e-05, "loss": 0.0669, "step": 10200 }, { "epoch": 16.61, "eval_loss": 0.10548633337020874, "eval_runtime": 1068.0412, "eval_samples_per_second": 10.869, "eval_steps_per_second": 0.68, "eval_wer": 0.28480363088993293, "step": 10200 }, { "epoch": 16.63, "learning_rate": 5.14569536423841e-05, "loss": 0.0738, "step": 10210 }, { "epoch": 16.64, "learning_rate": 5.120860927152317e-05, "loss": 0.0803, "step": 10220 }, { "epoch": 16.66, "learning_rate": 5.096026490066225e-05, "loss": 0.091, "step": 10230 }, { "epoch": 16.68, "learning_rate": 5.071192052980132e-05, "loss": 0.0707, "step": 10240 }, { "epoch": 16.69, "learning_rate": 5.0463576158940396e-05, "loss": 0.0635, "step": 10250 }, { "epoch": 16.71, "learning_rate": 5.0215231788079466e-05, "loss": 0.0712, "step": 10260 }, { "epoch": 16.72, "learning_rate": 4.9966887417218535e-05, "loss": 0.098, "step": 10270 }, { "epoch": 16.74, "learning_rate": 4.971854304635761e-05, "loss": 0.1061, "step": 10280 }, { "epoch": 16.76, "learning_rate": 4.947019867549668e-05, "loss": 0.069, "step": 10290 }, { "epoch": 16.77, "learning_rate": 4.922185430463576e-05, "loss": 0.072, "step": 10300 }, { "epoch": 16.77, "eval_loss": 0.10425380617380142, "eval_runtime": 1065.0981, "eval_samples_per_second": 10.899, "eval_steps_per_second": 0.682, "eval_wer": 0.28463257765816724, "step": 10300 }, { "epoch": 16.79, "learning_rate": 4.897350993377483e-05, "loss": 0.068, "step": 10310 }, { "epoch": 16.81, "learning_rate": 4.87251655629139e-05, "loss": 0.0797, "step": 10320 }, { "epoch": 16.82, "learning_rate": 4.847682119205298e-05, "loss": 0.083, "step": 10330 }, { "epoch": 16.84, "learning_rate": 4.822847682119205e-05, "loss": 0.0684, "step": 10340 }, { "epoch": 16.85, "learning_rate": 4.798013245033112e-05, "loss": 0.0671, "step": 10350 }, { "epoch": 16.87, "learning_rate": 4.77317880794702e-05, "loss": 0.0683, "step": 10360 }, { "epoch": 16.89, "learning_rate": 4.748344370860927e-05, "loss": 0.081, "step": 10370 }, { "epoch": 16.9, "learning_rate": 4.7235099337748345e-05, "loss": 0.092, "step": 10380 }, { "epoch": 16.92, "learning_rate": 4.6986754966887414e-05, "loss": 0.068, "step": 10390 }, { "epoch": 16.94, "learning_rate": 4.6738410596026484e-05, "loss": 0.0721, "step": 10400 }, { "epoch": 16.94, "eval_loss": 0.10198366641998291, "eval_runtime": 1065.8653, "eval_samples_per_second": 10.892, "eval_steps_per_second": 0.681, "eval_wer": 0.2820895862792501, "step": 10400 }, { "epoch": 16.95, "learning_rate": 4.649006622516556e-05, "loss": 0.0774, "step": 10410 }, { "epoch": 16.97, "learning_rate": 4.624172185430463e-05, "loss": 0.0917, "step": 10420 }, { "epoch": 16.98, "learning_rate": 4.5993377483443705e-05, "loss": 0.0816, "step": 10430 }, { "epoch": 17.0, "learning_rate": 4.574503311258278e-05, "loss": 0.0686, "step": 10440 }, { "epoch": 17.02, "learning_rate": 4.549668874172185e-05, "loss": 0.0662, "step": 10450 }, { "epoch": 17.03, "learning_rate": 4.524834437086092e-05, "loss": 0.0622, "step": 10460 }, { "epoch": 17.05, "learning_rate": 4.4999999999999996e-05, "loss": 0.0787, "step": 10470 }, { "epoch": 17.07, "learning_rate": 4.4751655629139065e-05, "loss": 0.0847, "step": 10480 }, { "epoch": 17.08, "learning_rate": 4.450331125827815e-05, "loss": 0.0917, "step": 10490 }, { "epoch": 17.1, "learning_rate": 4.425496688741722e-05, "loss": 0.0694, "step": 10500 }, { "epoch": 17.1, "eval_loss": 0.1046956330537796, "eval_runtime": 1061.7404, "eval_samples_per_second": 10.934, "eval_steps_per_second": 0.684, "eval_wer": 0.2802764220225334, "step": 10500 }, { "epoch": 17.11, "learning_rate": 4.400662251655629e-05, "loss": 0.0708, "step": 10510 }, { "epoch": 17.13, "learning_rate": 4.375827814569536e-05, "loss": 0.0676, "step": 10520 }, { "epoch": 17.15, "learning_rate": 4.350993377483443e-05, "loss": 0.0695, "step": 10530 }, { "epoch": 17.16, "learning_rate": 4.32615894039735e-05, "loss": 0.0889, "step": 10540 }, { "epoch": 17.18, "learning_rate": 4.301324503311258e-05, "loss": 0.0671, "step": 10550 }, { "epoch": 17.2, "learning_rate": 4.2764900662251654e-05, "loss": 0.0589, "step": 10560 }, { "epoch": 17.21, "learning_rate": 4.251655629139073e-05, "loss": 0.0769, "step": 10570 }, { "epoch": 17.23, "learning_rate": 4.22682119205298e-05, "loss": 0.0735, "step": 10580 }, { "epoch": 17.24, "learning_rate": 4.201986754966887e-05, "loss": 0.0944, "step": 10590 }, { "epoch": 17.26, "learning_rate": 4.1771523178807945e-05, "loss": 0.0574, "step": 10600 }, { "epoch": 17.26, "eval_loss": 0.10528343170881271, "eval_runtime": 1049.4398, "eval_samples_per_second": 11.062, "eval_steps_per_second": 0.692, "eval_wer": 0.2829790630844319, "step": 10600 }, { "epoch": 17.28, "learning_rate": 4.1523178807947014e-05, "loss": 0.0693, "step": 10610 }, { "epoch": 17.29, "learning_rate": 4.127483443708608e-05, "loss": 0.0716, "step": 10620 }, { "epoch": 17.31, "learning_rate": 4.1026490066225166e-05, "loss": 0.0773, "step": 10630 }, { "epoch": 17.33, "learning_rate": 4.0778145695364236e-05, "loss": 0.0825, "step": 10640 }, { "epoch": 17.34, "learning_rate": 4.0529801324503305e-05, "loss": 0.0624, "step": 10650 }, { "epoch": 17.36, "learning_rate": 4.028145695364238e-05, "loss": 0.077, "step": 10660 }, { "epoch": 17.38, "learning_rate": 4.003311258278145e-05, "loss": 0.0673, "step": 10670 }, { "epoch": 17.39, "learning_rate": 3.978476821192052e-05, "loss": 0.0698, "step": 10680 }, { "epoch": 17.41, "learning_rate": 3.95364238410596e-05, "loss": 0.0917, "step": 10690 }, { "epoch": 17.42, "learning_rate": 3.928807947019867e-05, "loss": 0.0578, "step": 10700 }, { "epoch": 17.42, "eval_loss": 0.10418736189603806, "eval_runtime": 1051.2916, "eval_samples_per_second": 11.043, "eval_steps_per_second": 0.691, "eval_wer": 0.28060712493728046, "step": 10700 }, { "epoch": 17.44, "learning_rate": 3.903973509933775e-05, "loss": 0.0729, "step": 10710 }, { "epoch": 17.46, "learning_rate": 3.879139072847682e-05, "loss": 0.0667, "step": 10720 }, { "epoch": 17.47, "learning_rate": 3.8543046357615887e-05, "loss": 0.0706, "step": 10730 }, { "epoch": 17.49, "learning_rate": 3.829470198675496e-05, "loss": 0.0918, "step": 10740 }, { "epoch": 17.51, "learning_rate": 3.804635761589403e-05, "loss": 0.0677, "step": 10750 }, { "epoch": 17.52, "learning_rate": 3.7798013245033115e-05, "loss": 0.0637, "step": 10760 }, { "epoch": 17.54, "learning_rate": 3.7549668874172184e-05, "loss": 0.0747, "step": 10770 }, { "epoch": 17.55, "learning_rate": 3.7301324503311254e-05, "loss": 0.0792, "step": 10780 }, { "epoch": 17.57, "learning_rate": 3.705298013245033e-05, "loss": 0.0811, "step": 10790 }, { "epoch": 17.59, "learning_rate": 3.68046357615894e-05, "loss": 0.0663, "step": 10800 }, { "epoch": 17.59, "eval_loss": 0.10351713746786118, "eval_runtime": 1048.2807, "eval_samples_per_second": 11.074, "eval_steps_per_second": 0.693, "eval_wer": 0.2800939652419833, "step": 10800 }, { "epoch": 17.6, "learning_rate": 3.6556291390728475e-05, "loss": 0.0697, "step": 10810 }, { "epoch": 17.62, "learning_rate": 3.6307947019867544e-05, "loss": 0.0663, "step": 10820 }, { "epoch": 17.64, "learning_rate": 3.605960264900662e-05, "loss": 0.0702, "step": 10830 }, { "epoch": 17.65, "learning_rate": 3.5811258278145697e-05, "loss": 0.0877, "step": 10840 }, { "epoch": 17.67, "learning_rate": 3.5562913907284766e-05, "loss": 0.0669, "step": 10850 }, { "epoch": 17.68, "learning_rate": 3.5314569536423835e-05, "loss": 0.0696, "step": 10860 }, { "epoch": 17.7, "learning_rate": 3.506622516556291e-05, "loss": 0.0677, "step": 10870 }, { "epoch": 17.72, "learning_rate": 3.481788079470199e-05, "loss": 0.0791, "step": 10880 }, { "epoch": 17.73, "learning_rate": 3.456953642384106e-05, "loss": 0.0886, "step": 10890 }, { "epoch": 17.75, "learning_rate": 3.4321192052980126e-05, "loss": 0.0615, "step": 10900 }, { "epoch": 17.75, "eval_loss": 0.10250340402126312, "eval_runtime": 1058.0947, "eval_samples_per_second": 10.972, "eval_steps_per_second": 0.686, "eval_wer": 0.2784974684121699, "step": 10900 }, { "epoch": 17.77, "learning_rate": 3.40728476821192e-05, "loss": 0.0714, "step": 10910 }, { "epoch": 17.78, "learning_rate": 3.382450331125827e-05, "loss": 0.0706, "step": 10920 }, { "epoch": 17.8, "learning_rate": 3.357615894039735e-05, "loss": 0.0621, "step": 10930 }, { "epoch": 17.81, "learning_rate": 3.3327814569536424e-05, "loss": 0.0985, "step": 10940 }, { "epoch": 17.83, "learning_rate": 3.307947019867549e-05, "loss": 0.0594, "step": 10950 }, { "epoch": 17.85, "learning_rate": 3.283112582781456e-05, "loss": 0.0676, "step": 10960 }, { "epoch": 17.86, "learning_rate": 3.258278145695364e-05, "loss": 0.0621, "step": 10970 }, { "epoch": 17.88, "learning_rate": 3.2334437086092715e-05, "loss": 0.0702, "step": 10980 }, { "epoch": 17.9, "learning_rate": 3.2086092715231784e-05, "loss": 0.0865, "step": 10990 }, { "epoch": 17.91, "learning_rate": 3.183774834437086e-05, "loss": 0.0706, "step": 11000 }, { "epoch": 17.91, "eval_loss": 0.10284795612096786, "eval_runtime": 1051.8233, "eval_samples_per_second": 11.037, "eval_steps_per_second": 0.69, "eval_wer": 0.2791702777904484, "step": 11000 }, { "epoch": 17.93, "learning_rate": 3.158940397350993e-05, "loss": 10.0938, "step": 11010 }, { "epoch": 17.95, "learning_rate": 3.1341059602649005e-05, "loss": 4.2095, "step": 11020 }, { "epoch": 17.96, "learning_rate": 3.1092715231788075e-05, "loss": 3.0169, "step": 11030 }, { "epoch": 17.98, "learning_rate": 3.084437086092715e-05, "loss": 2.3365, "step": 11040 }, { "epoch": 18.0, "learning_rate": 3.059602649006622e-05, "loss": 1.521, "step": 11050 }, { "epoch": 18.01, "learning_rate": 3.0347682119205293e-05, "loss": 0.9603, "step": 11060 }, { "epoch": 18.03, "learning_rate": 3.009933774834437e-05, "loss": 0.5772, "step": 11070 }, { "epoch": 18.04, "learning_rate": 2.985099337748344e-05, "loss": 0.3719, "step": 11080 }, { "epoch": 18.06, "learning_rate": 2.9602649006622514e-05, "loss": 0.2586, "step": 11090 }, { "epoch": 18.08, "learning_rate": 2.9354304635761587e-05, "loss": 0.2373, "step": 11100 }, { "epoch": 18.08, "eval_loss": 0.16864430904388428, "eval_runtime": 1045.0461, "eval_samples_per_second": 11.109, "eval_steps_per_second": 0.695, "eval_wer": 0.33723714820052, "step": 11100 }, { "epoch": 18.09, "learning_rate": 2.910596026490066e-05, "loss": 0.1572, "step": 11110 }, { "epoch": 18.11, "learning_rate": 2.8857615894039733e-05, "loss": 0.146, "step": 11120 }, { "epoch": 18.13, "learning_rate": 2.860927152317881e-05, "loss": 0.1208, "step": 11130 }, { "epoch": 18.14, "learning_rate": 2.8360927152317878e-05, "loss": 0.129, "step": 11140 }, { "epoch": 18.16, "learning_rate": 2.811258278145695e-05, "loss": 0.1562, "step": 11150 }, { "epoch": 18.17, "learning_rate": 2.7864238410596023e-05, "loss": 0.1154, "step": 11160 }, { "epoch": 18.19, "learning_rate": 2.76158940397351e-05, "loss": 0.1085, "step": 11170 }, { "epoch": 18.21, "learning_rate": 2.736754966887417e-05, "loss": 0.1035, "step": 11180 }, { "epoch": 18.22, "learning_rate": 2.711920529801324e-05, "loss": 0.1074, "step": 11190 }, { "epoch": 18.24, "learning_rate": 2.6870860927152318e-05, "loss": 0.1137, "step": 11200 }, { "epoch": 18.24, "eval_loss": 0.12019266188144684, "eval_runtime": 1065.1164, "eval_samples_per_second": 10.899, "eval_steps_per_second": 0.682, "eval_wer": 0.29376682023445694, "step": 11200 }, { "epoch": 18.26, "learning_rate": 2.662251655629139e-05, "loss": 0.112, "step": 11210 }, { "epoch": 18.27, "learning_rate": 2.637417218543046e-05, "loss": 0.0884, "step": 11220 }, { "epoch": 18.29, "learning_rate": 2.6125827814569536e-05, "loss": 0.0971, "step": 11230 }, { "epoch": 18.31, "learning_rate": 2.587748344370861e-05, "loss": 0.0817, "step": 11240 }, { "epoch": 18.32, "learning_rate": 2.5629139072847678e-05, "loss": 0.1116, "step": 11250 }, { "epoch": 18.34, "learning_rate": 2.538079470198675e-05, "loss": 0.1028, "step": 11260 }, { "epoch": 18.35, "learning_rate": 2.5132450331125827e-05, "loss": 0.0838, "step": 11270 }, { "epoch": 18.37, "learning_rate": 2.48841059602649e-05, "loss": 0.0956, "step": 11280 }, { "epoch": 18.39, "learning_rate": 2.463576158940397e-05, "loss": 0.1153, "step": 11290 }, { "epoch": 18.4, "learning_rate": 2.4387417218543045e-05, "loss": 0.1008, "step": 11300 }, { "epoch": 18.4, "eval_loss": 0.11432518064975739, "eval_runtime": 1073.4142, "eval_samples_per_second": 10.815, "eval_steps_per_second": 0.676, "eval_wer": 0.2895132965378826, "step": 11300 }, { "epoch": 18.42, "learning_rate": 2.4139072847682117e-05, "loss": 0.1066, "step": 11310 }, { "epoch": 18.44, "learning_rate": 2.389072847682119e-05, "loss": 0.0792, "step": 11320 }, { "epoch": 18.45, "learning_rate": 2.3642384105960263e-05, "loss": 0.0937, "step": 11330 }, { "epoch": 18.47, "learning_rate": 2.3394039735099336e-05, "loss": 0.0853, "step": 11340 }, { "epoch": 18.48, "learning_rate": 2.3145695364238408e-05, "loss": 0.1179, "step": 11350 }, { "epoch": 18.5, "learning_rate": 2.2897350993377484e-05, "loss": 0.1024, "step": 11360 }, { "epoch": 18.52, "learning_rate": 2.2649006622516554e-05, "loss": 0.0802, "step": 11370 }, { "epoch": 18.53, "learning_rate": 2.2400662251655626e-05, "loss": 0.0818, "step": 11380 }, { "epoch": 18.55, "learning_rate": 2.21523178807947e-05, "loss": 0.0862, "step": 11390 }, { "epoch": 18.57, "learning_rate": 2.1903973509933775e-05, "loss": 0.1004, "step": 11400 }, { "epoch": 18.57, "eval_loss": 0.1126680076122284, "eval_runtime": 1063.6172, "eval_samples_per_second": 10.915, "eval_steps_per_second": 0.683, "eval_wer": 0.28743785065912514, "step": 11400 }, { "epoch": 18.58, "learning_rate": 2.1655629139072845e-05, "loss": 0.1031, "step": 11410 }, { "epoch": 18.6, "learning_rate": 2.1407284768211917e-05, "loss": 0.0796, "step": 11420 }, { "epoch": 18.61, "learning_rate": 2.1158940397350993e-05, "loss": 0.0751, "step": 11430 }, { "epoch": 18.63, "learning_rate": 2.0910596026490066e-05, "loss": 0.0799, "step": 11440 }, { "epoch": 18.65, "learning_rate": 2.0662251655629135e-05, "loss": 0.0893, "step": 11450 }, { "epoch": 18.66, "learning_rate": 2.041390728476821e-05, "loss": 0.0856, "step": 11460 }, { "epoch": 18.68, "learning_rate": 2.0165562913907284e-05, "loss": 0.0864, "step": 11470 }, { "epoch": 18.7, "learning_rate": 1.9917218543046354e-05, "loss": 0.0892, "step": 11480 }, { "epoch": 18.71, "learning_rate": 1.9668874172185426e-05, "loss": 0.0887, "step": 11490 }, { "epoch": 18.73, "learning_rate": 1.9420529801324502e-05, "loss": 0.0874, "step": 11500 }, { "epoch": 18.73, "eval_loss": 0.11079666018486023, "eval_runtime": 1041.9643, "eval_samples_per_second": 11.141, "eval_steps_per_second": 0.697, "eval_wer": 0.28611503900013685, "step": 11500 }, { "epoch": 18.74, "learning_rate": 1.9172185430463575e-05, "loss": 0.0881, "step": 11510 }, { "epoch": 18.76, "learning_rate": 1.8923841059602644e-05, "loss": 0.0699, "step": 11520 }, { "epoch": 18.78, "learning_rate": 1.867549668874172e-05, "loss": 0.0706, "step": 11530 }, { "epoch": 18.79, "learning_rate": 1.8427152317880793e-05, "loss": 0.0796, "step": 11540 }, { "epoch": 18.81, "learning_rate": 1.8178807947019866e-05, "loss": 0.0873, "step": 11550 }, { "epoch": 18.83, "learning_rate": 1.793046357615894e-05, "loss": 0.0909, "step": 11560 }, { "epoch": 18.84, "learning_rate": 1.768211920529801e-05, "loss": 0.0757, "step": 11570 }, { "epoch": 18.86, "learning_rate": 1.7433774834437084e-05, "loss": 0.0848, "step": 11580 }, { "epoch": 18.87, "learning_rate": 1.7185430463576157e-05, "loss": 0.0883, "step": 11590 }, { "epoch": 18.89, "learning_rate": 1.693708609271523e-05, "loss": 0.0926, "step": 11600 }, { "epoch": 18.89, "eval_loss": 0.11082259565591812, "eval_runtime": 1057.6926, "eval_samples_per_second": 10.976, "eval_steps_per_second": 0.686, "eval_wer": 0.2837773114993386, "step": 11600 }, { "epoch": 18.91, "learning_rate": 1.6688741721854302e-05, "loss": 0.0818, "step": 11610 }, { "epoch": 18.92, "learning_rate": 1.6440397350993375e-05, "loss": 0.0755, "step": 11620 }, { "epoch": 18.94, "learning_rate": 1.6192052980132448e-05, "loss": 0.0729, "step": 11630 }, { "epoch": 18.96, "learning_rate": 1.594370860927152e-05, "loss": 0.089, "step": 11640 }, { "epoch": 18.97, "learning_rate": 1.5695364238410596e-05, "loss": 0.0851, "step": 11650 }, { "epoch": 18.99, "learning_rate": 1.5447019867549666e-05, "loss": 0.0825, "step": 11660 }, { "epoch": 19.0, "learning_rate": 1.519867549668874e-05, "loss": 0.0837, "step": 11670 }, { "epoch": 19.02, "learning_rate": 1.4950331125827815e-05, "loss": 0.087, "step": 11680 }, { "epoch": 19.04, "learning_rate": 1.4701986754966886e-05, "loss": 0.0712, "step": 11690 }, { "epoch": 19.05, "learning_rate": 1.445364238410596e-05, "loss": 0.0703, "step": 11700 }, { "epoch": 19.05, "eval_loss": 0.11007437109947205, "eval_runtime": 1067.8676, "eval_samples_per_second": 10.871, "eval_steps_per_second": 0.68, "eval_wer": 0.28337818729188524, "step": 11700 }, { "epoch": 19.07, "learning_rate": 1.4205298013245031e-05, "loss": 0.0777, "step": 11710 }, { "epoch": 19.09, "learning_rate": 1.3956953642384105e-05, "loss": 0.0876, "step": 11720 }, { "epoch": 19.1, "learning_rate": 1.3708609271523178e-05, "loss": 0.0782, "step": 11730 }, { "epoch": 19.12, "learning_rate": 1.3460264900662251e-05, "loss": 0.0983, "step": 11740 }, { "epoch": 19.14, "learning_rate": 1.3211920529801324e-05, "loss": 0.0696, "step": 11750 }, { "epoch": 19.15, "learning_rate": 1.2963576158940395e-05, "loss": 0.0708, "step": 11760 }, { "epoch": 19.17, "learning_rate": 1.2715231788079469e-05, "loss": 0.0975, "step": 11770 }, { "epoch": 19.18, "learning_rate": 1.2466887417218543e-05, "loss": 0.0732, "step": 11780 }, { "epoch": 19.2, "learning_rate": 1.2218543046357614e-05, "loss": 0.0666, "step": 11790 }, { "epoch": 19.22, "learning_rate": 1.1970198675496689e-05, "loss": 0.0893, "step": 11800 }, { "epoch": 19.22, "eval_loss": 0.10971127450466156, "eval_runtime": 1068.6941, "eval_samples_per_second": 10.863, "eval_steps_per_second": 0.679, "eval_wer": 0.2823632714500753, "step": 11800 } ], "logging_steps": 10, "max_steps": 12280, "num_train_epochs": 20, "save_steps": 200, "total_flos": 1.2371923109384754e+20, "trial_name": null, "trial_params": null }