{ "best_metric": null, "best_model_checkpoint": null, "epoch": 7.816110659072416, "eval_steps": 100, "global_step": 4800, "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 } ], "logging_steps": 10, "max_steps": 9210, "num_train_epochs": 15, "save_steps": 200, "total_flos": 5.031639452063693e+19, "trial_name": null, "trial_params": null }