| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 39.96271826333176, | |
| "eval_steps": 100.0, | |
| "global_step": 42360, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.4719207173194903, | |
| "grad_norm": 60.370758056640625, | |
| "learning_rate": 1.188e-06, | |
| "loss": 28.5881, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.9438414346389806, | |
| "grad_norm": 47.990840911865234, | |
| "learning_rate": 2.3880000000000003e-06, | |
| "loss": 11.3302, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_cer": 1.1283783783783783, | |
| "eval_loss": 10.027889251708984, | |
| "eval_runtime": 143.8913, | |
| "eval_samples_per_second": 47.536, | |
| "eval_steps_per_second": 5.942, | |
| "eval_wer": 1.0, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 1.4152902312411515, | |
| "grad_norm": 40.47557067871094, | |
| "learning_rate": 3.588e-06, | |
| "loss": 9.2112, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.8872109485606419, | |
| "grad_norm": 25.79726219177246, | |
| "learning_rate": 4.788e-06, | |
| "loss": 7.0068, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_cer": 1.1283487434803223, | |
| "eval_loss": 5.3823137283325195, | |
| "eval_runtime": 138.9276, | |
| "eval_samples_per_second": 49.234, | |
| "eval_steps_per_second": 6.154, | |
| "eval_wer": 1.0, | |
| "step": 2120 | |
| }, | |
| { | |
| "epoch": 2.3586597451628126, | |
| "grad_norm": 10.715571403503418, | |
| "learning_rate": 5.988e-06, | |
| "loss": 4.9098, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.830580462482303, | |
| "grad_norm": 2.4069929122924805, | |
| "learning_rate": 7.1880000000000005e-06, | |
| "loss": 3.682, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_cer": 1.1283487434803223, | |
| "eval_loss": 3.189443349838257, | |
| "eval_runtime": 133.8287, | |
| "eval_samples_per_second": 51.11, | |
| "eval_steps_per_second": 6.389, | |
| "eval_wer": 1.0, | |
| "step": 3180 | |
| }, | |
| { | |
| "epoch": 3.302029259084474, | |
| "grad_norm": 1.6513175964355469, | |
| "learning_rate": 8.388e-06, | |
| "loss": 3.1332, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 3.773949976403964, | |
| "grad_norm": 1.9935938119888306, | |
| "learning_rate": 9.588e-06, | |
| "loss": 2.725, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_cer": 1.1283783783783783, | |
| "eval_loss": 2.161285400390625, | |
| "eval_runtime": 133.493, | |
| "eval_samples_per_second": 51.239, | |
| "eval_steps_per_second": 6.405, | |
| "eval_wer": 1.0, | |
| "step": 4240 | |
| }, | |
| { | |
| "epoch": 4.245398773006135, | |
| "grad_norm": 2.0250744819641113, | |
| "learning_rate": 1.0787999999999999e-05, | |
| "loss": 2.2125, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 4.717319490325625, | |
| "grad_norm": 6.084537506103516, | |
| "learning_rate": 1.1988000000000001e-05, | |
| "loss": 1.4317, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 1.1232219061166429, | |
| "eval_loss": 0.7817353010177612, | |
| "eval_runtime": 140.0025, | |
| "eval_samples_per_second": 48.856, | |
| "eval_steps_per_second": 6.107, | |
| "eval_wer": 1.0, | |
| "step": 5300 | |
| }, | |
| { | |
| "epoch": 5.188768286927796, | |
| "grad_norm": 8.009575843811035, | |
| "learning_rate": 1.3188e-05, | |
| "loss": 0.8716, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 5.660689004247287, | |
| "grad_norm": 6.703843116760254, | |
| "learning_rate": 1.4388000000000002e-05, | |
| "loss": 0.5734, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_cer": 1.0337541488857278, | |
| "eval_loss": 0.2728799879550934, | |
| "eval_runtime": 137.2562, | |
| "eval_samples_per_second": 49.834, | |
| "eval_steps_per_second": 6.229, | |
| "eval_wer": 0.9991228070175439, | |
| "step": 6360 | |
| }, | |
| { | |
| "epoch": 6.132137800849457, | |
| "grad_norm": 6.002254009246826, | |
| "learning_rate": 1.5588e-05, | |
| "loss": 0.4288, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 6.604058518168948, | |
| "grad_norm": 8.313450813293457, | |
| "learning_rate": 1.6788e-05, | |
| "loss": 0.3485, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_cer": 1.0258712660028448, | |
| "eval_loss": 0.16624058783054352, | |
| "eval_runtime": 68.7385, | |
| "eval_samples_per_second": 99.508, | |
| "eval_steps_per_second": 12.438, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 7420 | |
| }, | |
| { | |
| "epoch": 7.0755073147711185, | |
| "grad_norm": 2.5701239109039307, | |
| "learning_rate": 1.7988e-05, | |
| "loss": 0.2995, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 7.547428032090608, | |
| "grad_norm": 5.836677551269531, | |
| "learning_rate": 1.9188e-05, | |
| "loss": 0.2631, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_cer": 1.0223743480322427, | |
| "eval_loss": 0.083324134349823, | |
| "eval_runtime": 70.5926, | |
| "eval_samples_per_second": 96.894, | |
| "eval_steps_per_second": 12.112, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 8480 | |
| }, | |
| { | |
| "epoch": 8.018876828692779, | |
| "grad_norm": 5.063877105712891, | |
| "learning_rate": 2.0388e-05, | |
| "loss": 0.2388, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 8.49079754601227, | |
| "grad_norm": 3.660646677017212, | |
| "learning_rate": 2.1588e-05, | |
| "loss": 0.218, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 8.96271826333176, | |
| "grad_norm": 6.175236225128174, | |
| "learning_rate": 2.2788000000000003e-05, | |
| "loss": 0.2068, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_cer": 1.0209518729255571, | |
| "eval_loss": 0.06159723177552223, | |
| "eval_runtime": 68.413, | |
| "eval_samples_per_second": 99.981, | |
| "eval_steps_per_second": 12.498, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 9540 | |
| }, | |
| { | |
| "epoch": 9.43416705993393, | |
| "grad_norm": 0.2773754894733429, | |
| "learning_rate": 2.3988e-05, | |
| "loss": 0.2023, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 9.906087777253422, | |
| "grad_norm": 1.5312055349349976, | |
| "learning_rate": 2.5188e-05, | |
| "loss": 0.1846, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_cer": 1.0198850165955429, | |
| "eval_loss": 0.07861880213022232, | |
| "eval_runtime": 59.5147, | |
| "eval_samples_per_second": 114.929, | |
| "eval_steps_per_second": 14.366, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 10600 | |
| }, | |
| { | |
| "epoch": 10.377536573855592, | |
| "grad_norm": 6.2259297370910645, | |
| "learning_rate": 2.6388000000000002e-05, | |
| "loss": 0.1781, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 10.849457291175083, | |
| "grad_norm": 5.571898460388184, | |
| "learning_rate": 2.7585600000000002e-05, | |
| "loss": 0.1792, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_cer": 1.019825746799431, | |
| "eval_loss": 0.0471838042140007, | |
| "eval_runtime": 65.0557, | |
| "eval_samples_per_second": 105.141, | |
| "eval_steps_per_second": 13.143, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 11660 | |
| }, | |
| { | |
| "epoch": 11.320906087777253, | |
| "grad_norm": 3.878492832183838, | |
| "learning_rate": 2.87856e-05, | |
| "loss": 0.1662, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 11.792826805096745, | |
| "grad_norm": 6.27423095703125, | |
| "learning_rate": 2.99856e-05, | |
| "loss": 0.1669, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_cer": 1.0206555239449977, | |
| "eval_loss": 0.05153690651059151, | |
| "eval_runtime": 65.9815, | |
| "eval_samples_per_second": 103.665, | |
| "eval_steps_per_second": 12.958, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 12720 | |
| }, | |
| { | |
| "epoch": 12.264275601698914, | |
| "grad_norm": 3.428331136703491, | |
| "learning_rate": 2.9979744753439937e-05, | |
| "loss": 0.159, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 12.736196319018404, | |
| "grad_norm": 3.60664701461792, | |
| "learning_rate": 2.991821289408868e-05, | |
| "loss": 0.1609, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_cer": 1.0218705547652915, | |
| "eval_loss": 0.05291549116373062, | |
| "eval_runtime": 81.1439, | |
| "eval_samples_per_second": 84.295, | |
| "eval_steps_per_second": 10.537, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 13780 | |
| }, | |
| { | |
| "epoch": 13.207645115620576, | |
| "grad_norm": 3.0601279735565186, | |
| "learning_rate": 2.981532510892707e-05, | |
| "loss": 0.156, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 13.679565832940066, | |
| "grad_norm": 1.1705658435821533, | |
| "learning_rate": 2.9671447959906427e-05, | |
| "loss": 0.1508, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_cer": 1.0183143669985775, | |
| "eval_loss": 0.043218065053224564, | |
| "eval_runtime": 72.3333, | |
| "eval_samples_per_second": 94.562, | |
| "eval_steps_per_second": 11.82, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 14840 | |
| }, | |
| { | |
| "epoch": 14.151014629542237, | |
| "grad_norm": 3.3974997997283936, | |
| "learning_rate": 2.9486979510030355e-05, | |
| "loss": 0.1454, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 14.622935346861727, | |
| "grad_norm": 1.8044121265411377, | |
| "learning_rate": 2.9262430125717204e-05, | |
| "loss": 0.1427, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_cer": 1.0147878141299194, | |
| "eval_loss": 0.08599487692117691, | |
| "eval_runtime": 59.8056, | |
| "eval_samples_per_second": 114.37, | |
| "eval_steps_per_second": 14.296, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 15900 | |
| }, | |
| { | |
| "epoch": 15.094384143463898, | |
| "grad_norm": 2.6493678092956543, | |
| "learning_rate": 2.899842106477605e-05, | |
| "loss": 0.1398, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 15.566304860783388, | |
| "grad_norm": 4.24639892578125, | |
| "learning_rate": 2.869568275758094e-05, | |
| "loss": 0.1316, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_cer": 1.0185218112849692, | |
| "eval_loss": 0.0349772572517395, | |
| "eval_runtime": 60.1896, | |
| "eval_samples_per_second": 113.641, | |
| "eval_steps_per_second": 14.205, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 16960 | |
| }, | |
| { | |
| "epoch": 16.037753657385558, | |
| "grad_norm": 4.061686992645264, | |
| "learning_rate": 2.8355052786198856e-05, | |
| "loss": 0.1306, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 16.50967437470505, | |
| "grad_norm": 2.1003165245056152, | |
| "learning_rate": 2.7977473567062487e-05, | |
| "loss": 0.1248, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 16.98159509202454, | |
| "grad_norm": 3.3991761207580566, | |
| "learning_rate": 2.7563989743599158e-05, | |
| "loss": 0.1296, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_cer": 1.0190552394499763, | |
| "eval_loss": 0.04494578763842583, | |
| "eval_runtime": 86.2502, | |
| "eval_samples_per_second": 79.304, | |
| "eval_steps_per_second": 9.913, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 18020 | |
| }, | |
| { | |
| "epoch": 17.45304388862671, | |
| "grad_norm": 3.5397164821624756, | |
| "learning_rate": 2.7116675667417045e-05, | |
| "loss": 0.1239, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 17.9249646059462, | |
| "grad_norm": 3.1016619205474854, | |
| "learning_rate": 2.663497648875447e-05, | |
| "loss": 0.1236, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_cer": 1.0170400663821717, | |
| "eval_loss": 0.04502090439200401, | |
| "eval_runtime": 67.5469, | |
| "eval_samples_per_second": 101.263, | |
| "eval_steps_per_second": 12.658, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 19080 | |
| }, | |
| { | |
| "epoch": 18.396413402548372, | |
| "grad_norm": 1.3774545192718506, | |
| "learning_rate": 2.6121086974660007e-05, | |
| "loss": 0.1168, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 18.86833411986786, | |
| "grad_norm": 3.927860736846924, | |
| "learning_rate": 2.557642889644962e-05, | |
| "loss": 0.1149, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_cer": 1.018136557610242, | |
| "eval_loss": 0.03580431640148163, | |
| "eval_runtime": 62.0438, | |
| "eval_samples_per_second": 110.245, | |
| "eval_steps_per_second": 13.781, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 20140 | |
| }, | |
| { | |
| "epoch": 19.339782916470032, | |
| "grad_norm": 1.4613689184188843, | |
| "learning_rate": 2.500368515269978e-05, | |
| "loss": 0.1097, | |
| "step": 20500 | |
| }, | |
| { | |
| "epoch": 19.811703633789524, | |
| "grad_norm": 1.7840094566345215, | |
| "learning_rate": 2.440214531030174e-05, | |
| "loss": 0.1149, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_cer": 1.0176327643432908, | |
| "eval_loss": 0.0317009836435318, | |
| "eval_runtime": 64.2537, | |
| "eval_samples_per_second": 106.453, | |
| "eval_steps_per_second": 13.307, | |
| "eval_wer": 0.9989766081871345, | |
| "step": 21200 | |
| }, | |
| { | |
| "epoch": 20.283152430391695, | |
| "grad_norm": 9.317741394042969, | |
| "learning_rate": 2.3774592676935842e-05, | |
| "loss": 0.1096, | |
| "step": 21500 | |
| }, | |
| { | |
| "epoch": 20.755073147711183, | |
| "grad_norm": 2.2778842449188232, | |
| "learning_rate": 2.3122763494165503e-05, | |
| "loss": 0.106, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_cer": 1.0170104314841157, | |
| "eval_loss": 0.03687233105301857, | |
| "eval_runtime": 70.0875, | |
| "eval_samples_per_second": 97.592, | |
| "eval_steps_per_second": 12.199, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 22260 | |
| }, | |
| { | |
| "epoch": 21.226521944313355, | |
| "grad_norm": 5.864170551300049, | |
| "learning_rate": 2.2449830971856908e-05, | |
| "loss": 0.1042, | |
| "step": 22500 | |
| }, | |
| { | |
| "epoch": 21.698442661632846, | |
| "grad_norm": 1.5493875741958618, | |
| "learning_rate": 2.1754960402430945e-05, | |
| "loss": 0.102, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_cer": 1.018018018018018, | |
| "eval_loss": 0.034220367670059204, | |
| "eval_runtime": 60.8599, | |
| "eval_samples_per_second": 112.389, | |
| "eval_steps_per_second": 14.049, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 23320 | |
| }, | |
| { | |
| "epoch": 22.169891458235018, | |
| "grad_norm": 2.6867339611053467, | |
| "learning_rate": 2.1041400973479267e-05, | |
| "loss": 0.099, | |
| "step": 23500 | |
| }, | |
| { | |
| "epoch": 22.641812175554506, | |
| "grad_norm": 6.666200637817383, | |
| "learning_rate": 2.031112688042812e-05, | |
| "loss": 0.1011, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_cer": 1.0178994784257942, | |
| "eval_loss": 0.04105741158127785, | |
| "eval_runtime": 63.1164, | |
| "eval_samples_per_second": 108.371, | |
| "eval_steps_per_second": 13.546, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 24380 | |
| }, | |
| { | |
| "epoch": 23.113260972156677, | |
| "grad_norm": 2.3484959602355957, | |
| "learning_rate": 1.956766180067445e-05, | |
| "loss": 0.0991, | |
| "step": 24500 | |
| }, | |
| { | |
| "epoch": 23.58518168947617, | |
| "grad_norm": 3.396759510040283, | |
| "learning_rate": 1.881008353783706e-05, | |
| "loss": 0.0948, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_cer": 1.0163288288288288, | |
| "eval_loss": 0.031411658972501755, | |
| "eval_runtime": 60.6777, | |
| "eval_samples_per_second": 112.727, | |
| "eval_steps_per_second": 14.091, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 25440 | |
| }, | |
| { | |
| "epoch": 24.05663048607834, | |
| "grad_norm": 0.44371920824050903, | |
| "learning_rate": 1.8041963967312025e-05, | |
| "loss": 0.0963, | |
| "step": 25500 | |
| }, | |
| { | |
| "epoch": 24.52855120339783, | |
| "grad_norm": 2.3664116859436035, | |
| "learning_rate": 1.7265428235347026e-05, | |
| "loss": 0.0906, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "grad_norm": 2.171415090560913, | |
| "learning_rate": 1.6482624773072205e-05, | |
| "loss": 0.0934, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_cer": 1.017454954954955, | |
| "eval_loss": 0.030156882479786873, | |
| "eval_runtime": 61.4594, | |
| "eval_samples_per_second": 111.293, | |
| "eval_steps_per_second": 13.912, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 26500 | |
| }, | |
| { | |
| "epoch": 25.47192071731949, | |
| "grad_norm": 3.737889051437378, | |
| "learning_rate": 1.5697295811286228e-05, | |
| "loss": 0.0844, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 25.94384143463898, | |
| "grad_norm": 5.188897609710693, | |
| "learning_rate": 1.4908467225519327e-05, | |
| "loss": 0.0843, | |
| "step": 27500 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_cer": 1.0172475106685632, | |
| "eval_loss": 0.04401924088597298, | |
| "eval_runtime": 59.6879, | |
| "eval_samples_per_second": 114.596, | |
| "eval_steps_per_second": 14.325, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 27560 | |
| }, | |
| { | |
| "epoch": 26.41529023124115, | |
| "grad_norm": 3.3139688968658447, | |
| "learning_rate": 1.4119891882267897e-05, | |
| "loss": 0.0789, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 26.887210948560643, | |
| "grad_norm": 3.2081096172332764, | |
| "learning_rate": 1.3333751522495247e-05, | |
| "loss": 0.0833, | |
| "step": 28500 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_cer": 1.016743717401612, | |
| "eval_loss": 0.034057337790727615, | |
| "eval_runtime": 60.1277, | |
| "eval_samples_per_second": 113.758, | |
| "eval_steps_per_second": 14.22, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 28620 | |
| }, | |
| { | |
| "epoch": 27.35865974516281, | |
| "grad_norm": 1.3224974870681763, | |
| "learning_rate": 1.2552221150328132e-05, | |
| "loss": 0.0749, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 27.830580462482303, | |
| "grad_norm": 0.9007667303085327, | |
| "learning_rate": 1.1779004344640104e-05, | |
| "loss": 0.0781, | |
| "step": 29500 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_cer": 1.0157064959696538, | |
| "eval_loss": 0.05653638020157814, | |
| "eval_runtime": 63.0029, | |
| "eval_samples_per_second": 108.566, | |
| "eval_steps_per_second": 13.571, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 29680 | |
| }, | |
| { | |
| "epoch": 28.302029259084474, | |
| "grad_norm": 1.9431039094924927, | |
| "learning_rate": 1.101314200491722e-05, | |
| "loss": 0.072, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 28.773949976403966, | |
| "grad_norm": 0.08316487073898315, | |
| "learning_rate": 1.0258310052432775e-05, | |
| "loss": 0.0741, | |
| "step": 30500 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "eval_cer": 1.0160917496443813, | |
| "eval_loss": 0.03570393845438957, | |
| "eval_runtime": 61.5019, | |
| "eval_samples_per_second": 111.216, | |
| "eval_steps_per_second": 13.902, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 30740 | |
| }, | |
| { | |
| "epoch": 29.245398773006134, | |
| "grad_norm": 0.031967077404260635, | |
| "learning_rate": 9.516596870757504e-06, | |
| "loss": 0.0717, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 29.717319490325625, | |
| "grad_norm": 0.1671179085969925, | |
| "learning_rate": 8.79005454793263e-06, | |
| "loss": 0.0704, | |
| "step": 31500 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_cer": 1.016239924134661, | |
| "eval_loss": 0.030619077384471893, | |
| "eval_runtime": 59.7711, | |
| "eval_samples_per_second": 114.437, | |
| "eval_steps_per_second": 14.305, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 31800 | |
| }, | |
| { | |
| "epoch": 30.188768286927797, | |
| "grad_norm": 4.69877815246582, | |
| "learning_rate": 8.080693198985019e-06, | |
| "loss": 0.0723, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 30.660689004247285, | |
| "grad_norm": 1.8107829093933105, | |
| "learning_rate": 7.39319559220429e-06, | |
| "loss": 0.0678, | |
| "step": 32500 | |
| }, | |
| { | |
| "epoch": 31.0, | |
| "eval_cer": 1.0159435751541015, | |
| "eval_loss": 0.02890847437083721, | |
| "eval_runtime": 62.096, | |
| "eval_samples_per_second": 110.152, | |
| "eval_steps_per_second": 13.769, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 32860 | |
| }, | |
| { | |
| "epoch": 31.132137800849456, | |
| "grad_norm": 2.0504305362701416, | |
| "learning_rate": 6.7239430275857026e-06, | |
| "loss": 0.0628, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 31.604058518168948, | |
| "grad_norm": 3.536367654800415, | |
| "learning_rate": 6.077587720237289e-06, | |
| "loss": 0.067, | |
| "step": 33500 | |
| }, | |
| { | |
| "epoch": 32.0, | |
| "eval_cer": 1.0158843053579896, | |
| "eval_loss": 0.027865121141076088, | |
| "eval_runtime": 60.0098, | |
| "eval_samples_per_second": 113.981, | |
| "eval_steps_per_second": 14.248, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 33920 | |
| }, | |
| { | |
| "epoch": 32.075507314771116, | |
| "grad_norm": 0.30863332748413086, | |
| "learning_rate": 5.4559179328316745e-06, | |
| "loss": 0.0617, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 32.54742803209061, | |
| "grad_norm": 1.2579513788223267, | |
| "learning_rate": 4.860653630940242e-06, | |
| "loss": 0.0641, | |
| "step": 34500 | |
| }, | |
| { | |
| "epoch": 33.0, | |
| "eval_cer": 1.0159732100521575, | |
| "eval_loss": 0.032531462609767914, | |
| "eval_runtime": 61.6714, | |
| "eval_samples_per_second": 110.91, | |
| "eval_steps_per_second": 13.864, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 34980 | |
| }, | |
| { | |
| "epoch": 33.01887682869278, | |
| "grad_norm": 1.1170421838760376, | |
| "learning_rate": 4.293441724427136e-06, | |
| "loss": 0.0582, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 33.49079754601227, | |
| "grad_norm": 2.8375914096832275, | |
| "learning_rate": 3.7558515109655663e-06, | |
| "loss": 0.0609, | |
| "step": 35500 | |
| }, | |
| { | |
| "epoch": 33.96271826333176, | |
| "grad_norm": 0.19227160513401031, | |
| "learning_rate": 3.249370334282695e-06, | |
| "loss": 0.0595, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 34.0, | |
| "eval_cer": 1.0166251778093884, | |
| "eval_loss": 0.033033788204193115, | |
| "eval_runtime": 59.4597, | |
| "eval_samples_per_second": 115.036, | |
| "eval_steps_per_second": 14.379, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 36040 | |
| }, | |
| { | |
| "epoch": 34.434167059933934, | |
| "grad_norm": 1.7270156145095825, | |
| "learning_rate": 2.775399469145443e-06, | |
| "loss": 0.0614, | |
| "step": 36500 | |
| }, | |
| { | |
| "epoch": 34.90608777725342, | |
| "grad_norm": 0.2758292257785797, | |
| "learning_rate": 2.336095963596342e-06, | |
| "loss": 0.055, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 35.0, | |
| "eval_cer": 1.0164177335229967, | |
| "eval_loss": 0.030884480103850365, | |
| "eval_runtime": 58.7794, | |
| "eval_samples_per_second": 116.367, | |
| "eval_steps_per_second": 14.546, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 37100 | |
| }, | |
| { | |
| "epoch": 35.37753657385559, | |
| "grad_norm": 2.685821771621704, | |
| "learning_rate": 1.9309149042102644e-06, | |
| "loss": 0.0564, | |
| "step": 37500 | |
| }, | |
| { | |
| "epoch": 35.849457291175085, | |
| "grad_norm": 2.9631125926971436, | |
| "learning_rate": 1.5618919095549582e-06, | |
| "loss": 0.055, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 36.0, | |
| "eval_cer": 1.016714082503556, | |
| "eval_loss": 0.03032144159078598, | |
| "eval_runtime": 71.6691, | |
| "eval_samples_per_second": 95.439, | |
| "eval_steps_per_second": 11.93, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 38160 | |
| }, | |
| { | |
| "epoch": 36.32090608777725, | |
| "grad_norm": 0.7306509613990784, | |
| "learning_rate": 1.2300479506647383e-06, | |
| "loss": 0.0554, | |
| "step": 38500 | |
| }, | |
| { | |
| "epoch": 36.792826805096745, | |
| "grad_norm": 1.8589800596237183, | |
| "learning_rate": 9.363011358278406e-07, | |
| "loss": 0.0555, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 37.0, | |
| "eval_cer": 1.0165955429113325, | |
| "eval_loss": 0.03112892434000969, | |
| "eval_runtime": 70.4338, | |
| "eval_samples_per_second": 97.112, | |
| "eval_steps_per_second": 12.139, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 39220 | |
| }, | |
| { | |
| "epoch": 37.26427560169891, | |
| "grad_norm": 1.8065155744552612, | |
| "learning_rate": 6.814641704683633e-07, | |
| "loss": 0.0568, | |
| "step": 39500 | |
| }, | |
| { | |
| "epoch": 37.736196319018404, | |
| "grad_norm": 1.0842024087905884, | |
| "learning_rate": 4.6663260289360743e-07, | |
| "loss": 0.0533, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 38.0, | |
| "eval_cer": 1.0165659080132765, | |
| "eval_loss": 0.031128019094467163, | |
| "eval_runtime": 62.073, | |
| "eval_samples_per_second": 110.193, | |
| "eval_steps_per_second": 13.774, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 40280 | |
| }, | |
| { | |
| "epoch": 38.20764511562058, | |
| "grad_norm": 1.8371340036392212, | |
| "learning_rate": 2.915399553636205e-07, | |
| "loss": 0.0571, | |
| "step": 40500 | |
| }, | |
| { | |
| "epoch": 38.679565832940064, | |
| "grad_norm": 1.5724213123321533, | |
| "learning_rate": 1.5714100953393373e-07, | |
| "loss": 0.0528, | |
| "step": 41000 | |
| }, | |
| { | |
| "epoch": 39.0, | |
| "eval_cer": 1.0165659080132765, | |
| "eval_loss": 0.031000742688775063, | |
| "eval_runtime": 61.7575, | |
| "eval_samples_per_second": 110.756, | |
| "eval_steps_per_second": 13.844, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 41340 | |
| }, | |
| { | |
| "epoch": 39.15101462954224, | |
| "grad_norm": 1.2510491609573364, | |
| "learning_rate": 6.380760518884532e-08, | |
| "loss": 0.0568, | |
| "step": 41500 | |
| }, | |
| { | |
| "epoch": 39.62293534686173, | |
| "grad_norm": 0.06529413163661957, | |
| "learning_rate": 1.1860630053380229e-08, | |
| "loss": 0.0509, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 39.96271826333176, | |
| "eval_cer": 1.0165659080132765, | |
| "eval_loss": 0.03096253238618374, | |
| "eval_runtime": 109.6526, | |
| "eval_samples_per_second": 62.379, | |
| "eval_steps_per_second": 7.797, | |
| "eval_wer": 0.9988304093567252, | |
| "step": 42360 | |
| }, | |
| { | |
| "epoch": 39.96271826333176, | |
| "step": 42360, | |
| "total_flos": 2.2101401353322127e+19, | |
| "train_loss": 0.9936401144753547, | |
| "train_runtime": 36758.8562, | |
| "train_samples_per_second": 73.778, | |
| "train_steps_per_second": 1.152 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 42360, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 40, | |
| "save_steps": 400, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.2101401353322127e+19, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |