| { | |
| "best_metric": 78.86568308105001, | |
| "best_model_checkpoint": "./whisper-small-ha-adam-v4/checkpoint-2000", | |
| "epoch": 15.0, | |
| "eval_steps": 500, | |
| "global_step": 2355, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1592356687898089, | |
| "grad_norm": 21.07309341430664, | |
| "learning_rate": 2.5e-05, | |
| "loss": 4.4409, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.3184713375796178, | |
| "grad_norm": 20.01163673400879, | |
| "learning_rate": 5e-05, | |
| "loss": 2.732, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.47770700636942676, | |
| "grad_norm": 20.838537216186523, | |
| "learning_rate": 5e-05, | |
| "loss": 2.1765, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.6369426751592356, | |
| "grad_norm": 13.77207088470459, | |
| "learning_rate": 5e-05, | |
| "loss": 1.908, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.7961783439490446, | |
| "grad_norm": 13.532313346862793, | |
| "learning_rate": 5e-05, | |
| "loss": 1.745, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.9554140127388535, | |
| "grad_norm": 13.651493072509766, | |
| "learning_rate": 5e-05, | |
| "loss": 1.6779, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.1146496815286624, | |
| "grad_norm": 9.819719314575195, | |
| "learning_rate": 5e-05, | |
| "loss": 1.0426, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.2738853503184713, | |
| "grad_norm": 9.96407699584961, | |
| "learning_rate": 5e-05, | |
| "loss": 0.7765, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.4331210191082802, | |
| "grad_norm": 11.112582206726074, | |
| "learning_rate": 5e-05, | |
| "loss": 0.7978, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.5923566878980893, | |
| "grad_norm": 10.045536041259766, | |
| "learning_rate": 5e-05, | |
| "loss": 0.8032, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.7515923566878981, | |
| "grad_norm": 11.228167533874512, | |
| "learning_rate": 5e-05, | |
| "loss": 0.7916, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 1.910828025477707, | |
| "grad_norm": 10.779911994934082, | |
| "learning_rate": 5e-05, | |
| "loss": 0.8241, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.070063694267516, | |
| "grad_norm": 6.224853038787842, | |
| "learning_rate": 5e-05, | |
| "loss": 0.5693, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 2.229299363057325, | |
| "grad_norm": 6.078026294708252, | |
| "learning_rate": 5e-05, | |
| "loss": 0.2365, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.388535031847134, | |
| "grad_norm": 5.989273548126221, | |
| "learning_rate": 5e-05, | |
| "loss": 0.2271, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 2.5477707006369426, | |
| "grad_norm": 8.927431106567383, | |
| "learning_rate": 5e-05, | |
| "loss": 0.2478, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.7070063694267517, | |
| "grad_norm": 7.507767200469971, | |
| "learning_rate": 5e-05, | |
| "loss": 0.2846, | |
| "step": 425 | |
| }, | |
| { | |
| "epoch": 2.8662420382165603, | |
| "grad_norm": 6.9506330490112305, | |
| "learning_rate": 5e-05, | |
| "loss": 0.2722, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 3.0254777070063694, | |
| "grad_norm": 4.77834415435791, | |
| "learning_rate": 5e-05, | |
| "loss": 0.2588, | |
| "step": 475 | |
| }, | |
| { | |
| "epoch": 3.1847133757961785, | |
| "grad_norm": 4.616095066070557, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0995, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.1847133757961785, | |
| "eval_loss": 1.7910184860229492, | |
| "eval_runtime": 264.8902, | |
| "eval_samples_per_second": 2.492, | |
| "eval_steps_per_second": 0.159, | |
| "eval_wer": 88.17781184134891, | |
| "eval_wer_ortho": 90.4296875, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.343949044585987, | |
| "grad_norm": 4.324921607971191, | |
| "learning_rate": 5e-05, | |
| "loss": 0.1136, | |
| "step": 525 | |
| }, | |
| { | |
| "epoch": 3.5031847133757963, | |
| "grad_norm": 3.88926100730896, | |
| "learning_rate": 5e-05, | |
| "loss": 0.117, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 3.662420382165605, | |
| "grad_norm": 4.270689964294434, | |
| "learning_rate": 5e-05, | |
| "loss": 0.1093, | |
| "step": 575 | |
| }, | |
| { | |
| "epoch": 3.821656050955414, | |
| "grad_norm": 5.100738048553467, | |
| "learning_rate": 5e-05, | |
| "loss": 0.1185, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 3.980891719745223, | |
| "grad_norm": 5.05330753326416, | |
| "learning_rate": 5e-05, | |
| "loss": 0.1377, | |
| "step": 625 | |
| }, | |
| { | |
| "epoch": 4.140127388535032, | |
| "grad_norm": 3.5940771102905273, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0823, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 4.2993630573248405, | |
| "grad_norm": 3.304124355316162, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0763, | |
| "step": 675 | |
| }, | |
| { | |
| "epoch": 4.45859872611465, | |
| "grad_norm": 3.3643264770507812, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0758, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 4.617834394904459, | |
| "grad_norm": 3.0554709434509277, | |
| "learning_rate": 5e-05, | |
| "loss": 0.084, | |
| "step": 725 | |
| }, | |
| { | |
| "epoch": 4.777070063694268, | |
| "grad_norm": 5.5209174156188965, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0816, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 4.936305732484076, | |
| "grad_norm": 5.73225736618042, | |
| "learning_rate": 5e-05, | |
| "loss": 0.1064, | |
| "step": 775 | |
| }, | |
| { | |
| "epoch": 5.095541401273885, | |
| "grad_norm": 2.312955379486084, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0697, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 5.254777070063694, | |
| "grad_norm": 5.729402542114258, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0639, | |
| "step": 825 | |
| }, | |
| { | |
| "epoch": 5.414012738853503, | |
| "grad_norm": 3.784353733062744, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0733, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 5.573248407643312, | |
| "grad_norm": 2.1651360988616943, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0774, | |
| "step": 875 | |
| }, | |
| { | |
| "epoch": 5.732484076433121, | |
| "grad_norm": 4.008174419403076, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0653, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 5.89171974522293, | |
| "grad_norm": 4.1395978927612305, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0892, | |
| "step": 925 | |
| }, | |
| { | |
| "epoch": 6.050955414012739, | |
| "grad_norm": 3.262603998184204, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0573, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 6.210191082802548, | |
| "grad_norm": 2.6815080642700195, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0646, | |
| "step": 975 | |
| }, | |
| { | |
| "epoch": 6.369426751592357, | |
| "grad_norm": 1.4305006265640259, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0468, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 6.369426751592357, | |
| "eval_loss": 1.959425926208496, | |
| "eval_runtime": 259.4756, | |
| "eval_samples_per_second": 2.544, | |
| "eval_steps_per_second": 0.162, | |
| "eval_wer": 81.14581337420962, | |
| "eval_wer_ortho": 82.83203125, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 6.528662420382165, | |
| "grad_norm": 6.745048522949219, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0515, | |
| "step": 1025 | |
| }, | |
| { | |
| "epoch": 6.687898089171974, | |
| "grad_norm": 2.1589980125427246, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0507, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 6.8471337579617835, | |
| "grad_norm": 3.8626465797424316, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0456, | |
| "step": 1075 | |
| }, | |
| { | |
| "epoch": 7.006369426751593, | |
| "grad_norm": 2.852226734161377, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0507, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 7.165605095541402, | |
| "grad_norm": 1.2226496934890747, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0606, | |
| "step": 1125 | |
| }, | |
| { | |
| "epoch": 7.32484076433121, | |
| "grad_norm": 2.6874589920043945, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0488, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 7.484076433121019, | |
| "grad_norm": 0.9016306400299072, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0526, | |
| "step": 1175 | |
| }, | |
| { | |
| "epoch": 7.643312101910828, | |
| "grad_norm": 2.1149990558624268, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0484, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 7.802547770700637, | |
| "grad_norm": 2.9689130783081055, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0466, | |
| "step": 1225 | |
| }, | |
| { | |
| "epoch": 7.961783439490446, | |
| "grad_norm": 2.814160108566284, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0443, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 8.121019108280255, | |
| "grad_norm": 4.792329788208008, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0374, | |
| "step": 1275 | |
| }, | |
| { | |
| "epoch": 8.280254777070065, | |
| "grad_norm": 3.2325730323791504, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0347, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 8.439490445859873, | |
| "grad_norm": 3.5987560749053955, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0421, | |
| "step": 1325 | |
| }, | |
| { | |
| "epoch": 8.598726114649681, | |
| "grad_norm": 2.2664973735809326, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0337, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 8.757961783439491, | |
| "grad_norm": 2.501434326171875, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0363, | |
| "step": 1375 | |
| }, | |
| { | |
| "epoch": 8.9171974522293, | |
| "grad_norm": 3.5731797218322754, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0511, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 9.07643312101911, | |
| "grad_norm": 2.30594539642334, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0535, | |
| "step": 1425 | |
| }, | |
| { | |
| "epoch": 9.235668789808917, | |
| "grad_norm": 1.6527944803237915, | |
| "learning_rate": 5e-05, | |
| "loss": 0.04, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 9.394904458598726, | |
| "grad_norm": 2.126922130584717, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0299, | |
| "step": 1475 | |
| }, | |
| { | |
| "epoch": 9.554140127388536, | |
| "grad_norm": 2.613511323928833, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0394, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 9.554140127388536, | |
| "eval_loss": 2.0775890350341797, | |
| "eval_runtime": 268.0164, | |
| "eval_samples_per_second": 2.463, | |
| "eval_steps_per_second": 0.157, | |
| "eval_wer": 87.75627514849587, | |
| "eval_wer_ortho": 89.84375, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 9.713375796178344, | |
| "grad_norm": 2.170787811279297, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0327, | |
| "step": 1525 | |
| }, | |
| { | |
| "epoch": 9.872611464968152, | |
| "grad_norm": 2.1923575401306152, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0472, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 10.031847133757962, | |
| "grad_norm": 4.220789909362793, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0463, | |
| "step": 1575 | |
| }, | |
| { | |
| "epoch": 10.19108280254777, | |
| "grad_norm": 1.4491336345672607, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0447, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 10.35031847133758, | |
| "grad_norm": 3.3499913215637207, | |
| "learning_rate": 5e-05, | |
| "loss": 0.043, | |
| "step": 1625 | |
| }, | |
| { | |
| "epoch": 10.509554140127388, | |
| "grad_norm": 2.196830987930298, | |
| "learning_rate": 5e-05, | |
| "loss": 0.035, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 10.668789808917197, | |
| "grad_norm": 2.2914416790008545, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0281, | |
| "step": 1675 | |
| }, | |
| { | |
| "epoch": 10.828025477707007, | |
| "grad_norm": 2.437507152557373, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0431, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 10.987261146496815, | |
| "grad_norm": 5.599733352661133, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0434, | |
| "step": 1725 | |
| }, | |
| { | |
| "epoch": 11.146496815286625, | |
| "grad_norm": 3.811133861541748, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0287, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 11.305732484076433, | |
| "grad_norm": 3.298198938369751, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0477, | |
| "step": 1775 | |
| }, | |
| { | |
| "epoch": 11.464968152866241, | |
| "grad_norm": 1.9741543531417847, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0423, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 11.624203821656051, | |
| "grad_norm": 1.3877679109573364, | |
| "learning_rate": 5e-05, | |
| "loss": 0.033, | |
| "step": 1825 | |
| }, | |
| { | |
| "epoch": 11.78343949044586, | |
| "grad_norm": 1.5006356239318848, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0292, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 11.94267515923567, | |
| "grad_norm": 2.8492929935455322, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0273, | |
| "step": 1875 | |
| }, | |
| { | |
| "epoch": 12.101910828025478, | |
| "grad_norm": 3.423665761947632, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0234, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 12.261146496815286, | |
| "grad_norm": 2.3133459091186523, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0358, | |
| "step": 1925 | |
| }, | |
| { | |
| "epoch": 12.420382165605096, | |
| "grad_norm": 3.495283603668213, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0355, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 12.579617834394904, | |
| "grad_norm": 1.5722227096557617, | |
| "learning_rate": 5e-05, | |
| "loss": 0.039, | |
| "step": 1975 | |
| }, | |
| { | |
| "epoch": 12.738853503184714, | |
| "grad_norm": 3.156038522720337, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0314, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 12.738853503184714, | |
| "eval_loss": 2.215003490447998, | |
| "eval_runtime": 260.0556, | |
| "eval_samples_per_second": 2.538, | |
| "eval_steps_per_second": 0.162, | |
| "eval_wer": 78.86568308105001, | |
| "eval_wer_ortho": 81.0546875, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 12.898089171974522, | |
| "grad_norm": 2.6695284843444824, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0364, | |
| "step": 2025 | |
| }, | |
| { | |
| "epoch": 13.05732484076433, | |
| "grad_norm": 1.3920303583145142, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0314, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 13.21656050955414, | |
| "grad_norm": 2.2318477630615234, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0226, | |
| "step": 2075 | |
| }, | |
| { | |
| "epoch": 13.375796178343949, | |
| "grad_norm": 2.457688093185425, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0374, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 13.535031847133759, | |
| "grad_norm": 3.3562824726104736, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0263, | |
| "step": 2125 | |
| }, | |
| { | |
| "epoch": 13.694267515923567, | |
| "grad_norm": 2.763430118560791, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0359, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 13.853503184713375, | |
| "grad_norm": 5.378473281860352, | |
| "learning_rate": 5e-05, | |
| "loss": 0.039, | |
| "step": 2175 | |
| }, | |
| { | |
| "epoch": 14.012738853503185, | |
| "grad_norm": 2.072021007537842, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0373, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 14.171974522292993, | |
| "grad_norm": 1.6606969833374023, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0263, | |
| "step": 2225 | |
| }, | |
| { | |
| "epoch": 14.331210191082803, | |
| "grad_norm": 3.090102195739746, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0257, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 14.490445859872612, | |
| "grad_norm": 6.045629024505615, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0232, | |
| "step": 2275 | |
| }, | |
| { | |
| "epoch": 14.64968152866242, | |
| "grad_norm": 0.8971702456474304, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0263, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 14.80891719745223, | |
| "grad_norm": 0.5613566637039185, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0295, | |
| "step": 2325 | |
| }, | |
| { | |
| "epoch": 14.968152866242038, | |
| "grad_norm": 3.578268051147461, | |
| "learning_rate": 5e-05, | |
| "loss": 0.0246, | |
| "step": 2350 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 2355, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 15, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.08565827600384e+19, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |