| { | |
| "best_metric": 0.15353840099691743, | |
| "best_model_checkpoint": "/scratch/lingjzhu_root/lingjzhu1/lingjzhu/g2p/byt5_small_finetuned_from_pretrained/checkpoint-5000", | |
| "epoch": 9.646302250803858, | |
| "global_step": 135000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 0.0003, | |
| "loss": 1.3832, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 0.0002999616623572683, | |
| "loss": 0.2586, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 0.00029984666902607135, | |
| "loss": 0.1824, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 0.0002996550787873857, | |
| "loss": 0.152, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 0.0002993869895761197, | |
| "loss": 0.1328, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "eval_cer": 0.15353840099691743, | |
| "eval_loss": 0.16419939696788788, | |
| "eval_runtime": 90.9116, | |
| "eval_samples_per_second": 54.449, | |
| "eval_steps_per_second": 0.429, | |
| "eval_wer": 0.40606060606060607, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 0.000299042538431052, | |
| "loss": 0.1171, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 0.00029862190142478177, | |
| "loss": 0.1067, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 0.00029812529357372587, | |
| "loss": 0.0979, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 0.00029755296872820933, | |
| "loss": 0.0911, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 0.0002969052194427048, | |
| "loss": 0.0857, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "eval_cer": 0.1219912113858464, | |
| "eval_loss": 0.10491432994604111, | |
| "eval_runtime": 90.8023, | |
| "eval_samples_per_second": 54.514, | |
| "eval_steps_per_second": 0.43, | |
| "eval_wer": 0.3408080808080808, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 0.0002961823768262882, | |
| "loss": 0.082, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 0.0002953848103733858, | |
| "loss": 0.0774, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 0.00029451292777490066, | |
| "loss": 0.0745, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 0.0002935671747098137, | |
| "loss": 0.0708, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 0.00029254803461736643, | |
| "loss": 0.0673, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "eval_cer": 0.11350866837192015, | |
| "eval_loss": 0.08455779403448105, | |
| "eval_runtime": 91.1047, | |
| "eval_samples_per_second": 54.333, | |
| "eval_steps_per_second": 0.428, | |
| "eval_wer": 0.31757575757575757, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.14, | |
| "learning_rate": 0.00029145602844994243, | |
| "loss": 0.065, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 0.0002902917144067724, | |
| "loss": 0.0625, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "learning_rate": 0.00028905568764860047, | |
| "loss": 0.0611, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.36, | |
| "learning_rate": 0.00028774857999345685, | |
| "loss": 0.0599, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "learning_rate": 0.0002863710595936922, | |
| "loss": 0.0585, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "eval_cer": 0.10635972103801841, | |
| "eval_loss": 0.0741606280207634, | |
| "eval_runtime": 90.9345, | |
| "eval_samples_per_second": 54.435, | |
| "eval_steps_per_second": 0.429, | |
| "eval_wer": 0.29757575757575755, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 0.0002849238305944389, | |
| "loss": 0.057, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "learning_rate": 0.00028340763277367477, | |
| "loss": 0.0557, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 1.64, | |
| "learning_rate": 0.0002818232411640713, | |
| "loss": 0.0548, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 0.00028017146565682144, | |
| "loss": 0.0538, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "learning_rate": 0.00027845315058764886, | |
| "loss": 0.0526, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "eval_cer": 0.10283990293172428, | |
| "eval_loss": 0.06704817712306976, | |
| "eval_runtime": 91.2941, | |
| "eval_samples_per_second": 54.22, | |
| "eval_steps_per_second": 0.427, | |
| "eval_wer": 0.2898989898989899, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.86, | |
| "learning_rate": 0.00027666917430520975, | |
| "loss": 0.0519, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "learning_rate": 0.00027482044872210895, | |
| "loss": 0.0506, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 0.0002729079188487587, | |
| "loss": 0.0498, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 2.07, | |
| "learning_rate": 0.00027093256231031885, | |
| "loss": 0.0468, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "learning_rate": 0.00026889538884696597, | |
| "loss": 0.0463, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "eval_cer": 0.10310224962287663, | |
| "eval_loss": 0.0634964108467102, | |
| "eval_runtime": 91.3121, | |
| "eval_samples_per_second": 54.21, | |
| "eval_steps_per_second": 0.427, | |
| "eval_wer": 0.2896969696969697, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "learning_rate": 0.0002667974397977457, | |
| "loss": 0.0461, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 2.29, | |
| "learning_rate": 0.0002646397875682729, | |
| "loss": 0.0454, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "learning_rate": 0.00026242353508255185, | |
| "loss": 0.0448, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 2.43, | |
| "learning_rate": 0.0002601498152191957, | |
| "loss": 0.0446, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "learning_rate": 0.0002578197902323352, | |
| "loss": 0.0443, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "eval_cer": 0.1002382982444634, | |
| "eval_loss": 0.06088022515177727, | |
| "eval_runtime": 91.0524, | |
| "eval_samples_per_second": 54.364, | |
| "eval_steps_per_second": 0.428, | |
| "eval_wer": 0.282020202020202, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 0.00025543465115751026, | |
| "loss": 0.0434, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 2.64, | |
| "learning_rate": 0.0002529956172028505, | |
| "loss": 0.0431, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 2.72, | |
| "learning_rate": 0.0002505039351258541, | |
| "loss": 0.0426, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 2.79, | |
| "learning_rate": 0.0002479608785960846, | |
| "loss": 0.0417, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "learning_rate": 0.0002453677475441111, | |
| "loss": 0.0415, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "eval_cer": 0.09859863142476115, | |
| "eval_loss": 0.05760909989476204, | |
| "eval_runtime": 91.2335, | |
| "eval_samples_per_second": 54.256, | |
| "eval_steps_per_second": 0.427, | |
| "eval_wer": 0.27494949494949494, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 2.93, | |
| "learning_rate": 0.00024272586749702474, | |
| "loss": 0.0411, | |
| "step": 41000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 0.0002400365889008706, | |
| "loss": 0.0408, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 3.07, | |
| "learning_rate": 0.00023730128643034235, | |
| "loss": 0.0377, | |
| "step": 43000 | |
| }, | |
| { | |
| "epoch": 3.14, | |
| "learning_rate": 0.00023452135828609167, | |
| "loss": 0.0381, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 3.22, | |
| "learning_rate": 0.0002316982254800121, | |
| "loss": 0.0379, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 3.22, | |
| "eval_cer": 0.09984477820773485, | |
| "eval_loss": 0.056990817189216614, | |
| "eval_runtime": 91.2394, | |
| "eval_samples_per_second": 54.253, | |
| "eval_steps_per_second": 0.427, | |
| "eval_wer": 0.27595959595959596, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "learning_rate": 0.00022883333110886237, | |
| "loss": 0.0375, | |
| "step": 46000 | |
| }, | |
| { | |
| "epoch": 3.36, | |
| "learning_rate": 0.00022592813961660067, | |
| "loss": 0.0372, | |
| "step": 47000 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "learning_rate": 0.00022298413604580696, | |
| "loss": 0.037, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 3.5, | |
| "learning_rate": 0.00022000282527857588, | |
| "loss": 0.0367, | |
| "step": 49000 | |
| }, | |
| { | |
| "epoch": 3.57, | |
| "learning_rate": 0.0002169857312672683, | |
| "loss": 0.0366, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 3.57, | |
| "eval_cer": 0.09800835136966835, | |
| "eval_loss": 0.05443936586380005, | |
| "eval_runtime": 91.3792, | |
| "eval_samples_per_second": 54.17, | |
| "eval_steps_per_second": 0.427, | |
| "eval_wer": 0.27434343434343433, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "learning_rate": 0.00021393439625551483, | |
| "loss": 0.0365, | |
| "step": 51000 | |
| }, | |
| { | |
| "epoch": 3.72, | |
| "learning_rate": 0.00021085037998986924, | |
| "loss": 0.036, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 3.79, | |
| "learning_rate": 0.00020773525892251514, | |
| "loss": 0.0361, | |
| "step": 53000 | |
| }, | |
| { | |
| "epoch": 3.86, | |
| "learning_rate": 0.00020459062540543316, | |
| "loss": 0.0358, | |
| "step": 54000 | |
| }, | |
| { | |
| "epoch": 3.93, | |
| "learning_rate": 0.00020141808687644067, | |
| "loss": 0.0355, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 3.93, | |
| "eval_cer": 0.09678406681095734, | |
| "eval_loss": 0.052644312381744385, | |
| "eval_runtime": 91.3207, | |
| "eval_samples_per_second": 54.205, | |
| "eval_steps_per_second": 0.427, | |
| "eval_wer": 0.27090909090909093, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 0.00019821926503751995, | |
| "loss": 0.0351, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 4.07, | |
| "learning_rate": 0.00019499579502585537, | |
| "loss": 0.0327, | |
| "step": 57000 | |
| }, | |
| { | |
| "epoch": 4.14, | |
| "learning_rate": 0.00019174932457800242, | |
| "loss": 0.0326, | |
| "step": 58000 | |
| }, | |
| { | |
| "epoch": 4.22, | |
| "learning_rate": 0.0001884815131876167, | |
| "loss": 0.0327, | |
| "step": 59000 | |
| }, | |
| { | |
| "epoch": 4.29, | |
| "learning_rate": 0.00018519403125717278, | |
| "loss": 0.0324, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 4.29, | |
| "eval_cer": 0.09630309787717803, | |
| "eval_loss": 0.05228239297866821, | |
| "eval_runtime": 91.3618, | |
| "eval_samples_per_second": 54.18, | |
| "eval_steps_per_second": 0.427, | |
| "eval_wer": 0.26686868686868687, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 4.36, | |
| "learning_rate": 0.00018188855924410722, | |
| "loss": 0.0326, | |
| "step": 61000 | |
| }, | |
| { | |
| "epoch": 4.43, | |
| "learning_rate": 0.00017856678680182127, | |
| "loss": 0.0323, | |
| "step": 62000 | |
| }, | |
| { | |
| "epoch": 4.5, | |
| "learning_rate": 0.0001752304119159834, | |
| "loss": 0.0322, | |
| "step": 63000 | |
| }, | |
| { | |
| "epoch": 4.57, | |
| "learning_rate": 0.00017188114003657205, | |
| "loss": 0.0323, | |
| "step": 64000 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "learning_rate": 0.00016852068320610358, | |
| "loss": 0.0316, | |
| "step": 65000 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "eval_cer": 0.09446667103911152, | |
| "eval_loss": 0.050846464931964874, | |
| "eval_runtime": 91.2629, | |
| "eval_samples_per_second": 54.239, | |
| "eval_steps_per_second": 0.427, | |
| "eval_wer": 0.2638383838383838, | |
| "step": 65000 | |
| }, | |
| { | |
| "epoch": 4.72, | |
| "learning_rate": 0.00016515075918448972, | |
| "loss": 0.0318, | |
| "step": 66000 | |
| }, | |
| { | |
| "epoch": 4.79, | |
| "learning_rate": 0.00016177309057097285, | |
| "loss": 0.0315, | |
| "step": 67000 | |
| }, | |
| { | |
| "epoch": 4.86, | |
| "learning_rate": 0.00015838940392358722, | |
| "loss": 0.0315, | |
| "step": 68000 | |
| }, | |
| { | |
| "epoch": 4.93, | |
| "learning_rate": 0.00015500142887659688, | |
| "loss": 0.0313, | |
| "step": 69000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "learning_rate": 0.00015161089725636095, | |
| "loss": 0.0313, | |
| "step": 70000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_cer": 0.09420432434795917, | |
| "eval_loss": 0.048644062131643295, | |
| "eval_runtime": 91.4526, | |
| "eval_samples_per_second": 54.126, | |
| "eval_steps_per_second": 0.426, | |
| "eval_wer": 0.26141414141414143, | |
| "step": 70000 | |
| }, | |
| { | |
| "epoch": 5.07, | |
| "learning_rate": 0.00014821954219607845, | |
| "loss": 0.0289, | |
| "step": 71000 | |
| }, | |
| { | |
| "epoch": 5.14, | |
| "learning_rate": 0.0001448290972498651, | |
| "loss": 0.029, | |
| "step": 72000 | |
| }, | |
| { | |
| "epoch": 5.22, | |
| "learning_rate": 0.00014144129550661485, | |
| "loss": 0.0288, | |
| "step": 73000 | |
| }, | |
| { | |
| "epoch": 5.29, | |
| "learning_rate": 0.0001380578687040995, | |
| "loss": 0.0289, | |
| "step": 74000 | |
| }, | |
| { | |
| "epoch": 5.36, | |
| "learning_rate": 0.00013468054634375843, | |
| "loss": 0.0292, | |
| "step": 75000 | |
| }, | |
| { | |
| "epoch": 5.36, | |
| "eval_cer": 0.0939638398810695, | |
| "eval_loss": 0.049041230231523514, | |
| "eval_runtime": 91.4249, | |
| "eval_samples_per_second": 54.143, | |
| "eval_steps_per_second": 0.427, | |
| "eval_wer": 0.258989898989899, | |
| "step": 75000 | |
| }, | |
| { | |
| "epoch": 5.43, | |
| "learning_rate": 0.00013131105480663235, | |
| "loss": 0.0289, | |
| "step": 76000 | |
| }, | |
| { | |
| "epoch": 5.5, | |
| "learning_rate": 0.000127951116470891, | |
| "loss": 0.0288, | |
| "step": 77000 | |
| }, | |
| { | |
| "epoch": 5.57, | |
| "learning_rate": 0.00012460244883140783, | |
| "loss": 0.0287, | |
| "step": 78000 | |
| }, | |
| { | |
| "epoch": 5.64, | |
| "learning_rate": 0.0001212667636218309, | |
| "loss": 0.0284, | |
| "step": 79000 | |
| }, | |
| { | |
| "epoch": 5.72, | |
| "learning_rate": 0.00011794576593959775, | |
| "loss": 0.0283, | |
| "step": 80000 | |
| }, | |
| { | |
| "epoch": 5.72, | |
| "eval_cer": 0.0939419776568068, | |
| "eval_loss": 0.04810198023915291, | |
| "eval_runtime": 91.3828, | |
| "eval_samples_per_second": 54.168, | |
| "eval_steps_per_second": 0.427, | |
| "eval_wer": 0.2604040404040404, | |
| "step": 80000 | |
| }, | |
| { | |
| "epoch": 5.79, | |
| "learning_rate": 0.00011464115337434394, | |
| "loss": 0.029, | |
| "step": 81000 | |
| }, | |
| { | |
| "epoch": 5.86, | |
| "learning_rate": 0.00011135461514014796, | |
| "loss": 0.0289, | |
| "step": 82000 | |
| }, | |
| { | |
| "epoch": 5.93, | |
| "learning_rate": 0.00010808783121205837, | |
| "loss": 0.0285, | |
| "step": 83000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "learning_rate": 0.00010484247146734352, | |
| "loss": 0.029, | |
| "step": 84000 | |
| }, | |
| { | |
| "epoch": 6.07, | |
| "learning_rate": 0.00010162019483190237, | |
| "loss": 0.0282, | |
| "step": 85000 | |
| }, | |
| { | |
| "epoch": 6.07, | |
| "eval_cer": 0.09354845762007827, | |
| "eval_loss": 0.048511628061532974, | |
| "eval_runtime": 93.8707, | |
| "eval_samples_per_second": 52.732, | |
| "eval_steps_per_second": 0.415, | |
| "eval_wer": 0.25696969696969696, | |
| "step": 85000 | |
| }, | |
| { | |
| "epoch": 6.15, | |
| "learning_rate": 9.842264843227404e-05, | |
| "loss": 0.0282, | |
| "step": 86000 | |
| }, | |
| { | |
| "epoch": 6.22, | |
| "learning_rate": 9.52514667536784e-05, | |
| "loss": 0.0281, | |
| "step": 87000 | |
| }, | |
| { | |
| "epoch": 6.29, | |
| "learning_rate": 9.210827080451842e-05, | |
| "loss": 0.0281, | |
| "step": 88000 | |
| }, | |
| { | |
| "epoch": 6.36, | |
| "learning_rate": 8.899466728777203e-05, | |
| "loss": 0.0279, | |
| "step": 89000 | |
| }, | |
| { | |
| "epoch": 6.43, | |
| "learning_rate": 8.591224777969557e-05, | |
| "loss": 0.028, | |
| "step": 90000 | |
| }, | |
| { | |
| "epoch": 6.43, | |
| "eval_cer": 0.09381080431123062, | |
| "eval_loss": 0.047929223626852036, | |
| "eval_runtime": 93.5944, | |
| "eval_samples_per_second": 52.888, | |
| "eval_steps_per_second": 0.417, | |
| "eval_wer": 0.2608080808080808, | |
| "step": 90000 | |
| }, | |
| { | |
| "epoch": 6.5, | |
| "learning_rate": 8.286258791626041e-05, | |
| "loss": 0.028, | |
| "step": 91000 | |
| }, | |
| { | |
| "epoch": 6.57, | |
| "learning_rate": 7.984724658773716e-05, | |
| "loss": 0.0276, | |
| "step": 92000 | |
| }, | |
| { | |
| "epoch": 6.65, | |
| "learning_rate": 7.686776514184009e-05, | |
| "loss": 0.0275, | |
| "step": 93000 | |
| }, | |
| { | |
| "epoch": 6.72, | |
| "learning_rate": 7.392566659583846e-05, | |
| "loss": 0.0274, | |
| "step": 94000 | |
| }, | |
| { | |
| "epoch": 6.79, | |
| "learning_rate": 7.102245485803813e-05, | |
| "loss": 0.0276, | |
| "step": 95000 | |
| }, | |
| { | |
| "epoch": 6.79, | |
| "eval_cer": 0.09346100872302748, | |
| "eval_loss": 0.04663492366671562, | |
| "eval_runtime": 93.7458, | |
| "eval_samples_per_second": 52.802, | |
| "eval_steps_per_second": 0.416, | |
| "eval_wer": 0.25737373737373737, | |
| "step": 95000 | |
| }, | |
| { | |
| "epoch": 6.86, | |
| "learning_rate": 6.81596139590308e-05, | |
| "loss": 0.0273, | |
| "step": 96000 | |
| }, | |
| { | |
| "epoch": 6.93, | |
| "learning_rate": 6.533860729310434e-05, | |
| "loss": 0.0275, | |
| "step": 97000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "learning_rate": 6.256087687020127e-05, | |
| "loss": 0.0271, | |
| "step": 98000 | |
| }, | |
| { | |
| "epoch": 7.07, | |
| "learning_rate": 5.98278425788092e-05, | |
| "loss": 0.0256, | |
| "step": 99000 | |
| }, | |
| { | |
| "epoch": 7.15, | |
| "learning_rate": 5.71409014601578e-05, | |
| "loss": 0.0257, | |
| "step": 100000 | |
| }, | |
| { | |
| "epoch": 7.15, | |
| "eval_cer": 0.09276141754662119, | |
| "eval_loss": 0.04688685014843941, | |
| "eval_runtime": 93.8485, | |
| "eval_samples_per_second": 52.745, | |
| "eval_steps_per_second": 0.416, | |
| "eval_wer": 0.25696969696969696, | |
| "step": 100000 | |
| }, | |
| { | |
| "epoch": 7.22, | |
| "learning_rate": 5.4501426994095876e-05, | |
| "loss": 0.0261, | |
| "step": 101000 | |
| }, | |
| { | |
| "epoch": 7.29, | |
| "learning_rate": 5.191076839701103e-05, | |
| "loss": 0.0262, | |
| "step": 102000 | |
| }, | |
| { | |
| "epoch": 7.36, | |
| "learning_rate": 4.9370249932153075e-05, | |
| "loss": 0.0261, | |
| "step": 103000 | |
| }, | |
| { | |
| "epoch": 7.43, | |
| "learning_rate": 4.6881170232712164e-05, | |
| "loss": 0.0259, | |
| "step": 104000 | |
| }, | |
| { | |
| "epoch": 7.5, | |
| "learning_rate": 4.444480163799822e-05, | |
| "loss": 0.0261, | |
| "step": 105000 | |
| }, | |
| { | |
| "epoch": 7.5, | |
| "eval_cer": 0.09184320412758794, | |
| "eval_loss": 0.04592433199286461, | |
| "eval_runtime": 93.414, | |
| "eval_samples_per_second": 52.99, | |
| "eval_steps_per_second": 0.417, | |
| "eval_wer": 0.25515151515151513, | |
| "step": 105000 | |
| }, | |
| { | |
| "epoch": 7.57, | |
| "learning_rate": 4.2062389543061265e-05, | |
| "loss": 0.0257, | |
| "step": 106000 | |
| }, | |
| { | |
| "epoch": 7.65, | |
| "learning_rate": 3.9735151762084384e-05, | |
| "loss": 0.0258, | |
| "step": 107000 | |
| }, | |
| { | |
| "epoch": 7.72, | |
| "learning_rate": 3.746427790587557e-05, | |
| "loss": 0.0258, | |
| "step": 108000 | |
| }, | |
| { | |
| "epoch": 7.79, | |
| "learning_rate": 3.525092877377602e-05, | |
| "loss": 0.0258, | |
| "step": 109000 | |
| }, | |
| { | |
| "epoch": 7.86, | |
| "learning_rate": 3.309623576029597e-05, | |
| "loss": 0.0258, | |
| "step": 110000 | |
| }, | |
| { | |
| "epoch": 7.86, | |
| "eval_cer": 0.09195251524890143, | |
| "eval_loss": 0.04556591063737869, | |
| "eval_runtime": 93.7142, | |
| "eval_samples_per_second": 52.82, | |
| "eval_steps_per_second": 0.416, | |
| "eval_wer": 0.25333333333333335, | |
| "step": 110000 | |
| }, | |
| { | |
| "epoch": 7.93, | |
| "learning_rate": 3.1001300276781274e-05, | |
| "loss": 0.0258, | |
| "step": 111000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "learning_rate": 2.8967193188406938e-05, | |
| "loss": 0.0255, | |
| "step": 112000 | |
| }, | |
| { | |
| "epoch": 8.07, | |
| "learning_rate": 2.699495426678389e-05, | |
| "loss": 0.0241, | |
| "step": 113000 | |
| }, | |
| { | |
| "epoch": 8.15, | |
| "learning_rate": 2.5085591658461056e-05, | |
| "loss": 0.0243, | |
| "step": 114000 | |
| }, | |
| { | |
| "epoch": 8.22, | |
| "learning_rate": 2.3240081369591984e-05, | |
| "loss": 0.0249, | |
| "step": 115000 | |
| }, | |
| { | |
| "epoch": 8.22, | |
| "eval_cer": 0.09116547517544435, | |
| "eval_loss": 0.04581727460026741, | |
| "eval_runtime": 93.453, | |
| "eval_samples_per_second": 52.968, | |
| "eval_steps_per_second": 0.417, | |
| "eval_wer": 0.25292929292929295, | |
| "step": 115000 | |
| }, | |
| { | |
| "epoch": 8.29, | |
| "learning_rate": 2.1459366767031522e-05, | |
| "loss": 0.0247, | |
| "step": 116000 | |
| }, | |
| { | |
| "epoch": 8.36, | |
| "learning_rate": 1.9744358096116225e-05, | |
| "loss": 0.0249, | |
| "step": 117000 | |
| }, | |
| { | |
| "epoch": 8.43, | |
| "learning_rate": 1.8095932015375496e-05, | |
| "loss": 0.0247, | |
| "step": 118000 | |
| }, | |
| { | |
| "epoch": 8.5, | |
| "learning_rate": 1.65149311484114e-05, | |
| "loss": 0.0246, | |
| "step": 119000 | |
| }, | |
| { | |
| "epoch": 8.57, | |
| "learning_rate": 1.500216365317587e-05, | |
| "loss": 0.025, | |
| "step": 120000 | |
| }, | |
| { | |
| "epoch": 8.57, | |
| "eval_cer": 0.09134037296954592, | |
| "eval_loss": 0.045206133276224136, | |
| "eval_runtime": 93.5201, | |
| "eval_samples_per_second": 52.93, | |
| "eval_steps_per_second": 0.417, | |
| "eval_wer": 0.25191919191919193, | |
| "step": 120000 | |
| }, | |
| { | |
| "epoch": 8.65, | |
| "learning_rate": 1.355840280886582e-05, | |
| "loss": 0.025, | |
| "step": 121000 | |
| }, | |
| { | |
| "epoch": 8.72, | |
| "learning_rate": 1.2184386620647097e-05, | |
| "loss": 0.0247, | |
| "step": 122000 | |
| }, | |
| { | |
| "epoch": 8.79, | |
| "learning_rate": 1.0880817442409478e-05, | |
| "loss": 0.0248, | |
| "step": 123000 | |
| }, | |
| { | |
| "epoch": 8.86, | |
| "learning_rate": 9.648361617745371e-06, | |
| "loss": 0.0248, | |
| "step": 124000 | |
| }, | |
| { | |
| "epoch": 8.93, | |
| "learning_rate": 8.487649139335962e-06, | |
| "loss": 0.0248, | |
| "step": 125000 | |
| }, | |
| { | |
| "epoch": 8.93, | |
| "eval_cer": 0.09169016855774907, | |
| "eval_loss": 0.04524415731430054, | |
| "eval_runtime": 93.2177, | |
| "eval_samples_per_second": 53.101, | |
| "eval_steps_per_second": 0.418, | |
| "eval_wer": 0.25292929292929295, | |
| "step": 125000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "learning_rate": 7.399273326918692e-06, | |
| "loss": 0.0246, | |
| "step": 126000 | |
| }, | |
| { | |
| "epoch": 9.07, | |
| "learning_rate": 6.383790524001009e-06, | |
| "loss": 0.024, | |
| "step": 127000 | |
| }, | |
| { | |
| "epoch": 9.15, | |
| "learning_rate": 5.441719813474849e-06, | |
| "loss": 0.0241, | |
| "step": 128000 | |
| }, | |
| { | |
| "epoch": 9.22, | |
| "learning_rate": 4.57354275227797e-06, | |
| "loss": 0.0242, | |
| "step": 129000 | |
| }, | |
| { | |
| "epoch": 9.29, | |
| "learning_rate": 3.7797031252369767e-06, | |
| "loss": 0.0243, | |
| "step": 130000 | |
| }, | |
| { | |
| "epoch": 9.29, | |
| "eval_cer": 0.09138409741807131, | |
| "eval_loss": 0.0453341044485569, | |
| "eval_runtime": 93.6155, | |
| "eval_samples_per_second": 52.876, | |
| "eval_steps_per_second": 0.417, | |
| "eval_wer": 0.25212121212121213, | |
| "step": 130000 | |
| }, | |
| { | |
| "epoch": 9.36, | |
| "learning_rate": 3.0606067182186776e-06, | |
| "loss": 0.024, | |
| "step": 131000 | |
| }, | |
| { | |
| "epoch": 9.43, | |
| "learning_rate": 2.4166211107049584e-06, | |
| "loss": 0.0242, | |
| "step": 132000 | |
| }, | |
| { | |
| "epoch": 9.5, | |
| "learning_rate": 1.8480754878977489e-06, | |
| "loss": 0.0241, | |
| "step": 133000 | |
| }, | |
| { | |
| "epoch": 9.57, | |
| "learning_rate": 1.3552604724498928e-06, | |
| "loss": 0.0241, | |
| "step": 134000 | |
| }, | |
| { | |
| "epoch": 9.65, | |
| "learning_rate": 9.384279759080127e-07, | |
| "loss": 0.024, | |
| "step": 135000 | |
| }, | |
| { | |
| "epoch": 9.65, | |
| "eval_cer": 0.09129664852102053, | |
| "eval_loss": 0.04531262069940567, | |
| "eval_runtime": 93.4097, | |
| "eval_samples_per_second": 52.992, | |
| "eval_steps_per_second": 0.418, | |
| "eval_wer": 0.25212121212121213, | |
| "step": 135000 | |
| } | |
| ], | |
| "max_steps": 139950, | |
| "num_train_epochs": 10, | |
| "total_flos": 4.19385405764928e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |