| { | |
| "best_metric": 0.07587812840938568, | |
| "best_model_checkpoint": "/scratch/lingjzhu_root/lingjzhu1/lingjzhu/g2p/byt5_8_layers_baseline/checkpoint-135000", | |
| "epoch": 9.917791237204689, | |
| "global_step": 140000, | |
| "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": 8.3596, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 0.0002999623214094654, | |
| "loss": 1.7981, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 0.0002998493045668776, | |
| "loss": 0.7044, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 0.00029966100624977436, | |
| "loss": 0.4284, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 0.00029939752105569156, | |
| "loss": 0.3384, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "eval_cer": 0.2570559951702279, | |
| "eval_loss": 0.38358163833618164, | |
| "eval_runtime": 77.3539, | |
| "eval_samples_per_second": 64.638, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.5968, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 0.00029905898135463893, | |
| "loss": 0.2935, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 0.0002986455572226002, | |
| "loss": 0.2654, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 0.00029815745635609, | |
| "loss": 0.2443, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 0.00029759492396781067, | |
| "loss": 0.229, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 0.00029695824266346256, | |
| "loss": 0.2169, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "eval_cer": 0.1943552038638177, | |
| "eval_loss": 0.24924039840698242, | |
| "eval_runtime": 76.8393, | |
| "eval_samples_per_second": 65.071, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.4916, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 0.000296247732299768, | |
| "loss": 0.2066, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 0.000295463749823781, | |
| "loss": 0.1979, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 0.0002946066890935644, | |
| "loss": 0.1906, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 0.0002936769806803222, | |
| "loss": 0.184, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "learning_rate": 0.00029267509165208945, | |
| "loss": 0.1769, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.06, | |
| "eval_cer": 0.1750576769658682, | |
| "eval_loss": 0.19950397312641144, | |
| "eval_runtime": 77.3025, | |
| "eval_samples_per_second": 64.681, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.4484, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 1.13, | |
| "learning_rate": 0.00029160152533908566, | |
| "loss": 0.1726, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "learning_rate": 0.0002904568210808509, | |
| "loss": 0.1676, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 1.28, | |
| "learning_rate": 0.00028924155395529237, | |
| "loss": 0.1627, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 1.35, | |
| "learning_rate": 0.00028795633448977547, | |
| "loss": 0.1583, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "learning_rate": 0.0002866018083544069, | |
| "loss": 0.1559, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.42, | |
| "eval_cer": 0.15757131460359214, | |
| "eval_loss": 0.1710461676120758, | |
| "eval_runtime": 77.109, | |
| "eval_samples_per_second": 64.843, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.4136, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 1.49, | |
| "learning_rate": 0.0002851786560376614, | |
| "loss": 0.1511, | |
| "step": 21000 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 0.00028368759250451786, | |
| "loss": 0.1473, | |
| "step": 22000 | |
| }, | |
| { | |
| "epoch": 1.63, | |
| "learning_rate": 0.00028212936683727404, | |
| "loss": 0.1443, | |
| "step": 23000 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "learning_rate": 0.0002805047618592216, | |
| "loss": 0.1414, | |
| "step": 24000 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "learning_rate": 0.0002788145937413704, | |
| "loss": 0.1387, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.77, | |
| "eval_cer": 0.1498307423618448, | |
| "eval_loss": 0.15107452869415283, | |
| "eval_runtime": 76.8475, | |
| "eval_samples_per_second": 65.064, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.4002, | |
| "step": 25000 | |
| }, | |
| { | |
| "epoch": 1.84, | |
| "learning_rate": 0.00027705971159241994, | |
| "loss": 0.1363, | |
| "step": 26000 | |
| }, | |
| { | |
| "epoch": 1.91, | |
| "learning_rate": 0.0002752409970321824, | |
| "loss": 0.1339, | |
| "step": 27000 | |
| }, | |
| { | |
| "epoch": 1.98, | |
| "learning_rate": 0.000273359363748674, | |
| "loss": 0.132, | |
| "step": 28000 | |
| }, | |
| { | |
| "epoch": 2.05, | |
| "learning_rate": 0.0002714157570390952, | |
| "loss": 0.1284, | |
| "step": 29000 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "learning_rate": 0.0002694111533349308, | |
| "loss": 0.126, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 2.13, | |
| "eval_cer": 0.14055930485780202, | |
| "eval_loss": 0.13569805026054382, | |
| "eval_runtime": 77.0438, | |
| "eval_samples_per_second": 64.898, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.3784, | |
| "step": 30000 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "learning_rate": 0.00026734655971140956, | |
| "loss": 0.1241, | |
| "step": 31000 | |
| }, | |
| { | |
| "epoch": 2.27, | |
| "learning_rate": 0.00026522301338156835, | |
| "loss": 0.1221, | |
| "step": 32000 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 0.00026304158117517563, | |
| "loss": 0.1207, | |
| "step": 33000 | |
| }, | |
| { | |
| "epoch": 2.41, | |
| "learning_rate": 0.0002608033590027766, | |
| "loss": 0.1195, | |
| "step": 34000 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "learning_rate": 0.000258509471305128, | |
| "loss": 0.1173, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 2.48, | |
| "eval_cer": 0.13616076241402358, | |
| "eval_loss": 0.12640255689620972, | |
| "eval_runtime": 76.9784, | |
| "eval_samples_per_second": 64.953, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.369, | |
| "step": 35000 | |
| }, | |
| { | |
| "epoch": 2.55, | |
| "learning_rate": 0.00025616107048830045, | |
| "loss": 0.1156, | |
| "step": 36000 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "learning_rate": 0.0002537593363447311, | |
| "loss": 0.1141, | |
| "step": 37000 | |
| }, | |
| { | |
| "epoch": 2.69, | |
| "learning_rate": 0.0002513054754605181, | |
| "loss": 0.113, | |
| "step": 38000 | |
| }, | |
| { | |
| "epoch": 2.76, | |
| "learning_rate": 0.0002488007206092547, | |
| "loss": 0.1116, | |
| "step": 39000 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "learning_rate": 0.00024624633013270664, | |
| "loss": 0.1101, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 2.83, | |
| "eval_cer": 0.13396149119213438, | |
| "eval_loss": 0.1196521520614624, | |
| "eval_runtime": 77.1165, | |
| "eval_samples_per_second": 64.837, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.3636, | |
| "step": 40000 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "learning_rate": 0.0002436435873086449, | |
| "loss": 0.1091, | |
| "step": 41000 | |
| }, | |
| { | |
| "epoch": 2.98, | |
| "learning_rate": 0.00024099379970615137, | |
| "loss": 0.1073, | |
| "step": 42000 | |
| }, | |
| { | |
| "epoch": 3.05, | |
| "learning_rate": 0.00023829829852872037, | |
| "loss": 0.1057, | |
| "step": 43000 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "learning_rate": 0.00023555843794548724, | |
| "loss": 0.1046, | |
| "step": 44000 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "learning_rate": 0.00023277559441091917, | |
| "loss": 0.1035, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 3.19, | |
| "eval_cer": 0.1291748420621402, | |
| "eval_loss": 0.11245186626911163, | |
| "eval_runtime": 77.1099, | |
| "eval_samples_per_second": 64.843, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.3536, | |
| "step": 45000 | |
| }, | |
| { | |
| "epoch": 3.26, | |
| "learning_rate": 0.0002299511659733103, | |
| "loss": 0.102, | |
| "step": 46000 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "learning_rate": 0.0002270865715724285, | |
| "loss": 0.1011, | |
| "step": 47000 | |
| }, | |
| { | |
| "epoch": 3.4, | |
| "learning_rate": 0.0002241832503266669, | |
| "loss": 0.101, | |
| "step": 48000 | |
| }, | |
| { | |
| "epoch": 3.47, | |
| "learning_rate": 0.00022124266081005755, | |
| "loss": 0.0999, | |
| "step": 49000 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "learning_rate": 0.0002182662803195115, | |
| "loss": 0.0993, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 3.54, | |
| "eval_cer": 0.12630716488065719, | |
| "eval_loss": 0.10715476423501968, | |
| "eval_runtime": 77.1106, | |
| "eval_samples_per_second": 64.842, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.3478, | |
| "step": 50000 | |
| }, | |
| { | |
| "epoch": 3.61, | |
| "learning_rate": 0.00021525560413265246, | |
| "loss": 0.0975, | |
| "step": 51000 | |
| }, | |
| { | |
| "epoch": 3.68, | |
| "learning_rate": 0.00021221214475661745, | |
| "loss": 0.0968, | |
| "step": 52000 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "learning_rate": 0.00020913743116820165, | |
| "loss": 0.0966, | |
| "step": 53000 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "learning_rate": 0.00020603300804572917, | |
| "loss": 0.0955, | |
| "step": 54000 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "learning_rate": 0.00020290043499303573, | |
| "loss": 0.0949, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "eval_cer": 0.12458224627525388, | |
| "eval_loss": 0.10156673938035965, | |
| "eval_runtime": 77.2608, | |
| "eval_samples_per_second": 64.716, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.3422, | |
| "step": 55000 | |
| }, | |
| { | |
| "epoch": 3.97, | |
| "learning_rate": 0.00019974128575595291, | |
| "loss": 0.0937, | |
| "step": 56000 | |
| }, | |
| { | |
| "epoch": 4.04, | |
| "learning_rate": 0.000196557147431688, | |
| "loss": 0.0925, | |
| "step": 57000 | |
| }, | |
| { | |
| "epoch": 4.11, | |
| "learning_rate": 0.00019334961967149597, | |
| "loss": 0.0917, | |
| "step": 58000 | |
| }, | |
| { | |
| "epoch": 4.18, | |
| "learning_rate": 0.00019012031387704483, | |
| "loss": 0.091, | |
| "step": 59000 | |
| }, | |
| { | |
| "epoch": 4.25, | |
| "learning_rate": 0.00018687085239087786, | |
| "loss": 0.0905, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 4.25, | |
| "eval_cer": 0.12152051575066301, | |
| "eval_loss": 0.09692572802305222, | |
| "eval_runtime": 77.031, | |
| "eval_samples_per_second": 64.909, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.3398, | |
| "step": 60000 | |
| }, | |
| { | |
| "epoch": 4.32, | |
| "learning_rate": 0.00018360286768137894, | |
| "loss": 0.0898, | |
| "step": 61000 | |
| }, | |
| { | |
| "epoch": 4.39, | |
| "learning_rate": 0.00018031800152265137, | |
| "loss": 0.0886, | |
| "step": 62000 | |
| }, | |
| { | |
| "epoch": 4.46, | |
| "learning_rate": 0.00017701790416972107, | |
| "loss": 0.0886, | |
| "step": 63000 | |
| }, | |
| { | |
| "epoch": 4.53, | |
| "learning_rate": 0.00017370423352948002, | |
| "loss": 0.0882, | |
| "step": 64000 | |
| }, | |
| { | |
| "epoch": 4.6, | |
| "learning_rate": 0.00017037865432778442, | |
| "loss": 0.0875, | |
| "step": 65000 | |
| }, | |
| { | |
| "epoch": 4.6, | |
| "eval_cer": 0.12011901938377283, | |
| "eval_loss": 0.09379537403583527, | |
| "eval_runtime": 77.3356, | |
| "eval_samples_per_second": 64.653, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.3314, | |
| "step": 65000 | |
| }, | |
| { | |
| "epoch": 4.68, | |
| "learning_rate": 0.00016704283727312792, | |
| "loss": 0.0867, | |
| "step": 66000 | |
| }, | |
| { | |
| "epoch": 4.75, | |
| "learning_rate": 0.00016369845821730933, | |
| "loss": 0.0865, | |
| "step": 67000 | |
| }, | |
| { | |
| "epoch": 4.82, | |
| "learning_rate": 0.00016034719731351576, | |
| "loss": 0.0858, | |
| "step": 68000 | |
| }, | |
| { | |
| "epoch": 4.89, | |
| "learning_rate": 0.00015699073817224542, | |
| "loss": 0.0854, | |
| "step": 69000 | |
| }, | |
| { | |
| "epoch": 4.96, | |
| "learning_rate": 0.00015363076701549341, | |
| "loss": 0.0849, | |
| "step": 70000 | |
| }, | |
| { | |
| "epoch": 4.96, | |
| "eval_cer": 0.11830785484809936, | |
| "eval_loss": 0.09131144732236862, | |
| "eval_runtime": 77.6068, | |
| "eval_samples_per_second": 64.427, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.3296, | |
| "step": 70000 | |
| }, | |
| { | |
| "epoch": 5.03, | |
| "learning_rate": 0.00015026897182962535, | |
| "loss": 0.0841, | |
| "step": 71000 | |
| }, | |
| { | |
| "epoch": 5.1, | |
| "learning_rate": 0.00014690704151736484, | |
| "loss": 0.0825, | |
| "step": 72000 | |
| }, | |
| { | |
| "epoch": 5.17, | |
| "learning_rate": 0.0001435466650493204, | |
| "loss": 0.0822, | |
| "step": 73000 | |
| }, | |
| { | |
| "epoch": 5.24, | |
| "learning_rate": 0.0001401895306154785, | |
| "loss": 0.082, | |
| "step": 74000 | |
| }, | |
| { | |
| "epoch": 5.31, | |
| "learning_rate": 0.0001368373247770885, | |
| "loss": 0.082, | |
| "step": 75000 | |
| }, | |
| { | |
| "epoch": 5.31, | |
| "eval_cer": 0.11716509627201967, | |
| "eval_loss": 0.08901260048151016, | |
| "eval_runtime": 77.1989, | |
| "eval_samples_per_second": 64.768, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.328, | |
| "step": 75000 | |
| }, | |
| { | |
| "epoch": 5.38, | |
| "learning_rate": 0.00013349173161936601, | |
| "loss": 0.0814, | |
| "step": 76000 | |
| }, | |
| { | |
| "epoch": 5.45, | |
| "learning_rate": 0.00013015443190544018, | |
| "loss": 0.0806, | |
| "step": 77000 | |
| }, | |
| { | |
| "epoch": 5.53, | |
| "learning_rate": 0.00012682710223196978, | |
| "loss": 0.0804, | |
| "step": 78000 | |
| }, | |
| { | |
| "epoch": 5.6, | |
| "learning_rate": 0.00012351141418685273, | |
| "loss": 0.0802, | |
| "step": 79000 | |
| }, | |
| { | |
| "epoch": 5.67, | |
| "learning_rate": 0.00012020903350945151, | |
| "loss": 0.0795, | |
| "step": 80000 | |
| }, | |
| { | |
| "epoch": 5.67, | |
| "eval_cer": 0.11610858362621014, | |
| "eval_loss": 0.08586645126342773, | |
| "eval_runtime": 77.3472, | |
| "eval_samples_per_second": 64.644, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.323, | |
| "step": 80000 | |
| }, | |
| { | |
| "epoch": 5.74, | |
| "learning_rate": 0.00011692161925375738, | |
| "loss": 0.0796, | |
| "step": 81000 | |
| }, | |
| { | |
| "epoch": 5.81, | |
| "learning_rate": 0.00011365082295491233, | |
| "loss": 0.079, | |
| "step": 82000 | |
| }, | |
| { | |
| "epoch": 5.88, | |
| "learning_rate": 0.00011039828779950928, | |
| "loss": 0.0786, | |
| "step": 83000 | |
| }, | |
| { | |
| "epoch": 5.95, | |
| "learning_rate": 0.0001071656478000858, | |
| "loss": 0.0785, | |
| "step": 84000 | |
| }, | |
| { | |
| "epoch": 6.02, | |
| "learning_rate": 0.00010395452697422703, | |
| "loss": 0.0778, | |
| "step": 85000 | |
| }, | |
| { | |
| "epoch": 6.02, | |
| "eval_cer": 0.11559110804458915, | |
| "eval_loss": 0.08337103575468063, | |
| "eval_runtime": 77.0153, | |
| "eval_samples_per_second": 64.922, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.322, | |
| "step": 85000 | |
| }, | |
| { | |
| "epoch": 6.09, | |
| "learning_rate": 0.00010076653852868962, | |
| "loss": 0.0766, | |
| "step": 86000 | |
| }, | |
| { | |
| "epoch": 6.16, | |
| "learning_rate": 9.760328404895728e-05, | |
| "loss": 0.0762, | |
| "step": 87000 | |
| }, | |
| { | |
| "epoch": 6.23, | |
| "learning_rate": 9.446635269463389e-05, | |
| "loss": 0.0759, | |
| "step": 88000 | |
| }, | |
| { | |
| "epoch": 6.3, | |
| "learning_rate": 9.135732040107998e-05, | |
| "loss": 0.0759, | |
| "step": 89000 | |
| }, | |
| { | |
| "epoch": 6.38, | |
| "learning_rate": 8.82777490876921e-05, | |
| "loss": 0.0755, | |
| "step": 90000 | |
| }, | |
| { | |
| "epoch": 6.38, | |
| "eval_cer": 0.11524612432350849, | |
| "eval_loss": 0.08183012902736664, | |
| "eval_runtime": 77.0479, | |
| "eval_samples_per_second": 64.895, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.3196, | |
| "step": 90000 | |
| }, | |
| { | |
| "epoch": 6.45, | |
| "learning_rate": 8.522918587322418e-05, | |
| "loss": 0.0752, | |
| "step": 91000 | |
| }, | |
| { | |
| "epoch": 6.52, | |
| "learning_rate": 8.22131622985438e-05, | |
| "loss": 0.0754, | |
| "step": 92000 | |
| }, | |
| { | |
| "epoch": 6.59, | |
| "learning_rate": 7.923119355721526e-05, | |
| "loss": 0.0752, | |
| "step": 93000 | |
| }, | |
| { | |
| "epoch": 6.66, | |
| "learning_rate": 7.628477773429438e-05, | |
| "loss": 0.0745, | |
| "step": 94000 | |
| }, | |
| { | |
| "epoch": 6.73, | |
| "learning_rate": 7.3375395053719e-05, | |
| "loss": 0.0747, | |
| "step": 95000 | |
| }, | |
| { | |
| "epoch": 6.73, | |
| "eval_cer": 0.11298216865391664, | |
| "eval_loss": 0.0807877779006958, | |
| "eval_runtime": 77.4517, | |
| "eval_samples_per_second": 64.556, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.318, | |
| "step": 95000 | |
| }, | |
| { | |
| "epoch": 6.8, | |
| "learning_rate": 7.050450713467216e-05, | |
| "loss": 0.0737, | |
| "step": 96000 | |
| }, | |
| { | |
| "epoch": 6.87, | |
| "learning_rate": 6.767355625729215e-05, | |
| "loss": 0.0735, | |
| "step": 97000 | |
| }, | |
| { | |
| "epoch": 6.94, | |
| "learning_rate": 6.488396463809805e-05, | |
| "loss": 0.0736, | |
| "step": 98000 | |
| }, | |
| { | |
| "epoch": 7.01, | |
| "learning_rate": 6.213713371549497e-05, | |
| "loss": 0.0729, | |
| "step": 99000 | |
| }, | |
| { | |
| "epoch": 7.08, | |
| "learning_rate": 5.943444344571755e-05, | |
| "loss": 0.0717, | |
| "step": 100000 | |
| }, | |
| { | |
| "epoch": 7.08, | |
| "eval_cer": 0.11276655382824123, | |
| "eval_loss": 0.07980009913444519, | |
| "eval_runtime": 77.3354, | |
| "eval_samples_per_second": 64.653, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.3174, | |
| "step": 100000 | |
| }, | |
| { | |
| "epoch": 7.15, | |
| "learning_rate": 5.6777251609566e-05, | |
| "loss": 0.0722, | |
| "step": 101000 | |
| }, | |
| { | |
| "epoch": 7.23, | |
| "learning_rate": 5.416689313028253e-05, | |
| "loss": 0.0717, | |
| "step": 102000 | |
| }, | |
| { | |
| "epoch": 7.3, | |
| "learning_rate": 5.1604679402911015e-05, | |
| "loss": 0.0717, | |
| "step": 103000 | |
| }, | |
| { | |
| "epoch": 7.37, | |
| "learning_rate": 4.9091897635476726e-05, | |
| "loss": 0.0716, | |
| "step": 104000 | |
| }, | |
| { | |
| "epoch": 7.44, | |
| "learning_rate": 4.6629810202317166e-05, | |
| "loss": 0.0716, | |
| "step": 105000 | |
| }, | |
| { | |
| "epoch": 7.44, | |
| "eval_cer": 0.11278811531080878, | |
| "eval_loss": 0.07881917804479599, | |
| "eval_runtime": 77.1771, | |
| "eval_samples_per_second": 64.786, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.3126, | |
| "step": 105000 | |
| }, | |
| { | |
| "epoch": 7.51, | |
| "learning_rate": 4.421965400988918e-05, | |
| "loss": 0.0713, | |
| "step": 106000 | |
| }, | |
| { | |
| "epoch": 7.58, | |
| "learning_rate": 4.1862639875370016e-05, | |
| "loss": 0.0712, | |
| "step": 107000 | |
| }, | |
| { | |
| "epoch": 7.65, | |
| "learning_rate": 3.955995191836581e-05, | |
| "loss": 0.0713, | |
| "step": 108000 | |
| }, | |
| { | |
| "epoch": 7.72, | |
| "learning_rate": 3.731274696603201e-05, | |
| "loss": 0.071, | |
| "step": 109000 | |
| }, | |
| { | |
| "epoch": 7.79, | |
| "learning_rate": 3.512215397190527e-05, | |
| "loss": 0.0707, | |
| "step": 110000 | |
| }, | |
| { | |
| "epoch": 7.79, | |
| "eval_cer": 0.11227063972918779, | |
| "eval_loss": 0.07775964587926865, | |
| "eval_runtime": 77.2915, | |
| "eval_samples_per_second": 64.69, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.3126, | |
| "step": 110000 | |
| }, | |
| { | |
| "epoch": 7.86, | |
| "learning_rate": 3.298927344873821e-05, | |
| "loss": 0.0703, | |
| "step": 111000 | |
| }, | |
| { | |
| "epoch": 7.93, | |
| "learning_rate": 3.091517691562284e-05, | |
| "loss": 0.0703, | |
| "step": 112000 | |
| }, | |
| { | |
| "epoch": 8.01, | |
| "learning_rate": 2.890090635967912e-05, | |
| "loss": 0.0708, | |
| "step": 113000 | |
| }, | |
| { | |
| "epoch": 8.08, | |
| "learning_rate": 2.694747371258042e-05, | |
| "loss": 0.0694, | |
| "step": 114000 | |
| }, | |
| { | |
| "epoch": 8.15, | |
| "learning_rate": 2.505586034217802e-05, | |
| "loss": 0.069, | |
| "step": 115000 | |
| }, | |
| { | |
| "epoch": 8.15, | |
| "eval_cer": 0.11177472563013433, | |
| "eval_loss": 0.07695135474205017, | |
| "eval_runtime": 77.144, | |
| "eval_samples_per_second": 64.814, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.3118, | |
| "step": 115000 | |
| }, | |
| { | |
| "epoch": 8.22, | |
| "learning_rate": 2.3227016559480356e-05, | |
| "loss": 0.0695, | |
| "step": 116000 | |
| }, | |
| { | |
| "epoch": 8.29, | |
| "learning_rate": 2.1461861141234628e-05, | |
| "loss": 0.0693, | |
| "step": 117000 | |
| }, | |
| { | |
| "epoch": 8.36, | |
| "learning_rate": 1.9761280868350654e-05, | |
| "loss": 0.0692, | |
| "step": 118000 | |
| }, | |
| { | |
| "epoch": 8.43, | |
| "learning_rate": 1.8126130080398725e-05, | |
| "loss": 0.0693, | |
| "step": 119000 | |
| }, | |
| { | |
| "epoch": 8.5, | |
| "learning_rate": 1.655723024640553e-05, | |
| "loss": 0.069, | |
| "step": 120000 | |
| }, | |
| { | |
| "epoch": 8.5, | |
| "eval_cer": 0.11201190193837729, | |
| "eval_loss": 0.0767173245549202, | |
| "eval_runtime": 77.276, | |
| "eval_samples_per_second": 64.703, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.315, | |
| "step": 120000 | |
| }, | |
| { | |
| "epoch": 8.57, | |
| "learning_rate": 1.5055369552163509e-05, | |
| "loss": 0.0689, | |
| "step": 121000 | |
| }, | |
| { | |
| "epoch": 8.64, | |
| "learning_rate": 1.3621302504261194e-05, | |
| "loss": 0.0687, | |
| "step": 122000 | |
| }, | |
| { | |
| "epoch": 8.71, | |
| "learning_rate": 1.2255749551033205e-05, | |
| "loss": 0.0686, | |
| "step": 123000 | |
| }, | |
| { | |
| "epoch": 8.78, | |
| "learning_rate": 1.0959396720620555e-05, | |
| "loss": 0.0688, | |
| "step": 124000 | |
| }, | |
| { | |
| "epoch": 8.86, | |
| "learning_rate": 9.732895276323099e-06, | |
| "loss": 0.0686, | |
| "step": 125000 | |
| }, | |
| { | |
| "epoch": 8.86, | |
| "eval_cer": 0.11177472563013433, | |
| "eval_loss": 0.07622109353542328, | |
| "eval_runtime": 77.3587, | |
| "eval_samples_per_second": 64.634, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.3114, | |
| "step": 125000 | |
| }, | |
| { | |
| "epoch": 8.93, | |
| "learning_rate": 8.576861389416939e-06, | |
| "loss": 0.0683, | |
| "step": 126000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "learning_rate": 7.4918758296016734e-06, | |
| "loss": 0.0684, | |
| "step": 127000 | |
| }, | |
| { | |
| "epoch": 9.07, | |
| "learning_rate": 6.478483673232537e-06, | |
| "loss": 0.0682, | |
| "step": 128000 | |
| }, | |
| { | |
| "epoch": 9.14, | |
| "learning_rate": 5.537194029484349e-06, | |
| "loss": 0.0677, | |
| "step": 129000 | |
| }, | |
| { | |
| "epoch": 9.21, | |
| "learning_rate": 4.668479784584583e-06, | |
| "loss": 0.0677, | |
| "step": 130000 | |
| }, | |
| { | |
| "epoch": 9.21, | |
| "eval_cer": 0.1117100411824317, | |
| "eval_loss": 0.07608392089605331, | |
| "eval_runtime": 77.3029, | |
| "eval_samples_per_second": 64.681, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.312, | |
| "step": 130000 | |
| }, | |
| { | |
| "epoch": 9.28, | |
| "learning_rate": 3.872777364244245e-06, | |
| "loss": 0.0683, | |
| "step": 131000 | |
| }, | |
| { | |
| "epoch": 9.35, | |
| "learning_rate": 3.150486514405798e-06, | |
| "loss": 0.068, | |
| "step": 132000 | |
| }, | |
| { | |
| "epoch": 9.42, | |
| "learning_rate": 2.50197010041826e-06, | |
| "loss": 0.0679, | |
| "step": 133000 | |
| }, | |
| { | |
| "epoch": 9.49, | |
| "learning_rate": 1.9275539247405437e-06, | |
| "loss": 0.0677, | |
| "step": 134000 | |
| }, | |
| { | |
| "epoch": 9.56, | |
| "learning_rate": 1.4275265632643806e-06, | |
| "loss": 0.0686, | |
| "step": 135000 | |
| }, | |
| { | |
| "epoch": 9.56, | |
| "eval_cer": 0.11110631967054055, | |
| "eval_loss": 0.07587812840938568, | |
| "eval_runtime": 77.3541, | |
| "eval_samples_per_second": 64.638, | |
| "eval_steps_per_second": 0.129, | |
| "eval_wer": 0.3122, | |
| "step": 135000 | |
| }, | |
| { | |
| "epoch": 9.63, | |
| "learning_rate": 1.0021392203391932e-06, | |
| "loss": 0.0674, | |
| "step": 136000 | |
| }, | |
| { | |
| "epoch": 9.71, | |
| "learning_rate": 6.516056025718585e-07, | |
| "loss": 0.0679, | |
| "step": 137000 | |
| }, | |
| { | |
| "epoch": 9.78, | |
| "learning_rate": 3.76101811464391e-07, | |
| "loss": 0.0679, | |
| "step": 138000 | |
| }, | |
| { | |
| "epoch": 9.85, | |
| "learning_rate": 1.7576625494392648e-07, | |
| "loss": 0.0683, | |
| "step": 139000 | |
| }, | |
| { | |
| "epoch": 9.92, | |
| "learning_rate": 5.0699577829210744e-08, | |
| "loss": 0.0683, | |
| "step": 140000 | |
| }, | |
| { | |
| "epoch": 9.92, | |
| "eval_cer": 0.11117100411824317, | |
| "eval_loss": 0.07590720802545547, | |
| "eval_runtime": 76.9386, | |
| "eval_samples_per_second": 64.987, | |
| "eval_steps_per_second": 0.13, | |
| "eval_wer": 0.312, | |
| "step": 140000 | |
| } | |
| ], | |
| "max_steps": 141160, | |
| "num_train_epochs": 10, | |
| "total_flos": 1.2397348217772442e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |