| { | |
| "best_metric": 50.2114, | |
| "best_model_checkpoint": "./jako_13p_tokenie_run1/checkpoint-19200", | |
| "epoch": 9.997403271877435, | |
| "eval_steps": 1600, | |
| "global_step": 19250, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.9212598425196856e-05, | |
| "loss": 1.7571, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 4.79002624671916e-05, | |
| "loss": 1.3294, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 4.6587926509186354e-05, | |
| "loss": 1.2125, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "eval_bleu": 44.2732, | |
| "eval_gen_len": 18.9394, | |
| "eval_loss": 1.1356315612792969, | |
| "eval_runtime": 557.6241, | |
| "eval_samples_per_second": 13.81, | |
| "eval_steps_per_second": 0.864, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.04, | |
| "learning_rate": 4.52755905511811e-05, | |
| "loss": 1.1386, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "learning_rate": 4.396325459317586e-05, | |
| "loss": 0.9283, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 4.2650918635170604e-05, | |
| "loss": 0.8519, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.66, | |
| "eval_bleu": 47.1622, | |
| "eval_gen_len": 18.3936, | |
| "eval_loss": 1.061800241470337, | |
| "eval_runtime": 524.2089, | |
| "eval_samples_per_second": 14.691, | |
| "eval_steps_per_second": 0.919, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 1.82, | |
| "learning_rate": 4.133858267716536e-05, | |
| "loss": 0.8109, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.08, | |
| "learning_rate": 4.00262467191601e-05, | |
| "loss": 0.7727, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 3.871391076115486e-05, | |
| "loss": 0.6394, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.49, | |
| "eval_bleu": 47.7818, | |
| "eval_gen_len": 18.3397, | |
| "eval_loss": 1.0923182964324951, | |
| "eval_runtime": 516.7936, | |
| "eval_samples_per_second": 14.902, | |
| "eval_steps_per_second": 0.933, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "learning_rate": 3.740157480314961e-05, | |
| "loss": 0.5875, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "learning_rate": 3.608923884514436e-05, | |
| "loss": 0.5625, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 3.12, | |
| "learning_rate": 3.4776902887139105e-05, | |
| "loss": 0.532, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.32, | |
| "eval_bleu": 48.4283, | |
| "eval_gen_len": 18.3375, | |
| "eval_loss": 1.1293830871582031, | |
| "eval_runtime": 519.3842, | |
| "eval_samples_per_second": 14.827, | |
| "eval_steps_per_second": 0.928, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 3.38, | |
| "learning_rate": 3.3464566929133864e-05, | |
| "loss": 0.4299, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.64, | |
| "learning_rate": 3.215223097112861e-05, | |
| "loss": 0.3984, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.9, | |
| "learning_rate": 3.083989501312336e-05, | |
| "loss": 0.3857, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 4.15, | |
| "learning_rate": 2.952755905511811e-05, | |
| "loss": 0.3543, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 4.15, | |
| "eval_bleu": 47.7916, | |
| "eval_gen_len": 18.4422, | |
| "eval_loss": 1.176469087600708, | |
| "eval_runtime": 519.0077, | |
| "eval_samples_per_second": 14.838, | |
| "eval_steps_per_second": 0.929, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 4.41, | |
| "learning_rate": 2.8215223097112863e-05, | |
| "loss": 0.2836, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 4.67, | |
| "learning_rate": 2.6902887139107612e-05, | |
| "loss": 0.2648, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 4.93, | |
| "learning_rate": 2.5590551181102364e-05, | |
| "loss": 0.2569, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 4.99, | |
| "eval_bleu": 48.1268, | |
| "eval_gen_len": 18.5385, | |
| "eval_loss": 1.2102879285812378, | |
| "eval_runtime": 526.7602, | |
| "eval_samples_per_second": 14.62, | |
| "eval_steps_per_second": 0.915, | |
| "step": 9600 | |
| }, | |
| { | |
| "epoch": 5.19, | |
| "learning_rate": 2.4278215223097113e-05, | |
| "loss": 0.2268, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 5.45, | |
| "learning_rate": 2.2965879265091865e-05, | |
| "loss": 0.1854, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 5.71, | |
| "learning_rate": 2.1653543307086614e-05, | |
| "loss": 0.1732, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 5.82, | |
| "eval_bleu": 48.9329, | |
| "eval_gen_len": 18.2085, | |
| "eval_loss": 1.25494384765625, | |
| "eval_runtime": 505.0437, | |
| "eval_samples_per_second": 15.248, | |
| "eval_steps_per_second": 0.954, | |
| "step": 11200 | |
| }, | |
| { | |
| "epoch": 5.97, | |
| "learning_rate": 2.0341207349081366e-05, | |
| "loss": 0.1693, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 6.23, | |
| "learning_rate": 1.9028871391076115e-05, | |
| "loss": 0.1453, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 6.49, | |
| "learning_rate": 1.7716535433070868e-05, | |
| "loss": 0.1228, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 6.65, | |
| "eval_bleu": 49.0248, | |
| "eval_gen_len": 18.2133, | |
| "eval_loss": 1.3022269010543823, | |
| "eval_runtime": 504.7977, | |
| "eval_samples_per_second": 15.256, | |
| "eval_steps_per_second": 0.955, | |
| "step": 12800 | |
| }, | |
| { | |
| "epoch": 6.75, | |
| "learning_rate": 1.6404199475065617e-05, | |
| "loss": 0.1158, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 7.01, | |
| "learning_rate": 1.5091863517060367e-05, | |
| "loss": 0.1144, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 7.27, | |
| "learning_rate": 1.377952755905512e-05, | |
| "loss": 0.0937, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 7.48, | |
| "eval_bleu": 49.3503, | |
| "eval_gen_len": 18.1673, | |
| "eval_loss": 1.317897081375122, | |
| "eval_runtime": 503.3739, | |
| "eval_samples_per_second": 15.299, | |
| "eval_steps_per_second": 0.958, | |
| "step": 14400 | |
| }, | |
| { | |
| "epoch": 7.53, | |
| "learning_rate": 1.246719160104987e-05, | |
| "loss": 0.0829, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 7.79, | |
| "learning_rate": 1.115485564304462e-05, | |
| "loss": 0.0783, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 8.05, | |
| "learning_rate": 9.842519685039371e-06, | |
| "loss": 0.0779, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 8.31, | |
| "learning_rate": 8.530183727034122e-06, | |
| "loss": 0.0627, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 8.31, | |
| "eval_bleu": 49.5551, | |
| "eval_gen_len": 18.2672, | |
| "eval_loss": 1.3408894538879395, | |
| "eval_runtime": 506.5726, | |
| "eval_samples_per_second": 15.202, | |
| "eval_steps_per_second": 0.951, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 8.57, | |
| "learning_rate": 7.2178477690288725e-06, | |
| "loss": 0.0579, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 8.83, | |
| "learning_rate": 5.905511811023622e-06, | |
| "loss": 0.0551, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 9.09, | |
| "learning_rate": 4.593175853018373e-06, | |
| "loss": 0.0558, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 9.14, | |
| "eval_bleu": 49.7808, | |
| "eval_gen_len": 18.2815, | |
| "eval_loss": 1.3544921875, | |
| "eval_runtime": 505.3645, | |
| "eval_samples_per_second": 15.239, | |
| "eval_steps_per_second": 0.954, | |
| "step": 17600 | |
| }, | |
| { | |
| "epoch": 9.35, | |
| "learning_rate": 3.2808398950131235e-06, | |
| "loss": 0.0456, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 9.61, | |
| "learning_rate": 1.968503937007874e-06, | |
| "loss": 0.0433, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 9.87, | |
| "learning_rate": 6.561679790026247e-07, | |
| "loss": 0.0442, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 9.97, | |
| "eval_bleu": 50.2114, | |
| "eval_gen_len": 18.2159, | |
| "eval_loss": 1.3559678792953491, | |
| "eval_runtime": 503.7841, | |
| "eval_samples_per_second": 15.286, | |
| "eval_steps_per_second": 0.957, | |
| "step": 19200 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 19250, | |
| "total_flos": 1.334951937048576e+18, | |
| "train_loss": 0.4018082245665711, | |
| "train_runtime": 39148.2976, | |
| "train_samples_per_second": 15.739, | |
| "train_steps_per_second": 0.492 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 19250, | |
| "num_train_epochs": 10, | |
| "save_steps": 1600, | |
| "total_flos": 1.334951937048576e+18, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |