| { | |
| "best_metric": 76.43196644229354, | |
| "best_model_checkpoint": "/root/turkic_qa/en_kaz_models/en_kaz_mdeberta_base_squad_model/checkpoint-2764", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 6910, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "step": 691, | |
| "train_exact_match": 63.03696303696304, | |
| "train_f1": 78.13915659047835, | |
| "train_runtime": 17.4035, | |
| "train_samples_per_second": 74.468, | |
| "train_steps_per_second": 2.701 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 40.22502899169922, | |
| "learning_rate": 5e-06, | |
| "loss": 1.2642, | |
| "step": 691 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 58.5625, | |
| "eval_f1": 74.11433052314848, | |
| "eval_runtime": 54.8873, | |
| "eval_samples_per_second": 74.626, | |
| "eval_steps_per_second": 2.678, | |
| "step": 691 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1382, | |
| "train_exact_match": 67.33266733266733, | |
| "train_f1": 80.90558473928773, | |
| "train_runtime": 17.2817, | |
| "train_samples_per_second": 75.745, | |
| "train_steps_per_second": 2.72 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 28.721803665161133, | |
| "learning_rate": 1e-05, | |
| "loss": 1.0309, | |
| "step": 1382 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 60.84375, | |
| "eval_f1": 75.39680273889411, | |
| "eval_runtime": 54.2212, | |
| "eval_samples_per_second": 75.542, | |
| "eval_steps_per_second": 2.711, | |
| "step": 1382 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 2073, | |
| "train_exact_match": 72.22777222777223, | |
| "train_f1": 86.54756181813792, | |
| "train_runtime": 16.9462, | |
| "train_samples_per_second": 74.943, | |
| "train_steps_per_second": 2.714 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 25.19014549255371, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 0.8566, | |
| "step": 2073 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 62.1875, | |
| "eval_f1": 76.29400510573826, | |
| "eval_runtime": 54.2582, | |
| "eval_samples_per_second": 75.491, | |
| "eval_steps_per_second": 2.709, | |
| "step": 2073 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 2764, | |
| "train_exact_match": 75.52447552447552, | |
| "train_f1": 88.34455431353304, | |
| "train_runtime": 17.9414, | |
| "train_samples_per_second": 75.078, | |
| "train_steps_per_second": 2.731 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 31.350345611572266, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 0.7007, | |
| "step": 2764 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 62.21875, | |
| "eval_f1": 76.43196644229354, | |
| "eval_runtime": 54.0444, | |
| "eval_samples_per_second": 75.79, | |
| "eval_steps_per_second": 2.72, | |
| "step": 2764 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 3455, | |
| "train_exact_match": 80.01998001998003, | |
| "train_f1": 91.67585411210774, | |
| "train_runtime": 17.0774, | |
| "train_samples_per_second": 75.246, | |
| "train_steps_per_second": 2.694 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 41.42879867553711, | |
| "learning_rate": 6.25e-06, | |
| "loss": 0.5836, | |
| "step": 3455 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 62.0, | |
| "eval_f1": 75.89595740381746, | |
| "eval_runtime": 54.8769, | |
| "eval_samples_per_second": 74.64, | |
| "eval_steps_per_second": 2.679, | |
| "step": 3455 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 4146, | |
| "train_exact_match": 81.71828171828172, | |
| "train_f1": 92.46355571003183, | |
| "train_runtime": 16.6375, | |
| "train_samples_per_second": 75.252, | |
| "train_steps_per_second": 2.705 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 20.05802345275879, | |
| "learning_rate": 5e-06, | |
| "loss": 0.494, | |
| "step": 4146 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 62.46875, | |
| "eval_f1": 75.95799025269356, | |
| "eval_runtime": 54.1924, | |
| "eval_samples_per_second": 75.583, | |
| "eval_steps_per_second": 2.713, | |
| "step": 4146 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 4837, | |
| "train_exact_match": 82.11788211788212, | |
| "train_f1": 93.12695306139514, | |
| "train_runtime": 17.1308, | |
| "train_samples_per_second": 74.953, | |
| "train_steps_per_second": 2.685 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 36.17577362060547, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.4281, | |
| "step": 4837 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 62.09375, | |
| "eval_f1": 75.87457028727934, | |
| "eval_runtime": 54.4468, | |
| "eval_samples_per_second": 75.229, | |
| "eval_steps_per_second": 2.7, | |
| "step": 4837 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 5528, | |
| "train_exact_match": 86.01398601398601, | |
| "train_f1": 94.42525886876679, | |
| "train_runtime": 17.0114, | |
| "train_samples_per_second": 75.479, | |
| "train_steps_per_second": 2.704 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 12.529890060424805, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.3806, | |
| "step": 5528 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 62.0625, | |
| "eval_f1": 75.78487868224668, | |
| "eval_runtime": 54.4766, | |
| "eval_samples_per_second": 75.188, | |
| "eval_steps_per_second": 2.698, | |
| "step": 5528 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 6219, | |
| "train_exact_match": 86.11388611388611, | |
| "train_f1": 94.8958657040008, | |
| "train_runtime": 17.878, | |
| "train_samples_per_second": 72.1, | |
| "train_steps_per_second": 2.629 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 37.82954788208008, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.3457, | |
| "step": 6219 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 62.03125, | |
| "eval_f1": 75.65181251557605, | |
| "eval_runtime": 54.6102, | |
| "eval_samples_per_second": 75.004, | |
| "eval_steps_per_second": 2.692, | |
| "step": 6219 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 6910, | |
| "train_exact_match": 87.81218781218782, | |
| "train_f1": 95.32192987518101, | |
| "train_runtime": 16.6546, | |
| "train_samples_per_second": 74.814, | |
| "train_steps_per_second": 2.702 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 8.968611717224121, | |
| "learning_rate": 0.0, | |
| "loss": 0.3243, | |
| "step": 6910 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 61.9375, | |
| "eval_f1": 75.6746197198531, | |
| "eval_runtime": 54.8671, | |
| "eval_samples_per_second": 74.653, | |
| "eval_steps_per_second": 2.679, | |
| "step": 6910 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 6910, | |
| "total_flos": 3.788022162461184e+16, | |
| "train_loss": 0.6408887451878504, | |
| "train_runtime": 5720.422, | |
| "train_samples_per_second": 33.789, | |
| "train_steps_per_second": 1.208 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 6910, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 3.788022162461184e+16, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |