| { | |
| "best_metric": 78.92909907885071, | |
| "best_model_checkpoint": "/root/turkic_qa/tr_uzn_models/tr_uzn_xlm_roberta_base_model/checkpoint-7308", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 8120, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "step": 812, | |
| "train_exact_match": 28.97102897102897, | |
| "train_f1": 47.43836868671638, | |
| "train_runtime": 18.4368, | |
| "train_samples_per_second": 90.037, | |
| "train_steps_per_second": 3.254 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 62.82646942138672, | |
| "learning_rate": 5e-06, | |
| "loss": 4.1982, | |
| "step": 812 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 27.4375, | |
| "eval_f1": 46.68812257417232, | |
| "eval_runtime": 56.5084, | |
| "eval_samples_per_second": 90.447, | |
| "eval_steps_per_second": 3.238, | |
| "step": 812 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1624, | |
| "train_exact_match": 59.54045954045954, | |
| "train_f1": 75.73675798511488, | |
| "train_runtime": 18.1425, | |
| "train_samples_per_second": 89.679, | |
| "train_steps_per_second": 3.252 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 65.56416320800781, | |
| "learning_rate": 1e-05, | |
| "loss": 1.581, | |
| "step": 1624 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 52.5625, | |
| "eval_f1": 71.03718374348476, | |
| "eval_runtime": 56.6314, | |
| "eval_samples_per_second": 90.25, | |
| "eval_steps_per_second": 3.231, | |
| "step": 1624 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 2436, | |
| "train_exact_match": 66.03396603396604, | |
| "train_f1": 80.90860196742766, | |
| "train_runtime": 17.5378, | |
| "train_samples_per_second": 90.034, | |
| "train_steps_per_second": 3.25 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 56.48591613769531, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 1.0352, | |
| "step": 2436 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 59.1875, | |
| "eval_f1": 75.75121931796592, | |
| "eval_runtime": 56.5131, | |
| "eval_samples_per_second": 90.439, | |
| "eval_steps_per_second": 3.238, | |
| "step": 2436 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 3248, | |
| "train_exact_match": 70.92907092907093, | |
| "train_f1": 85.47419759277868, | |
| "train_runtime": 18.0502, | |
| "train_samples_per_second": 88.254, | |
| "train_steps_per_second": 3.158 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 23.379392623901367, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 0.8114, | |
| "step": 3248 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 61.375, | |
| "eval_f1": 77.2668497754134, | |
| "eval_runtime": 57.6273, | |
| "eval_samples_per_second": 88.691, | |
| "eval_steps_per_second": 3.176, | |
| "step": 3248 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 4060, | |
| "train_exact_match": 76.42357642357642, | |
| "train_f1": 87.93119243163716, | |
| "train_runtime": 18.0592, | |
| "train_samples_per_second": 88.376, | |
| "train_steps_per_second": 3.156 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 36.056884765625, | |
| "learning_rate": 6.25e-06, | |
| "loss": 0.6739, | |
| "step": 4060 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 62.0, | |
| "eval_f1": 77.40449528954417, | |
| "eval_runtime": 58.1956, | |
| "eval_samples_per_second": 87.824, | |
| "eval_steps_per_second": 3.145, | |
| "step": 4060 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 4872, | |
| "train_exact_match": 76.72327672327673, | |
| "train_f1": 89.14222582781746, | |
| "train_runtime": 18.7035, | |
| "train_samples_per_second": 87.203, | |
| "train_steps_per_second": 3.154 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 14.642334938049316, | |
| "learning_rate": 5e-06, | |
| "loss": 0.567, | |
| "step": 4872 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 62.71875, | |
| "eval_f1": 78.11173049836967, | |
| "eval_runtime": 57.9474, | |
| "eval_samples_per_second": 88.201, | |
| "eval_steps_per_second": 3.158, | |
| "step": 4872 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 5684, | |
| "train_exact_match": 78.52147852147853, | |
| "train_f1": 89.51285838266682, | |
| "train_runtime": 18.2594, | |
| "train_samples_per_second": 87.297, | |
| "train_steps_per_second": 3.122 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 17.107421875, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.4875, | |
| "step": 5684 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 62.875, | |
| "eval_f1": 78.08839073173891, | |
| "eval_runtime": 58.6231, | |
| "eval_samples_per_second": 87.184, | |
| "eval_steps_per_second": 3.122, | |
| "step": 5684 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 6496, | |
| "train_exact_match": 82.31768231768231, | |
| "train_f1": 91.06806991799819, | |
| "train_runtime": 17.4317, | |
| "train_samples_per_second": 88.23, | |
| "train_steps_per_second": 3.155 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 37.56465148925781, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.4315, | |
| "step": 6496 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 63.4375, | |
| "eval_f1": 78.70384115394582, | |
| "eval_runtime": 58.0199, | |
| "eval_samples_per_second": 88.09, | |
| "eval_steps_per_second": 3.154, | |
| "step": 6496 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 7308, | |
| "train_exact_match": 83.31668331668331, | |
| "train_f1": 92.69987725307878, | |
| "train_runtime": 18.674, | |
| "train_samples_per_second": 87.394, | |
| "train_steps_per_second": 3.159 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 52.06901168823242, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.3884, | |
| "step": 7308 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 63.78125, | |
| "eval_f1": 78.92909907885071, | |
| "eval_runtime": 58.0352, | |
| "eval_samples_per_second": 88.067, | |
| "eval_steps_per_second": 3.153, | |
| "step": 7308 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 8120, | |
| "train_exact_match": 82.61738261738262, | |
| "train_f1": 91.90869875798059, | |
| "train_runtime": 17.1824, | |
| "train_samples_per_second": 87.764, | |
| "train_steps_per_second": 3.143 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 24.732524871826172, | |
| "learning_rate": 0.0, | |
| "loss": 0.3607, | |
| "step": 8120 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 63.53125, | |
| "eval_f1": 78.82926814591693, | |
| "eval_runtime": 58.3589, | |
| "eval_samples_per_second": 87.579, | |
| "eval_steps_per_second": 3.136, | |
| "step": 8120 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 8120, | |
| "total_flos": 4.451320899376128e+16, | |
| "train_loss": 1.0534881366297528, | |
| "train_runtime": 5226.7321, | |
| "train_samples_per_second": 43.457, | |
| "train_steps_per_second": 1.554 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 8120, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 4.451320899376128e+16, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |