| { | |
| "best_metric": 73.01711252807637, | |
| "best_model_checkpoint": "/root/turkic_qa/en_uzn_models/en_uzn_xlm_roberta_base_squad_model/checkpoint-4140", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 6900, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "step": 690, | |
| "train_exact_match": 53.84615384615385, | |
| "train_f1": 69.79095478605717, | |
| "train_runtime": 14.0504, | |
| "train_samples_per_second": 89.82, | |
| "train_steps_per_second": 3.274 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 19.47218894958496, | |
| "learning_rate": 5e-06, | |
| "loss": 1.6813, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 52.0625, | |
| "eval_f1": 67.50170309345776, | |
| "eval_runtime": 44.891, | |
| "eval_samples_per_second": 90.263, | |
| "eval_steps_per_second": 3.23, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1380, | |
| "train_exact_match": 63.93606393606394, | |
| "train_f1": 76.38693770706172, | |
| "train_runtime": 14.2008, | |
| "train_samples_per_second": 89.643, | |
| "train_steps_per_second": 3.239 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 45.075042724609375, | |
| "learning_rate": 1e-05, | |
| "loss": 1.3266, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 56.15625, | |
| "eval_f1": 70.83430989837035, | |
| "eval_runtime": 45.0891, | |
| "eval_samples_per_second": 89.867, | |
| "eval_steps_per_second": 3.216, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 2070, | |
| "train_exact_match": 67.43256743256744, | |
| "train_f1": 80.45847460323496, | |
| "train_runtime": 14.4622, | |
| "train_samples_per_second": 89.544, | |
| "train_steps_per_second": 3.25 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 17.318368911743164, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 1.1036, | |
| "step": 2070 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 57.46875, | |
| "eval_f1": 71.70302249821802, | |
| "eval_runtime": 44.9476, | |
| "eval_samples_per_second": 90.149, | |
| "eval_steps_per_second": 3.226, | |
| "step": 2070 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 2760, | |
| "train_exact_match": 71.72827172827172, | |
| "train_f1": 84.90184305868037, | |
| "train_runtime": 14.4996, | |
| "train_samples_per_second": 89.313, | |
| "train_steps_per_second": 3.241 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 217.5539093017578, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 0.9255, | |
| "step": 2760 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 58.34375, | |
| "eval_f1": 72.47808689784853, | |
| "eval_runtime": 44.8378, | |
| "eval_samples_per_second": 90.37, | |
| "eval_steps_per_second": 3.234, | |
| "step": 2760 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 3450, | |
| "train_exact_match": 76.32367632367632, | |
| "train_f1": 88.37491424418958, | |
| "train_runtime": 14.4084, | |
| "train_samples_per_second": 85.992, | |
| "train_steps_per_second": 3.123 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 394.8098449707031, | |
| "learning_rate": 6.25e-06, | |
| "loss": 0.7806, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 58.21875, | |
| "eval_f1": 72.48290373070347, | |
| "eval_runtime": 45.4975, | |
| "eval_samples_per_second": 89.06, | |
| "eval_steps_per_second": 3.187, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 4140, | |
| "train_exact_match": 78.72127872127872, | |
| "train_f1": 88.97987588000849, | |
| "train_runtime": 14.2096, | |
| "train_samples_per_second": 88.813, | |
| "train_steps_per_second": 3.237 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 17.177759170532227, | |
| "learning_rate": 5e-06, | |
| "loss": 0.6664, | |
| "step": 4140 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 59.0, | |
| "eval_f1": 73.01711252807637, | |
| "eval_runtime": 45.351, | |
| "eval_samples_per_second": 89.348, | |
| "eval_steps_per_second": 3.197, | |
| "step": 4140 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 4830, | |
| "train_exact_match": 82.51748251748252, | |
| "train_f1": 91.72028497776134, | |
| "train_runtime": 13.942, | |
| "train_samples_per_second": 89.37, | |
| "train_steps_per_second": 3.228 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 354.42584228515625, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.5911, | |
| "step": 4830 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 58.53125, | |
| "eval_f1": 72.55506746037935, | |
| "eval_runtime": 45.1047, | |
| "eval_samples_per_second": 89.835, | |
| "eval_steps_per_second": 3.215, | |
| "step": 4830 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 5520, | |
| "train_exact_match": 81.91808191808192, | |
| "train_f1": 91.29001859546244, | |
| "train_runtime": 14.3482, | |
| "train_samples_per_second": 89.14, | |
| "train_steps_per_second": 3.206 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 184.72279357910156, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.5247, | |
| "step": 5520 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 58.25, | |
| "eval_f1": 72.07621980566323, | |
| "eval_runtime": 45.0133, | |
| "eval_samples_per_second": 90.018, | |
| "eval_steps_per_second": 3.221, | |
| "step": 5520 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 6210, | |
| "train_exact_match": 83.21678321678321, | |
| "train_f1": 92.6376023057958, | |
| "train_runtime": 14.5039, | |
| "train_samples_per_second": 89.355, | |
| "train_steps_per_second": 3.24 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 3.3138372898101807, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.4735, | |
| "step": 6210 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 58.25, | |
| "eval_f1": 72.13528491061855, | |
| "eval_runtime": 45.1402, | |
| "eval_samples_per_second": 89.765, | |
| "eval_steps_per_second": 3.212, | |
| "step": 6210 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 6900, | |
| "train_exact_match": 86.81318681318682, | |
| "train_f1": 94.26051176344673, | |
| "train_runtime": 14.1643, | |
| "train_samples_per_second": 89.309, | |
| "train_steps_per_second": 3.248 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 12.08403491973877, | |
| "learning_rate": 0.0, | |
| "loss": 0.4451, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 58.25, | |
| "eval_f1": 72.1039896331836, | |
| "eval_runtime": 45.0152, | |
| "eval_samples_per_second": 90.014, | |
| "eval_steps_per_second": 3.221, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 6900, | |
| "total_flos": 3.780898745780736e+16, | |
| "train_loss": 0.8518442347429801, | |
| "train_runtime": 4369.5998, | |
| "train_samples_per_second": 44.153, | |
| "train_steps_per_second": 1.579 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 6900, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 3.780898745780736e+16, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |