| { | |
| "best_metric": 69.86920835599972, | |
| "best_model_checkpoint": "/root/turkic_qa/en_uzn_models/en_uzn_xlm_roberta_base_model/checkpoint-6900", | |
| "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": 23.076923076923077, | |
| "train_f1": 37.566079561862445, | |
| "train_runtime": 14.0841, | |
| "train_samples_per_second": 89.604, | |
| "train_steps_per_second": 3.266 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 116.39849853515625, | |
| "learning_rate": 5e-06, | |
| "loss": 4.5006, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 21.46875, | |
| "eval_f1": 34.27477969652227, | |
| "eval_runtime": 44.8553, | |
| "eval_samples_per_second": 90.335, | |
| "eval_steps_per_second": 3.233, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1380, | |
| "train_exact_match": 51.04895104895105, | |
| "train_f1": 64.28880003049197, | |
| "train_runtime": 14.6272, | |
| "train_samples_per_second": 87.03, | |
| "train_steps_per_second": 3.145 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 242.94326782226562, | |
| "learning_rate": 1e-05, | |
| "loss": 2.1628, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 45.46875, | |
| "eval_f1": 60.28283429101314, | |
| "eval_runtime": 46.3266, | |
| "eval_samples_per_second": 87.466, | |
| "eval_steps_per_second": 3.13, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 2070, | |
| "train_exact_match": 57.142857142857146, | |
| "train_f1": 71.92623291529159, | |
| "train_runtime": 14.5466, | |
| "train_samples_per_second": 89.024, | |
| "train_steps_per_second": 3.231 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 18.747507095336914, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 1.4946, | |
| "step": 2070 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 52.40625, | |
| "eval_f1": 67.09147147337285, | |
| "eval_runtime": 45.3354, | |
| "eval_samples_per_second": 89.378, | |
| "eval_steps_per_second": 3.198, | |
| "step": 2070 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 2760, | |
| "train_exact_match": 64.13586413586414, | |
| "train_f1": 78.3731582090412, | |
| "train_runtime": 14.5689, | |
| "train_samples_per_second": 88.888, | |
| "train_steps_per_second": 3.226 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 269.3212890625, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 1.2186, | |
| "step": 2760 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 53.96875, | |
| "eval_f1": 68.38825732557214, | |
| "eval_runtime": 45.3562, | |
| "eval_samples_per_second": 89.337, | |
| "eval_steps_per_second": 3.197, | |
| "step": 2760 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 3450, | |
| "train_exact_match": 68.33166833166833, | |
| "train_f1": 82.36153774583077, | |
| "train_runtime": 13.878, | |
| "train_samples_per_second": 89.278, | |
| "train_steps_per_second": 3.243 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 259.8070068359375, | |
| "learning_rate": 6.25e-06, | |
| "loss": 1.0359, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 54.96875, | |
| "eval_f1": 69.02164576043303, | |
| "eval_runtime": 45.1308, | |
| "eval_samples_per_second": 89.783, | |
| "eval_steps_per_second": 3.213, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 4140, | |
| "train_exact_match": 71.92807192807193, | |
| "train_f1": 83.477467208777, | |
| "train_runtime": 14.0926, | |
| "train_samples_per_second": 89.55, | |
| "train_steps_per_second": 3.264 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 39.62363052368164, | |
| "learning_rate": 5e-06, | |
| "loss": 0.8945, | |
| "step": 4140 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 55.125, | |
| "eval_f1": 69.56052410993578, | |
| "eval_runtime": 44.9507, | |
| "eval_samples_per_second": 90.143, | |
| "eval_steps_per_second": 3.226, | |
| "step": 4140 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 4830, | |
| "train_exact_match": 75.62437562437563, | |
| "train_f1": 86.36686612245309, | |
| "train_runtime": 14.0278, | |
| "train_samples_per_second": 88.823, | |
| "train_steps_per_second": 3.208 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 122.51107788085938, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.7916, | |
| "step": 4830 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 55.4375, | |
| "eval_f1": 69.46738777882736, | |
| "eval_runtime": 45.1053, | |
| "eval_samples_per_second": 89.834, | |
| "eval_steps_per_second": 3.215, | |
| "step": 4830 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 5520, | |
| "train_exact_match": 75.92407592407592, | |
| "train_f1": 86.72722282271204, | |
| "train_runtime": 14.2506, | |
| "train_samples_per_second": 89.75, | |
| "train_steps_per_second": 3.228 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 202.73541259765625, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.7119, | |
| "step": 5520 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 56.0, | |
| "eval_f1": 69.69766614851221, | |
| "eval_runtime": 45.2302, | |
| "eval_samples_per_second": 89.586, | |
| "eval_steps_per_second": 3.206, | |
| "step": 5520 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 6210, | |
| "train_exact_match": 77.32267732267732, | |
| "train_f1": 88.72058931170982, | |
| "train_runtime": 14.7856, | |
| "train_samples_per_second": 87.653, | |
| "train_steps_per_second": 3.179 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 5.172795295715332, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.6445, | |
| "step": 6210 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 55.90625, | |
| "eval_f1": 69.57114295943043, | |
| "eval_runtime": 45.9508, | |
| "eval_samples_per_second": 88.181, | |
| "eval_steps_per_second": 3.156, | |
| "step": 6210 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 6900, | |
| "train_exact_match": 80.81918081918081, | |
| "train_f1": 90.37663245353303, | |
| "train_runtime": 14.2722, | |
| "train_samples_per_second": 88.634, | |
| "train_steps_per_second": 3.223 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 6.113234996795654, | |
| "learning_rate": 0.0, | |
| "loss": 0.6132, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 56.09375, | |
| "eval_f1": 69.86920835599972, | |
| "eval_runtime": 44.9954, | |
| "eval_samples_per_second": 90.054, | |
| "eval_steps_per_second": 3.223, | |
| "step": 6900 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 6900, | |
| "total_flos": 3.780898745780736e+16, | |
| "train_loss": 1.4068151059358016, | |
| "train_runtime": 4389.3737, | |
| "train_samples_per_second": 43.954, | |
| "train_steps_per_second": 1.572 | |
| } | |
| ], | |
| "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 | |
| } | |