| { | |
| "best_metric": 78.05994502395909, | |
| "best_model_checkpoint": "/root/turkic_qa/ru_kaz_models/ru_kaz_xlm_roberta_base_squad_model/checkpoint-5520", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 5520, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "step": 552, | |
| "train_exact_match": 56.34365634365634, | |
| "train_f1": 75.2715442869406, | |
| "train_runtime": 11.3355, | |
| "train_samples_per_second": 89.63, | |
| "train_steps_per_second": 3.264 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 50.90898132324219, | |
| "learning_rate": 5e-06, | |
| "loss": 1.5648, | |
| "step": 552 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 55.0625, | |
| "eval_f1": 73.73760370127802, | |
| "eval_runtime": 36.5495, | |
| "eval_samples_per_second": 89.769, | |
| "eval_steps_per_second": 3.229, | |
| "step": 552 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1104, | |
| "train_exact_match": 66.93306693306694, | |
| "train_f1": 82.49314886706897, | |
| "train_runtime": 11.3179, | |
| "train_samples_per_second": 90.83, | |
| "train_steps_per_second": 3.269 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 55.97765350341797, | |
| "learning_rate": 1e-05, | |
| "loss": 1.2158, | |
| "step": 1104 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 58.84375, | |
| "eval_f1": 76.49071665255673, | |
| "eval_runtime": 36.2092, | |
| "eval_samples_per_second": 90.612, | |
| "eval_steps_per_second": 3.259, | |
| "step": 1104 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 1656, | |
| "train_exact_match": 67.53246753246754, | |
| "train_f1": 85.19091955779871, | |
| "train_runtime": 11.374, | |
| "train_samples_per_second": 89.415, | |
| "train_steps_per_second": 3.253 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 57.926448822021484, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 1.028, | |
| "step": 1656 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 60.0, | |
| "eval_f1": 77.43391533868831, | |
| "eval_runtime": 36.533, | |
| "eval_samples_per_second": 89.809, | |
| "eval_steps_per_second": 3.23, | |
| "step": 1656 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 2208, | |
| "train_exact_match": 74.22577422577423, | |
| "train_f1": 89.15395273332221, | |
| "train_runtime": 11.3619, | |
| "train_samples_per_second": 89.246, | |
| "train_steps_per_second": 3.256 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 38.46842956542969, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 0.8582, | |
| "step": 2208 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 60.625, | |
| "eval_f1": 77.72152196665945, | |
| "eval_runtime": 36.465, | |
| "eval_samples_per_second": 89.977, | |
| "eval_steps_per_second": 3.236, | |
| "step": 2208 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 2760, | |
| "train_exact_match": 77.12287712287713, | |
| "train_f1": 90.14502448136521, | |
| "train_runtime": 11.5016, | |
| "train_samples_per_second": 88.857, | |
| "train_steps_per_second": 3.217 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 14.753046989440918, | |
| "learning_rate": 6.25e-06, | |
| "loss": 0.7281, | |
| "step": 2760 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 60.8125, | |
| "eval_f1": 77.62510042626374, | |
| "eval_runtime": 36.4524, | |
| "eval_samples_per_second": 90.008, | |
| "eval_steps_per_second": 3.237, | |
| "step": 2760 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 3312, | |
| "train_exact_match": 80.21978021978022, | |
| "train_f1": 92.23780908752109, | |
| "train_runtime": 11.7336, | |
| "train_samples_per_second": 87.101, | |
| "train_steps_per_second": 3.153 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 23.36362075805664, | |
| "learning_rate": 5e-06, | |
| "loss": 0.6422, | |
| "step": 3312 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 61.28125, | |
| "eval_f1": 77.91507488796293, | |
| "eval_runtime": 37.4038, | |
| "eval_samples_per_second": 87.718, | |
| "eval_steps_per_second": 3.155, | |
| "step": 3312 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 3864, | |
| "train_exact_match": 81.81818181818181, | |
| "train_f1": 93.62881852209286, | |
| "train_runtime": 11.6143, | |
| "train_samples_per_second": 88.512, | |
| "train_steps_per_second": 3.186 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 21.251754760742188, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.5646, | |
| "step": 3864 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 61.1875, | |
| "eval_f1": 77.88113274187097, | |
| "eval_runtime": 36.8966, | |
| "eval_samples_per_second": 88.924, | |
| "eval_steps_per_second": 3.198, | |
| "step": 3864 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 4416, | |
| "train_exact_match": 84.31568431568432, | |
| "train_f1": 93.90553661277687, | |
| "train_runtime": 11.4894, | |
| "train_samples_per_second": 89.126, | |
| "train_steps_per_second": 3.22 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 31.64153289794922, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.5156, | |
| "step": 4416 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 60.71875, | |
| "eval_f1": 77.56837719428701, | |
| "eval_runtime": 36.6849, | |
| "eval_samples_per_second": 89.437, | |
| "eval_steps_per_second": 3.217, | |
| "step": 4416 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 4968, | |
| "train_exact_match": 85.11488511488511, | |
| "train_f1": 94.79917066124965, | |
| "train_runtime": 11.5536, | |
| "train_samples_per_second": 89.496, | |
| "train_steps_per_second": 3.202 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 46.01136779785156, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.4658, | |
| "step": 4968 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 61.125, | |
| "eval_f1": 77.88301584766332, | |
| "eval_runtime": 36.5869, | |
| "eval_samples_per_second": 89.677, | |
| "eval_steps_per_second": 3.225, | |
| "step": 4968 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 5520, | |
| "train_exact_match": 84.51548451548452, | |
| "train_f1": 94.34363228489812, | |
| "train_runtime": 11.395, | |
| "train_samples_per_second": 90.479, | |
| "train_steps_per_second": 3.247 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 84.91465759277344, | |
| "learning_rate": 0.0, | |
| "loss": 0.4488, | |
| "step": 5520 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 61.0625, | |
| "eval_f1": 78.05994502395909, | |
| "eval_runtime": 36.4836, | |
| "eval_samples_per_second": 89.931, | |
| "eval_steps_per_second": 3.234, | |
| "step": 5520 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 5520, | |
| "total_flos": 3.025228525300224e+16, | |
| "train_loss": 0.8031932664954144, | |
| "train_runtime": 3531.0877, | |
| "train_samples_per_second": 43.717, | |
| "train_steps_per_second": 1.563 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5520, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 3.025228525300224e+16, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |