| { | |
| "best_metric": 74.36086976945779, | |
| "best_model_checkpoint": "/root/turkic_qa/ru_uzn_models/ru_uzn_xlm_roberta_base_model/checkpoint-3906", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 5580, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "step": 558, | |
| "train_exact_match": 23.176823176823177, | |
| "train_f1": 40.91790695060761, | |
| "train_runtime": 11.8072, | |
| "train_samples_per_second": 89.437, | |
| "train_steps_per_second": 3.218 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 69.84556579589844, | |
| "learning_rate": 5e-06, | |
| "loss": 4.7931, | |
| "step": 558 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 23.375, | |
| "eval_f1": 40.588302053792845, | |
| "eval_runtime": 37.5548, | |
| "eval_samples_per_second": 89.443, | |
| "eval_steps_per_second": 3.195, | |
| "step": 558 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1116, | |
| "train_exact_match": 48.15184815184815, | |
| "train_f1": 67.62163572060659, | |
| "train_runtime": 11.6905, | |
| "train_samples_per_second": 89.047, | |
| "train_steps_per_second": 3.251 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 141.2274932861328, | |
| "learning_rate": 1e-05, | |
| "loss": 2.3399, | |
| "step": 1116 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 46.0, | |
| "eval_f1": 64.31534088055481, | |
| "eval_runtime": 37.2801, | |
| "eval_samples_per_second": 90.102, | |
| "eval_steps_per_second": 3.219, | |
| "step": 1116 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 1674, | |
| "train_exact_match": 59.74025974025974, | |
| "train_f1": 77.9890564127976, | |
| "train_runtime": 12.2542, | |
| "train_samples_per_second": 85.848, | |
| "train_steps_per_second": 3.101 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 176.3468475341797, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 1.5669, | |
| "step": 1674 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 53.71875, | |
| "eval_f1": 71.6046462048705, | |
| "eval_runtime": 38.2025, | |
| "eval_samples_per_second": 87.926, | |
| "eval_steps_per_second": 3.141, | |
| "step": 1674 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 2232, | |
| "train_exact_match": 60.93906093906094, | |
| "train_f1": 78.733011923745, | |
| "train_runtime": 11.7993, | |
| "train_samples_per_second": 88.649, | |
| "train_steps_per_second": 3.221 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 37.7187385559082, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 1.2585, | |
| "step": 2232 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 54.65625, | |
| "eval_f1": 72.55858557669187, | |
| "eval_runtime": 37.5502, | |
| "eval_samples_per_second": 89.454, | |
| "eval_steps_per_second": 3.196, | |
| "step": 2232 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 2790, | |
| "train_exact_match": 67.03296703296704, | |
| "train_f1": 83.4692833669754, | |
| "train_runtime": 11.75, | |
| "train_samples_per_second": 88.681, | |
| "train_steps_per_second": 3.234 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 65.70742797851562, | |
| "learning_rate": 6.25e-06, | |
| "loss": 1.0746, | |
| "step": 2790 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 56.3125, | |
| "eval_f1": 73.3893893291572, | |
| "eval_runtime": 37.5867, | |
| "eval_samples_per_second": 89.367, | |
| "eval_steps_per_second": 3.193, | |
| "step": 2790 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 3348, | |
| "train_exact_match": 71.42857142857143, | |
| "train_f1": 86.60157934407324, | |
| "train_runtime": 11.7396, | |
| "train_samples_per_second": 88.93, | |
| "train_steps_per_second": 3.237 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 64.60488891601562, | |
| "learning_rate": 5e-06, | |
| "loss": 0.9341, | |
| "step": 3348 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 57.09375, | |
| "eval_f1": 73.83274981499065, | |
| "eval_runtime": 37.5525, | |
| "eval_samples_per_second": 89.448, | |
| "eval_steps_per_second": 3.196, | |
| "step": 3348 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 3906, | |
| "train_exact_match": 74.22577422577423, | |
| "train_f1": 88.79932331084886, | |
| "train_runtime": 11.745, | |
| "train_samples_per_second": 88.548, | |
| "train_steps_per_second": 3.235 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 63.03971862792969, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 0.8362, | |
| "step": 3906 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 56.9375, | |
| "eval_f1": 74.36086976945779, | |
| "eval_runtime": 37.3638, | |
| "eval_samples_per_second": 89.9, | |
| "eval_steps_per_second": 3.212, | |
| "step": 3906 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 4464, | |
| "train_exact_match": 75.22477522477523, | |
| "train_f1": 89.24938452782732, | |
| "train_runtime": 11.7145, | |
| "train_samples_per_second": 89.206, | |
| "train_steps_per_second": 3.244 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 115.92487335205078, | |
| "learning_rate": 2.5e-06, | |
| "loss": 0.7519, | |
| "step": 4464 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 56.875, | |
| "eval_f1": 73.74170941323764, | |
| "eval_runtime": 37.4071, | |
| "eval_samples_per_second": 89.796, | |
| "eval_steps_per_second": 3.208, | |
| "step": 4464 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 5022, | |
| "train_exact_match": 77.92207792207792, | |
| "train_f1": 90.77887866843285, | |
| "train_runtime": 11.7653, | |
| "train_samples_per_second": 88.735, | |
| "train_steps_per_second": 3.23 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 67.61051177978516, | |
| "learning_rate": 1.25e-06, | |
| "loss": 0.7002, | |
| "step": 5022 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 57.09375, | |
| "eval_f1": 73.99562834029655, | |
| "eval_runtime": 37.3049, | |
| "eval_samples_per_second": 90.042, | |
| "eval_steps_per_second": 3.217, | |
| "step": 5022 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 5580, | |
| "train_exact_match": 76.82317682317682, | |
| "train_f1": 90.5111904232365, | |
| "train_runtime": 11.6949, | |
| "train_samples_per_second": 89.526, | |
| "train_steps_per_second": 3.249 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 77.0350341796875, | |
| "learning_rate": 0.0, | |
| "loss": 0.6664, | |
| "step": 5580 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 57.09375, | |
| "eval_f1": 74.00492648190163, | |
| "eval_runtime": 37.3623, | |
| "eval_samples_per_second": 89.904, | |
| "eval_steps_per_second": 3.212, | |
| "step": 5580 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 5580, | |
| "total_flos": 3.056976081243648e+16, | |
| "train_loss": 1.4921705567281307, | |
| "train_runtime": 3588.4893, | |
| "train_samples_per_second": 43.47, | |
| "train_steps_per_second": 1.555 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5580, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 3.056976081243648e+16, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |