| { | |
| "best_metric": 36.31328557707662, | |
| "best_model_checkpoint": "/root/turkic_qa/tr_uzn_models/orig_uzn_roberta_base_model/checkpoint-7310", | |
| "epoch": 10.0, | |
| "eval_steps": 500, | |
| "global_step": 7310, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "step": 731, | |
| "train_exact_match": 6.893106893106893, | |
| "train_f1": 13.147750410374615, | |
| "train_runtime": 13.1312, | |
| "train_samples_per_second": 113.851, | |
| "train_steps_per_second": 4.112 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 12.239034652709961, | |
| "learning_rate": 5e-06, | |
| "loss": 4.418, | |
| "step": 731 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 4.3125, | |
| "eval_f1": 11.281225208823463, | |
| "eval_runtime": 40.2248, | |
| "eval_samples_per_second": 113.935, | |
| "eval_steps_per_second": 4.077, | |
| "step": 731 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 1462, | |
| "train_exact_match": 15.484515484515484, | |
| "train_f1": 23.00008673777926, | |
| "train_runtime": 12.8965, | |
| "train_samples_per_second": 113.364, | |
| "train_steps_per_second": 4.11 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 13.470973014831543, | |
| "learning_rate": 1e-05, | |
| "loss": 3.3027, | |
| "step": 1462 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 10.5, | |
| "eval_f1": 18.60249294088304, | |
| "eval_runtime": 40.2237, | |
| "eval_samples_per_second": 113.938, | |
| "eval_steps_per_second": 4.077, | |
| "step": 1462 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 2193, | |
| "train_exact_match": 24.475524475524477, | |
| "train_f1": 33.75749140151479, | |
| "train_runtime": 12.4836, | |
| "train_samples_per_second": 113.83, | |
| "train_steps_per_second": 4.085 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 11.94182014465332, | |
| "learning_rate": 8.750000000000001e-06, | |
| "loss": 2.8501, | |
| "step": 2193 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 14.96875, | |
| "eval_f1": 23.95818931657813, | |
| "eval_runtime": 40.3346, | |
| "eval_samples_per_second": 113.625, | |
| "eval_steps_per_second": 4.066, | |
| "step": 2193 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 2924, | |
| "train_exact_match": 32.56743256743257, | |
| "train_f1": 41.56263168632173, | |
| "train_runtime": 12.6655, | |
| "train_samples_per_second": 112.826, | |
| "train_steps_per_second": 4.106 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 13.756119728088379, | |
| "learning_rate": 7.500000000000001e-06, | |
| "loss": 2.5536, | |
| "step": 2924 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 17.875, | |
| "eval_f1": 27.247899577169083, | |
| "eval_runtime": 40.1853, | |
| "eval_samples_per_second": 114.047, | |
| "eval_steps_per_second": 4.081, | |
| "step": 2924 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 3655, | |
| "train_exact_match": 36.76323676323676, | |
| "train_f1": 46.68974245865671, | |
| "train_runtime": 12.7455, | |
| "train_samples_per_second": 112.981, | |
| "train_steps_per_second": 4.08 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 15.806905746459961, | |
| "learning_rate": 6.25e-06, | |
| "loss": 2.3273, | |
| "step": 3655 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 19.75, | |
| "eval_f1": 29.976243167371322, | |
| "eval_runtime": 40.5485, | |
| "eval_samples_per_second": 113.025, | |
| "eval_steps_per_second": 4.045, | |
| "step": 3655 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "step": 4386, | |
| "train_exact_match": 45.45454545454545, | |
| "train_f1": 55.6255817809433, | |
| "train_runtime": 13.0466, | |
| "train_samples_per_second": 113.517, | |
| "train_steps_per_second": 4.062 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 22.134902954101562, | |
| "learning_rate": 5e-06, | |
| "loss": 2.111, | |
| "step": 4386 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 21.875, | |
| "eval_f1": 32.46367810578293, | |
| "eval_runtime": 40.5576, | |
| "eval_samples_per_second": 113.0, | |
| "eval_steps_per_second": 4.044, | |
| "step": 4386 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "step": 5117, | |
| "train_exact_match": 48.451548451548454, | |
| "train_f1": 59.11155707778988, | |
| "train_runtime": 12.7339, | |
| "train_samples_per_second": 112.456, | |
| "train_steps_per_second": 4.084 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 18.71537971496582, | |
| "learning_rate": 3.7500000000000005e-06, | |
| "loss": 1.9313, | |
| "step": 5117 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 23.34375, | |
| "eval_f1": 34.34464862509141, | |
| "eval_runtime": 40.5488, | |
| "eval_samples_per_second": 113.024, | |
| "eval_steps_per_second": 4.045, | |
| "step": 5117 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "step": 5848, | |
| "train_exact_match": 52.34765234765235, | |
| "train_f1": 61.4699737458471, | |
| "train_runtime": 12.2502, | |
| "train_samples_per_second": 112.978, | |
| "train_steps_per_second": 4.082 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 21.410564422607422, | |
| "learning_rate": 2.5e-06, | |
| "loss": 1.7907, | |
| "step": 5848 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 24.65625, | |
| "eval_f1": 35.76160110698192, | |
| "eval_runtime": 40.4412, | |
| "eval_samples_per_second": 113.325, | |
| "eval_steps_per_second": 4.055, | |
| "step": 5848 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "step": 6579, | |
| "train_exact_match": 56.34365634365634, | |
| "train_f1": 65.84711342159365, | |
| "train_runtime": 12.9755, | |
| "train_samples_per_second": 113.213, | |
| "train_steps_per_second": 4.085 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 18.834157943725586, | |
| "learning_rate": 1.25e-06, | |
| "loss": 1.6986, | |
| "step": 6579 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 24.8125, | |
| "eval_f1": 35.999416160714, | |
| "eval_runtime": 40.5359, | |
| "eval_samples_per_second": 113.06, | |
| "eval_steps_per_second": 4.046, | |
| "step": 6579 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 7310, | |
| "train_exact_match": 56.64335664335665, | |
| "train_f1": 66.49918737685148, | |
| "train_runtime": 12.1264, | |
| "train_samples_per_second": 113.224, | |
| "train_steps_per_second": 4.123 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 26.555368423461914, | |
| "learning_rate": 0.0, | |
| "loss": 1.6365, | |
| "step": 7310 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 24.9375, | |
| "eval_f1": 36.31328557707662, | |
| "eval_runtime": 40.4168, | |
| "eval_samples_per_second": 113.394, | |
| "eval_steps_per_second": 4.058, | |
| "step": 7310 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "step": 7310, | |
| "total_flos": 2.004675840340992e+16, | |
| "train_loss": 2.4619675162715886, | |
| "train_runtime": 2688.9862, | |
| "train_samples_per_second": 76.081, | |
| "train_steps_per_second": 2.718 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 7310, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 2.004675840340992e+16, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |