{ "best_metric": 70.69724510197014, "best_model_checkpoint": "/home/jcanete/ft-data/all_results/mlqa/roberta_large_bne/epochs_3_bs_32_lr_1e-5/checkpoint-2400", "epoch": 3.0, "global_step": 7695, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.12, "eval_exact_match": 38.0, "eval_f1": 62.39487322160937, "step": 300 }, { "epoch": 0.19, "learning_rate": 9.35022742040286e-06, "loss": 2.4203, "step": 500 }, { "epoch": 0.23, "eval_exact_match": 40.6, "eval_f1": 65.9788457320324, "step": 600 }, { "epoch": 0.35, "eval_exact_match": 43.4, "eval_f1": 67.55424411452253, "step": 900 }, { "epoch": 0.39, "learning_rate": 8.700454840805719e-06, "loss": 1.913, "step": 1000 }, { "epoch": 0.47, "eval_exact_match": 43.8, "eval_f1": 67.76261698067435, "step": 1200 }, { "epoch": 0.58, "learning_rate": 8.050682261208577e-06, "loss": 1.7856, "step": 1500 }, { "epoch": 0.58, "eval_exact_match": 44.0, "eval_f1": 68.3763816678372, "step": 1500 }, { "epoch": 0.7, "eval_exact_match": 41.4, "eval_f1": 66.69667249640099, "step": 1800 }, { "epoch": 0.78, "learning_rate": 7.400909681611437e-06, "loss": 1.7603, "step": 2000 }, { "epoch": 0.82, "eval_exact_match": 43.4, "eval_f1": 68.6773344322754, "step": 2100 }, { "epoch": 0.94, "eval_exact_match": 46.6, "eval_f1": 70.69724510197014, "step": 2400 }, { "epoch": 0.97, "learning_rate": 6.7524366471734905e-06, "loss": 1.6929, "step": 2500 }, { "epoch": 1.05, "eval_exact_match": 46.0, "eval_f1": 69.30388852674767, "step": 2700 }, { "epoch": 1.17, "learning_rate": 6.103963612735543e-06, "loss": 1.3491, "step": 3000 }, { "epoch": 1.17, "eval_exact_match": 43.0, "eval_f1": 68.61157350646901, "step": 3000 }, { "epoch": 1.29, "eval_exact_match": 44.0, "eval_f1": 68.69707838640173, "step": 3300 }, { "epoch": 1.36, "learning_rate": 5.454191033138402e-06, "loss": 1.2655, "step": 3500 }, { "epoch": 1.4, "eval_exact_match": 42.2, "eval_f1": 67.63235840167032, "step": 3600 }, { "epoch": 1.52, "eval_exact_match": 44.6, "eval_f1": 68.81188025343606, "step": 3900 }, { "epoch": 1.56, "learning_rate": 4.804418453541261e-06, "loss": 1.2769, "step": 4000 }, { "epoch": 1.64, "eval_exact_match": 41.8, "eval_f1": 67.97025725384887, "step": 4200 }, { "epoch": 1.75, "learning_rate": 4.15464587394412e-06, "loss": 1.2644, "step": 4500 }, { "epoch": 1.75, "eval_exact_match": 41.4, "eval_f1": 67.39389717590865, "step": 4500 }, { "epoch": 1.87, "eval_exact_match": 42.2, "eval_f1": 67.0843891712357, "step": 4800 }, { "epoch": 1.95, "learning_rate": 3.5061728395061736e-06, "loss": 1.2759, "step": 5000 }, { "epoch": 1.99, "eval_exact_match": 44.4, "eval_f1": 68.50517059240434, "step": 5100 }, { "epoch": 2.11, "eval_exact_match": 41.8, "eval_f1": 66.50371716570872, "step": 5400 }, { "epoch": 2.14, "learning_rate": 2.8564002599090316e-06, "loss": 0.9914, "step": 5500 }, { "epoch": 2.22, "eval_exact_match": 38.6, "eval_f1": 65.02028461028927, "step": 5700 }, { "epoch": 2.34, "learning_rate": 2.206627680311891e-06, "loss": 0.8708, "step": 6000 }, { "epoch": 2.34, "eval_exact_match": 39.6, "eval_f1": 66.20180637430703, "step": 6000 }, { "epoch": 2.46, "eval_exact_match": 39.8, "eval_f1": 66.64472686734118, "step": 6300 }, { "epoch": 2.53, "learning_rate": 1.5581546458739443e-06, "loss": 0.8511, "step": 6500 }, { "epoch": 2.57, "eval_exact_match": 40.8, "eval_f1": 67.51042164555922, "step": 6600 }, { "epoch": 2.69, "eval_exact_match": 40.6, "eval_f1": 67.41323327360205, "step": 6900 }, { "epoch": 2.73, "learning_rate": 9.083820662768032e-07, "loss": 0.8376, "step": 7000 }, { "epoch": 2.81, "eval_exact_match": 40.4, "eval_f1": 66.72964354261293, "step": 7200 }, { "epoch": 2.92, "learning_rate": 2.5860948667966215e-07, "loss": 0.845, "step": 7500 }, { "epoch": 2.92, "eval_exact_match": 40.2, "eval_f1": 66.87800892788201, "step": 7500 }, { "epoch": 3.0, "step": 7695, "total_flos": 1.7670689327303856e+17, "train_loss": 1.3470741752837678, "train_runtime": 3160.2174, "train_samples_per_second": 77.899, "train_steps_per_second": 2.435 } ], "max_steps": 7695, "num_train_epochs": 3, "total_flos": 1.7670689327303856e+17, "trial_name": null, "trial_params": null }