| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 10.0, | |
| "eval_steps": 248, | |
| "global_step": 2480, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 3e-05, | |
| "loss": 2.8428, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3e-05, | |
| "loss": 2.1589, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 3e-05, | |
| "loss": 2.0527, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 3e-05, | |
| "loss": 1.9823, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.6398457059679767, | |
| "eval_loss": 1.6149228811264038, | |
| "eval_runtime": 17.4504, | |
| "eval_samples_per_second": 28.653, | |
| "eval_steps_per_second": 1.834, | |
| "step": 248 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_exact_match": 9.0, | |
| "eval_f1": 13.45666666666666, | |
| "eval_qa_bleu": 7.063432441196736, | |
| "eval_qa_exact_match": 0.09, | |
| "eval_recite_bleu": 15.344658855358656, | |
| "eval_recite_exact_match": 0.0, | |
| "step": 248 | |
| }, | |
| { | |
| "epoch": 1.01, | |
| "learning_rate": 3e-05, | |
| "loss": 1.9153, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.21, | |
| "learning_rate": 3e-05, | |
| "loss": 1.49, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "learning_rate": 3e-05, | |
| "loss": 1.4406, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.61, | |
| "learning_rate": 3e-05, | |
| "loss": 1.4161, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.81, | |
| "learning_rate": 3e-05, | |
| "loss": 1.3868, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.6700320232896652, | |
| "eval_loss": 1.1928881406784058, | |
| "eval_runtime": 15.4486, | |
| "eval_samples_per_second": 32.365, | |
| "eval_steps_per_second": 2.071, | |
| "step": 496 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_exact_match": 9.2, | |
| "eval_f1": 14.735098500533276, | |
| "eval_qa_bleu": 7.255412763224127, | |
| "eval_qa_exact_match": 0.088, | |
| "eval_recite_bleu": 17.979965007898493, | |
| "eval_recite_exact_match": 0.0, | |
| "step": 496 | |
| }, | |
| { | |
| "epoch": 2.02, | |
| "learning_rate": 3e-05, | |
| "loss": 1.309, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "learning_rate": 3e-05, | |
| "loss": 0.9401, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 2.42, | |
| "learning_rate": 3e-05, | |
| "loss": 0.9429, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "learning_rate": 3e-05, | |
| "loss": 0.9115, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "learning_rate": 3e-05, | |
| "loss": 0.9102, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.6992489082969432, | |
| "eval_loss": 0.8513368368148804, | |
| "eval_runtime": 15.7421, | |
| "eval_samples_per_second": 31.762, | |
| "eval_steps_per_second": 2.033, | |
| "step": 744 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_exact_match": 10.2, | |
| "eval_f1": 15.888450046685344, | |
| "eval_qa_bleu": 7.2379640900667646, | |
| "eval_qa_exact_match": 0.098, | |
| "eval_recite_bleu": 20.876580135938546, | |
| "eval_recite_exact_match": 0.0, | |
| "step": 744 | |
| }, | |
| { | |
| "epoch": 3.02, | |
| "learning_rate": 3e-05, | |
| "loss": 0.8302, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 3.23, | |
| "learning_rate": 3e-05, | |
| "loss": 0.5813, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.43, | |
| "learning_rate": 3e-05, | |
| "loss": 0.5844, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 3.63, | |
| "learning_rate": 3e-05, | |
| "loss": 0.5776, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 3.83, | |
| "learning_rate": 3e-05, | |
| "loss": 0.5869, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.721787481804949, | |
| "eval_loss": 0.6180558204650879, | |
| "eval_runtime": 15.5183, | |
| "eval_samples_per_second": 32.22, | |
| "eval_steps_per_second": 2.062, | |
| "step": 992 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_exact_match": 13.6, | |
| "eval_f1": 19.457685255920552, | |
| "eval_qa_bleu": 7.372812651620524, | |
| "eval_qa_exact_match": 0.126, | |
| "eval_recite_bleu": 24.119469757278573, | |
| "eval_recite_exact_match": 0.0, | |
| "step": 992 | |
| }, | |
| { | |
| "epoch": 4.03, | |
| "learning_rate": 3e-05, | |
| "loss": 0.5394, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.23, | |
| "learning_rate": 3e-05, | |
| "loss": 0.3787, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "learning_rate": 3e-05, | |
| "loss": 0.3865, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 4.64, | |
| "learning_rate": 3e-05, | |
| "loss": 0.3734, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 4.84, | |
| "learning_rate": 3e-05, | |
| "loss": 0.364, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.7352052401746725, | |
| "eval_loss": 0.49626532196998596, | |
| "eval_runtime": 15.762, | |
| "eval_samples_per_second": 31.722, | |
| "eval_steps_per_second": 2.03, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_exact_match": 11.6, | |
| "eval_f1": 17.050505050505052, | |
| "eval_qa_bleu": 9.3668737780184, | |
| "eval_qa_exact_match": 0.114, | |
| "eval_recite_bleu": 28.6471128712922, | |
| "eval_recite_exact_match": 0.004, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 5.04, | |
| "learning_rate": 3e-05, | |
| "loss": 0.345, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 5.24, | |
| "learning_rate": 3e-05, | |
| "loss": 0.2508, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 5.44, | |
| "learning_rate": 3e-05, | |
| "loss": 0.26, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 5.65, | |
| "learning_rate": 3e-05, | |
| "loss": 0.2618, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 5.85, | |
| "learning_rate": 3e-05, | |
| "loss": 0.2706, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.7419825327510917, | |
| "eval_loss": 0.441346138715744, | |
| "eval_runtime": 15.4217, | |
| "eval_samples_per_second": 32.422, | |
| "eval_steps_per_second": 2.075, | |
| "step": 1488 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_exact_match": 17.6, | |
| "eval_f1": 23.964126984126988, | |
| "eval_qa_bleu": 13.1735935431327, | |
| "eval_qa_exact_match": 0.164, | |
| "eval_recite_bleu": 35.20769002447571, | |
| "eval_recite_exact_match": 0.008, | |
| "step": 1488 | |
| }, | |
| { | |
| "epoch": 6.05, | |
| "learning_rate": 3e-05, | |
| "loss": 0.2459, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 6.25, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1802, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 6.45, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1853, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 6.65, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1905, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 6.85, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1906, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.7457903930131005, | |
| "eval_loss": 0.4084840714931488, | |
| "eval_runtime": 15.7338, | |
| "eval_samples_per_second": 31.779, | |
| "eval_steps_per_second": 2.034, | |
| "step": 1736 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_exact_match": 20.2, | |
| "eval_f1": 25.946536796536805, | |
| "eval_qa_bleu": 13.109935567750034, | |
| "eval_qa_exact_match": 0.198, | |
| "eval_recite_bleu": 41.370445416956926, | |
| "eval_recite_exact_match": 0.018, | |
| "step": 1736 | |
| }, | |
| { | |
| "epoch": 7.06, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1735, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 7.26, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1402, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 7.46, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1423, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 7.66, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1468, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 7.86, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1474, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.7476826783114993, | |
| "eval_loss": 0.4009522497653961, | |
| "eval_runtime": 15.0512, | |
| "eval_samples_per_second": 33.22, | |
| "eval_steps_per_second": 2.126, | |
| "step": 1984 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_exact_match": 17.4, | |
| "eval_f1": 24.613044733044738, | |
| "eval_qa_bleu": 11.328428780013391, | |
| "eval_qa_exact_match": 0.162, | |
| "eval_recite_bleu": 45.62279446830165, | |
| "eval_recite_exact_match": 0.036, | |
| "step": 1984 | |
| }, | |
| { | |
| "epoch": 8.06, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1383, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 8.27, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1181, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 8.47, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1212, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 8.67, | |
| "learning_rate": 3e-05, | |
| "loss": 0.118, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 8.87, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1264, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.7484366812227075, | |
| "eval_loss": 0.39788317680358887, | |
| "eval_runtime": 15.8038, | |
| "eval_samples_per_second": 31.638, | |
| "eval_steps_per_second": 2.025, | |
| "step": 2232 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_exact_match": 16.4, | |
| "eval_f1": 23.748124098124105, | |
| "eval_qa_bleu": 10.821376307326117, | |
| "eval_qa_exact_match": 0.148, | |
| "eval_recite_bleu": 46.63269966666572, | |
| "eval_recite_exact_match": 0.04, | |
| "step": 2232 | |
| }, | |
| { | |
| "epoch": 9.07, | |
| "learning_rate": 3e-05, | |
| "loss": 0.121, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 9.27, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1036, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 9.48, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1053, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 9.68, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1129, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 9.88, | |
| "learning_rate": 3e-05, | |
| "loss": 0.1091, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.7487336244541485, | |
| "eval_loss": 0.4059823751449585, | |
| "eval_runtime": 15.4716, | |
| "eval_samples_per_second": 32.317, | |
| "eval_steps_per_second": 2.068, | |
| "step": 2480 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_exact_match": 16.6, | |
| "eval_f1": 24.02516483516484, | |
| "eval_qa_bleu": 9.045796919947849, | |
| "eval_qa_exact_match": 0.156, | |
| "eval_recite_bleu": 50.48708434203323, | |
| "eval_recite_exact_match": 0.042, | |
| "step": 2480 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 4960, | |
| "num_train_epochs": 20, | |
| "save_steps": 500, | |
| "total_flos": 1.152564234928128e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |