| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 248, | |
| "global_step": 744, | |
| "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 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 4960, | |
| "num_train_epochs": 20, | |
| "save_steps": 500, | |
| "total_flos": 3.4656338231808e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |