| { | |
| "best_metric": 0.6162846088409424, | |
| "best_model_checkpoint": "BERT-WMM/run-13/checkpoint-533", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 2665, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 12.600181579589844, | |
| "learning_rate": 1.4607577277616791e-05, | |
| "loss": 0.7148, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7577464788732394, | |
| "eval_loss": 0.6162846088409424, | |
| "eval_runtime": 1.9944, | |
| "eval_samples_per_second": 1067.994, | |
| "eval_steps_per_second": 67.188, | |
| "step": 533 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 8.747657775878906, | |
| "learning_rate": 1.1234002848605984e-05, | |
| "loss": 0.4485, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7624413145539906, | |
| "eval_loss": 0.6197112202644348, | |
| "eval_runtime": 2.1341, | |
| "eval_samples_per_second": 998.1, | |
| "eval_steps_per_second": 62.791, | |
| "step": 1066 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "grad_norm": 34.570838928222656, | |
| "learning_rate": 7.860428419595179e-06, | |
| "loss": 0.2791, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7661971830985915, | |
| "eval_loss": 0.7673255205154419, | |
| "eval_runtime": 2.1145, | |
| "eval_samples_per_second": 1007.348, | |
| "eval_steps_per_second": 63.373, | |
| "step": 1599 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "grad_norm": 25.005464553833008, | |
| "learning_rate": 4.486853990584372e-06, | |
| "loss": 0.1691, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7680751173708921, | |
| "eval_loss": 1.1236116886138916, | |
| "eval_runtime": 2.0265, | |
| "eval_samples_per_second": 1051.061, | |
| "eval_steps_per_second": 66.123, | |
| "step": 2132 | |
| }, | |
| { | |
| "epoch": 4.69, | |
| "grad_norm": 0.5956600904464722, | |
| "learning_rate": 1.113279561573566e-06, | |
| "loss": 0.0966, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.7690140845070422, | |
| "eval_loss": 1.2481839656829834, | |
| "eval_runtime": 2.1691, | |
| "eval_samples_per_second": 981.987, | |
| "eval_steps_per_second": 61.778, | |
| "step": 2665 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2665, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 846139862236032.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 1.7981151706627596e-05, | |
| "num_train_epochs": 5, | |
| "per_device_train_batch_size": 16, | |
| "weight_decay": 0.009594844383188954 | |
| } | |
| } | |