| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.383378016085791, | |
| "eval_steps": 500, | |
| "global_step": 7000, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.6702412868632708, | |
| "grad_norm": 20.23386001586914, | |
| "learning_rate": 1.2656372669468497e-05, | |
| "loss": 0.5543, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7663425803184509, | |
| "eval_loss": 0.4670431315898895, | |
| "eval_runtime": 8.8872, | |
| "eval_samples_per_second": 335.653, | |
| "eval_steps_per_second": 21.042, | |
| "step": 746 | |
| }, | |
| { | |
| "epoch": 1.3404825737265416, | |
| "grad_norm": 15.927043914794922, | |
| "learning_rate": 1.1747150494937714e-05, | |
| "loss": 0.3962, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7676835656166077, | |
| "eval_loss": 0.5088892579078674, | |
| "eval_runtime": 8.9675, | |
| "eval_samples_per_second": 332.647, | |
| "eval_steps_per_second": 20.853, | |
| "step": 1492 | |
| }, | |
| { | |
| "epoch": 2.0107238605898123, | |
| "grad_norm": 13.621393203735352, | |
| "learning_rate": 1.0837928320406931e-05, | |
| "loss": 0.3095, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.680965147453083, | |
| "grad_norm": 11.860713005065918, | |
| "learning_rate": 9.92870614587615e-06, | |
| "loss": 0.1321, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7743881940841675, | |
| "eval_loss": 0.7907660007476807, | |
| "eval_runtime": 8.9795, | |
| "eval_samples_per_second": 332.2, | |
| "eval_steps_per_second": 20.825, | |
| "step": 2238 | |
| }, | |
| { | |
| "epoch": 3.351206434316354, | |
| "grad_norm": 4.521104335784912, | |
| "learning_rate": 9.019483971345365e-06, | |
| "loss": 0.0899, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7730472683906555, | |
| "eval_loss": 1.2473280429840088, | |
| "eval_runtime": 9.0219, | |
| "eval_samples_per_second": 330.638, | |
| "eval_steps_per_second": 20.727, | |
| "step": 2984 | |
| }, | |
| { | |
| "epoch": 4.021447721179625, | |
| "grad_norm": 1.990652084350586, | |
| "learning_rate": 8.110261796814582e-06, | |
| "loss": 0.0644, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.6916890080428955, | |
| "grad_norm": 28.1745662689209, | |
| "learning_rate": 7.2010396222838e-06, | |
| "loss": 0.0335, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.7787462472915649, | |
| "eval_loss": 1.3820961713790894, | |
| "eval_runtime": 8.9708, | |
| "eval_samples_per_second": 332.522, | |
| "eval_steps_per_second": 20.845, | |
| "step": 3730 | |
| }, | |
| { | |
| "epoch": 5.361930294906166, | |
| "grad_norm": 82.3304443359375, | |
| "learning_rate": 6.2918174477530164e-06, | |
| "loss": 0.0335, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.7723767757415771, | |
| "eval_loss": 1.3331011533737183, | |
| "eval_runtime": 9.0247, | |
| "eval_samples_per_second": 330.537, | |
| "eval_steps_per_second": 20.721, | |
| "step": 4476 | |
| }, | |
| { | |
| "epoch": 6.032171581769437, | |
| "grad_norm": 53.1148567199707, | |
| "learning_rate": 5.382595273222234e-06, | |
| "loss": 0.0246, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.702412868632708, | |
| "grad_norm": 0.35531890392303467, | |
| "learning_rate": 4.4733730986914515e-06, | |
| "loss": 0.0126, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.7686892151832581, | |
| "eval_loss": 1.553377389907837, | |
| "eval_runtime": 8.9304, | |
| "eval_samples_per_second": 334.029, | |
| "eval_steps_per_second": 20.94, | |
| "step": 5222 | |
| }, | |
| { | |
| "epoch": 7.372654155495979, | |
| "grad_norm": 0.009911920875310898, | |
| "learning_rate": 3.564150924160669e-06, | |
| "loss": 0.0113, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.7737177610397339, | |
| "eval_loss": 1.53861403465271, | |
| "eval_runtime": 8.9722, | |
| "eval_samples_per_second": 332.47, | |
| "eval_steps_per_second": 20.842, | |
| "step": 5968 | |
| }, | |
| { | |
| "epoch": 8.04289544235925, | |
| "grad_norm": 4.0014214515686035, | |
| "learning_rate": 2.6549287496298857e-06, | |
| "loss": 0.0085, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 8.71313672922252, | |
| "grad_norm": 5.749587216996588e-05, | |
| "learning_rate": 1.745706575099103e-06, | |
| "loss": 0.006, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.7707006335258484, | |
| "eval_loss": 1.7166153192520142, | |
| "eval_runtime": 8.9324, | |
| "eval_samples_per_second": 333.952, | |
| "eval_steps_per_second": 20.935, | |
| "step": 6714 | |
| }, | |
| { | |
| "epoch": 9.383378016085791, | |
| "grad_norm": 0.26535865664482117, | |
| "learning_rate": 8.364844005683202e-07, | |
| "loss": 0.005, | |
| "step": 7000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 7460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1.6607423106651636e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 1.356559484399928e-05, | |
| "per_device_train_batch_size": 16 | |
| } | |
| } | |