| { | |
| "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": 12.96187973022461, | |
| "learning_rate": 5.852478627669732e-06, | |
| "loss": 0.5779, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7619845867156982, | |
| "eval_loss": 0.48268476128578186, | |
| "eval_runtime": 9.469, | |
| "eval_samples_per_second": 315.027, | |
| "eval_steps_per_second": 19.749, | |
| "step": 746 | |
| }, | |
| { | |
| "epoch": 1.3404825737265416, | |
| "grad_norm": 19.64900779724121, | |
| "learning_rate": 5.432041944647482e-06, | |
| "loss": 0.464, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7680187821388245, | |
| "eval_loss": 0.5030698180198669, | |
| "eval_runtime": 9.5015, | |
| "eval_samples_per_second": 313.951, | |
| "eval_steps_per_second": 19.681, | |
| "step": 1492 | |
| }, | |
| { | |
| "epoch": 2.0107238605898123, | |
| "grad_norm": 9.11796760559082, | |
| "learning_rate": 5.0116052616252305e-06, | |
| "loss": 0.4025, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.680965147453083, | |
| "grad_norm": 11.002016067504883, | |
| "learning_rate": 4.59116857860298e-06, | |
| "loss": 0.2688, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7747234106063843, | |
| "eval_loss": 0.5899690389633179, | |
| "eval_runtime": 9.4806, | |
| "eval_samples_per_second": 314.643, | |
| "eval_steps_per_second": 19.725, | |
| "step": 2238 | |
| }, | |
| { | |
| "epoch": 3.351206434316354, | |
| "grad_norm": 3.2045211791992188, | |
| "learning_rate": 4.170731895580729e-06, | |
| "loss": 0.2182, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7673482894897461, | |
| "eval_loss": 0.7394132614135742, | |
| "eval_runtime": 9.6202, | |
| "eval_samples_per_second": 310.077, | |
| "eval_steps_per_second": 19.438, | |
| "step": 2984 | |
| }, | |
| { | |
| "epoch": 4.021447721179625, | |
| "grad_norm": 0.5618192553520203, | |
| "learning_rate": 3.7502952125584775e-06, | |
| "loss": 0.1644, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.6916890080428955, | |
| "grad_norm": 19.86639976501465, | |
| "learning_rate": 3.3298585295362272e-06, | |
| "loss": 0.1173, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.7646664381027222, | |
| "eval_loss": 0.7920505404472351, | |
| "eval_runtime": 9.4974, | |
| "eval_samples_per_second": 314.085, | |
| "eval_steps_per_second": 19.69, | |
| "step": 3730 | |
| }, | |
| { | |
| "epoch": 5.361930294906166, | |
| "grad_norm": 7.497531414031982, | |
| "learning_rate": 2.909421846513976e-06, | |
| "loss": 0.1026, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.7740529775619507, | |
| "eval_loss": 0.9123848676681519, | |
| "eval_runtime": 9.4955, | |
| "eval_samples_per_second": 314.147, | |
| "eval_steps_per_second": 19.693, | |
| "step": 4476 | |
| }, | |
| { | |
| "epoch": 6.032171581769437, | |
| "grad_norm": 21.95626449584961, | |
| "learning_rate": 2.488985163491725e-06, | |
| "loss": 0.082, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.702412868632708, | |
| "grad_norm": 0.22410695254802704, | |
| "learning_rate": 2.0685484804694742e-06, | |
| "loss": 0.065, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.7720415592193604, | |
| "eval_loss": 0.9457674026489258, | |
| "eval_runtime": 9.4723, | |
| "eval_samples_per_second": 314.918, | |
| "eval_steps_per_second": 19.742, | |
| "step": 5222 | |
| }, | |
| { | |
| "epoch": 7.372654155495979, | |
| "grad_norm": 53.26298904418945, | |
| "learning_rate": 1.6481117974472235e-06, | |
| "loss": 0.0629, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.7740529775619507, | |
| "eval_loss": 0.9820337891578674, | |
| "eval_runtime": 9.5123, | |
| "eval_samples_per_second": 313.596, | |
| "eval_steps_per_second": 19.659, | |
| "step": 5968 | |
| }, | |
| { | |
| "epoch": 8.04289544235925, | |
| "grad_norm": 13.058491706848145, | |
| "learning_rate": 1.2276751144249726e-06, | |
| "loss": 0.0518, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 8.71313672922252, | |
| "grad_norm": 2.8812756538391113, | |
| "learning_rate": 8.072384314027216e-07, | |
| "loss": 0.0435, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.7743881940841675, | |
| "eval_loss": 0.9872733354568481, | |
| "eval_runtime": 9.4773, | |
| "eval_samples_per_second": 314.753, | |
| "eval_steps_per_second": 19.731, | |
| "step": 6714 | |
| }, | |
| { | |
| "epoch": 9.383378016085791, | |
| "grad_norm": 2.8628196716308594, | |
| "learning_rate": 3.868017483804708e-07, | |
| "loss": 0.037, | |
| "step": 7000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 7460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1.8861065792123148e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 6.272915310691983e-06, | |
| "per_device_train_batch_size": 16 | |
| } | |
| } | |