| { | |
| "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": 20.911739349365234, | |
| "learning_rate": 1.9406388835210846e-05, | |
| "loss": 0.5519, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7670130729675293, | |
| "eval_loss": 0.4711463153362274, | |
| "eval_runtime": 8.8245, | |
| "eval_samples_per_second": 338.035, | |
| "eval_steps_per_second": 21.191, | |
| "step": 746 | |
| }, | |
| { | |
| "epoch": 1.3404825737265416, | |
| "grad_norm": 12.272890090942383, | |
| "learning_rate": 1.801225170624455e-05, | |
| "loss": 0.3843, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7767348289489746, | |
| "eval_loss": 0.5532987117767334, | |
| "eval_runtime": 8.9725, | |
| "eval_samples_per_second": 332.46, | |
| "eval_steps_per_second": 20.841, | |
| "step": 1492 | |
| }, | |
| { | |
| "epoch": 2.0107238605898123, | |
| "grad_norm": 15.580121994018555, | |
| "learning_rate": 1.6618114577278254e-05, | |
| "loss": 0.2751, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.680965147453083, | |
| "grad_norm": 8.400450706481934, | |
| "learning_rate": 1.5223977448311958e-05, | |
| "loss": 0.1119, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7800871729850769, | |
| "eval_loss": 0.8922383189201355, | |
| "eval_runtime": 8.9532, | |
| "eval_samples_per_second": 333.175, | |
| "eval_steps_per_second": 20.886, | |
| "step": 2238 | |
| }, | |
| { | |
| "epoch": 3.351206434316354, | |
| "grad_norm": 42.92605972290039, | |
| "learning_rate": 1.382984031934566e-05, | |
| "loss": 0.0723, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7700302004814148, | |
| "eval_loss": 1.4834158420562744, | |
| "eval_runtime": 8.9912, | |
| "eval_samples_per_second": 331.769, | |
| "eval_steps_per_second": 20.798, | |
| "step": 2984 | |
| }, | |
| { | |
| "epoch": 4.021447721179625, | |
| "grad_norm": 0.11555304378271103, | |
| "learning_rate": 1.2435703190379364e-05, | |
| "loss": 0.0542, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.6916890080428955, | |
| "grad_norm": 0.4538387954235077, | |
| "learning_rate": 1.1041566061413068e-05, | |
| "loss": 0.025, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.777405321598053, | |
| "eval_loss": 1.4542551040649414, | |
| "eval_runtime": 8.9782, | |
| "eval_samples_per_second": 332.249, | |
| "eval_steps_per_second": 20.828, | |
| "step": 3730 | |
| }, | |
| { | |
| "epoch": 5.361930294906166, | |
| "grad_norm": 33.408546447753906, | |
| "learning_rate": 9.64742893244677e-06, | |
| "loss": 0.0203, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.7753939032554626, | |
| "eval_loss": 1.549986720085144, | |
| "eval_runtime": 9.039, | |
| "eval_samples_per_second": 330.014, | |
| "eval_steps_per_second": 20.688, | |
| "step": 4476 | |
| }, | |
| { | |
| "epoch": 6.032171581769437, | |
| "grad_norm": 0.05445607751607895, | |
| "learning_rate": 8.253291803480475e-06, | |
| "loss": 0.0144, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.702412868632708, | |
| "grad_norm": 15.311173439025879, | |
| "learning_rate": 6.859154674514179e-06, | |
| "loss": 0.009, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.7804223895072937, | |
| "eval_loss": 1.5616616010665894, | |
| "eval_runtime": 8.9521, | |
| "eval_samples_per_second": 333.219, | |
| "eval_steps_per_second": 20.889, | |
| "step": 5222 | |
| }, | |
| { | |
| "epoch": 7.372654155495979, | |
| "grad_norm": 0.05285044014453888, | |
| "learning_rate": 5.465017545547883e-06, | |
| "loss": 0.0072, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.7757291197776794, | |
| "eval_loss": 1.6639212369918823, | |
| "eval_runtime": 8.9695, | |
| "eval_samples_per_second": 332.573, | |
| "eval_steps_per_second": 20.849, | |
| "step": 5968 | |
| }, | |
| { | |
| "epoch": 8.04289544235925, | |
| "grad_norm": 0.034013714641332626, | |
| "learning_rate": 4.070880416581586e-06, | |
| "loss": 0.0055, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 8.71313672922252, | |
| "grad_norm": 4.92729377746582, | |
| "learning_rate": 2.676743287615289e-06, | |
| "loss": 0.0047, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.7787462472915649, | |
| "eval_loss": 1.6556830406188965, | |
| "eval_runtime": 8.9515, | |
| "eval_samples_per_second": 333.241, | |
| "eval_steps_per_second": 20.89, | |
| "step": 6714 | |
| }, | |
| { | |
| "epoch": 9.383378016085791, | |
| "grad_norm": 0.033865608274936676, | |
| "learning_rate": 1.2826061586489927e-06, | |
| "loss": 0.0045, | |
| "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": 2.0800525964177143e-05, | |
| "per_device_train_batch_size": 16 | |
| } | |
| } | |