| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 7.861635220125786, | |
| "eval_steps": 500, | |
| "global_step": 2500, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.5958064516129032, | |
| "eval_loss": 0.1949465274810791, | |
| "eval_runtime": 1.357, | |
| "eval_samples_per_second": 2284.501, | |
| "eval_steps_per_second": 47.901, | |
| "step": 318 | |
| }, | |
| { | |
| "epoch": 1.57, | |
| "learning_rate": 1.606918238993711e-05, | |
| "loss": 0.3113, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8164516129032258, | |
| "eval_loss": 0.09700000286102295, | |
| "eval_runtime": 1.3602, | |
| "eval_samples_per_second": 2279.158, | |
| "eval_steps_per_second": 47.789, | |
| "step": 636 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8754838709677419, | |
| "eval_loss": 0.0647290050983429, | |
| "eval_runtime": 1.3748, | |
| "eval_samples_per_second": 2254.826, | |
| "eval_steps_per_second": 47.279, | |
| "step": 954 | |
| }, | |
| { | |
| "epoch": 3.14, | |
| "learning_rate": 1.2138364779874214e-05, | |
| "loss": 0.1105, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8951612903225806, | |
| "eval_loss": 0.05031874030828476, | |
| "eval_runtime": 1.3686, | |
| "eval_samples_per_second": 2265.046, | |
| "eval_steps_per_second": 47.493, | |
| "step": 1272 | |
| }, | |
| { | |
| "epoch": 4.72, | |
| "learning_rate": 8.207547169811321e-06, | |
| "loss": 0.0722, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9035483870967742, | |
| "eval_loss": 0.04259829595685005, | |
| "eval_runtime": 1.3733, | |
| "eval_samples_per_second": 2257.258, | |
| "eval_steps_per_second": 47.33, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.91, | |
| "eval_loss": 0.03827530890703201, | |
| "eval_runtime": 1.3751, | |
| "eval_samples_per_second": 2254.357, | |
| "eval_steps_per_second": 47.269, | |
| "step": 1908 | |
| }, | |
| { | |
| "epoch": 6.29, | |
| "learning_rate": 4.276729559748428e-06, | |
| "loss": 0.058, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.9119354838709678, | |
| "eval_loss": 0.035722751170396805, | |
| "eval_runtime": 1.3775, | |
| "eval_samples_per_second": 2250.492, | |
| "eval_steps_per_second": 47.188, | |
| "step": 2226 | |
| }, | |
| { | |
| "epoch": 7.86, | |
| "learning_rate": 3.459119496855346e-07, | |
| "loss": 0.0523, | |
| "step": 2500 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2544, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 500, | |
| "total_flos": 649131947564688.0, | |
| "train_batch_size": 48, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.05314446157998587, | |
| "num_train_epochs": 8, | |
| "temperature": 17 | |
| } | |
| } | |