| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 8.04289544235925, | |
| "eval_steps": 500, | |
| "global_step": 6000, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.6702412868632708, | |
| "grad_norm": 1.8121609687805176, | |
| "learning_rate": 4.400558387583305e-05, | |
| "loss": 0.6264, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.5072075128555298, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.1548, | |
| "eval_samples_per_second": 325.839, | |
| "eval_steps_per_second": 20.426, | |
| "step": 746 | |
| }, | |
| { | |
| "epoch": 1.3404825737265416, | |
| "grad_norm": 1.8386740684509277, | |
| "learning_rate": 4.0844263195097915e-05, | |
| "loss": 0.7039, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.49882668256759644, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.1028, | |
| "eval_samples_per_second": 327.702, | |
| "eval_steps_per_second": 20.543, | |
| "step": 1492 | |
| }, | |
| { | |
| "epoch": 2.0107238605898123, | |
| "grad_norm": 1.775593876838684, | |
| "learning_rate": 3.768294251436278e-05, | |
| "loss": 0.7008, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.680965147453083, | |
| "grad_norm": 1.6979082822799683, | |
| "learning_rate": 3.452162183362765e-05, | |
| "loss": 0.7034, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.5015085339546204, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.1903, | |
| "eval_samples_per_second": 324.583, | |
| "eval_steps_per_second": 20.348, | |
| "step": 2238 | |
| }, | |
| { | |
| "epoch": 3.351206434316354, | |
| "grad_norm": 1.7350631952285767, | |
| "learning_rate": 3.136030115289252e-05, | |
| "loss": 0.7017, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.5065370202064514, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.2185, | |
| "eval_samples_per_second": 323.589, | |
| "eval_steps_per_second": 20.285, | |
| "step": 2984 | |
| }, | |
| { | |
| "epoch": 4.021447721179625, | |
| "grad_norm": 1.841038465499878, | |
| "learning_rate": 2.8198980472157382e-05, | |
| "loss": 0.7019, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.6916890080428955, | |
| "grad_norm": 1.6272307634353638, | |
| "learning_rate": 2.5037659791422253e-05, | |
| "loss": 0.6994, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.5095541477203369, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.0976, | |
| "eval_samples_per_second": 327.89, | |
| "eval_steps_per_second": 20.555, | |
| "step": 3730 | |
| }, | |
| { | |
| "epoch": 5.361930294906166, | |
| "grad_norm": 1.7355244159698486, | |
| "learning_rate": 2.1876339110687117e-05, | |
| "loss": 0.7006, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.49446865916252136, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.0977, | |
| "eval_samples_per_second": 327.885, | |
| "eval_steps_per_second": 20.555, | |
| "step": 4476 | |
| }, | |
| { | |
| "epoch": 6.032171581769437, | |
| "grad_norm": 1.6792231798171997, | |
| "learning_rate": 1.8715018429951985e-05, | |
| "loss": 0.7005, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.702412868632708, | |
| "grad_norm": 1.7535431385040283, | |
| "learning_rate": 1.5553697749216852e-05, | |
| "loss": 0.701, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.4941334128379822, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.0431, | |
| "eval_samples_per_second": 329.865, | |
| "eval_steps_per_second": 20.679, | |
| "step": 5222 | |
| }, | |
| { | |
| "epoch": 7.372654155495979, | |
| "grad_norm": 1.7443852424621582, | |
| "learning_rate": 1.239237706848172e-05, | |
| "loss": 0.7015, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.5122359991073608, | |
| "eval_loss": 0.6931472420692444, | |
| "eval_runtime": 9.0713, | |
| "eval_samples_per_second": 328.839, | |
| "eval_steps_per_second": 20.614, | |
| "step": 5968 | |
| }, | |
| { | |
| "epoch": 8.04289544235925, | |
| "grad_norm": 1.6793931722640991, | |
| "learning_rate": 9.231056387746587e-06, | |
| "loss": 0.7003, | |
| "step": 6000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 7460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1.563011071614504e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 4.716690455656818e-05, | |
| "per_device_train_batch_size": 16 | |
| } | |
| } | |