| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 6.702412868632708, | |
| "eval_steps": 500, | |
| "global_step": 5000, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.6702412868632708, | |
| "grad_norm": 4.357646942138672, | |
| "learning_rate": 1.0524844447678954e-05, | |
| "loss": 0.552, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7542742490768433, | |
| "eval_loss": 0.4794575572013855, | |
| "eval_runtime": 8.8133, | |
| "eval_samples_per_second": 338.465, | |
| "eval_steps_per_second": 21.218, | |
| "step": 746 | |
| }, | |
| { | |
| "epoch": 1.3404825737265416, | |
| "grad_norm": 9.759268760681152, | |
| "learning_rate": 9.768749300575582e-06, | |
| "loss": 0.4163, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7660073637962341, | |
| "eval_loss": 0.5210239291191101, | |
| "eval_runtime": 8.8965, | |
| "eval_samples_per_second": 335.302, | |
| "eval_steps_per_second": 21.02, | |
| "step": 1492 | |
| }, | |
| { | |
| "epoch": 2.0107238605898123, | |
| "grad_norm": 5.988788604736328, | |
| "learning_rate": 9.01265415347221e-06, | |
| "loss": 0.3141, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.680965147453083, | |
| "grad_norm": 7.434717178344727, | |
| "learning_rate": 8.256559006368837e-06, | |
| "loss": 0.1535, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7717063426971436, | |
| "eval_loss": 0.7554643154144287, | |
| "eval_runtime": 8.8771, | |
| "eval_samples_per_second": 336.034, | |
| "eval_steps_per_second": 21.065, | |
| "step": 2238 | |
| }, | |
| { | |
| "epoch": 3.351206434316354, | |
| "grad_norm": 2.2768313884735107, | |
| "learning_rate": 7.5004638592654626e-06, | |
| "loss": 0.1157, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7656721472740173, | |
| "eval_loss": 0.8537822961807251, | |
| "eval_runtime": 8.9123, | |
| "eval_samples_per_second": 334.706, | |
| "eval_steps_per_second": 20.982, | |
| "step": 2984 | |
| }, | |
| { | |
| "epoch": 4.021447721179625, | |
| "grad_norm": 13.178425788879395, | |
| "learning_rate": 6.744368712162088e-06, | |
| "loss": 0.0762, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.6916890080428955, | |
| "grad_norm": 11.147052764892578, | |
| "learning_rate": 5.988273565058716e-06, | |
| "loss": 0.0402, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.7683539986610413, | |
| "eval_loss": 1.0469584465026855, | |
| "eval_runtime": 8.8798, | |
| "eval_samples_per_second": 335.931, | |
| "eval_steps_per_second": 21.059, | |
| "step": 3730 | |
| }, | |
| { | |
| "epoch": 5.361930294906166, | |
| "grad_norm": 6.590982437133789, | |
| "learning_rate": 5.232178417955343e-06, | |
| "loss": 0.0302, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.763660728931427, | |
| "eval_loss": 1.0408499240875244, | |
| "eval_runtime": 9.1073, | |
| "eval_samples_per_second": 327.539, | |
| "eval_steps_per_second": 20.533, | |
| "step": 4476 | |
| }, | |
| { | |
| "epoch": 6.032171581769437, | |
| "grad_norm": 6.727069854736328, | |
| "learning_rate": 4.476083270851969e-06, | |
| "loss": 0.0248, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.702412868632708, | |
| "grad_norm": 0.06561928987503052, | |
| "learning_rate": 3.719988123748596e-06, | |
| "loss": 0.017, | |
| "step": 5000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 7460, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1.3910619432415356e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 1.1280939594782328e-05, | |
| "per_device_train_batch_size": 16 | |
| } | |
| } | |