| { | |
| "best_metric": 0.8038990825688074, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-4/checkpoint-2635", | |
| "epoch": 6.0, | |
| "eval_steps": 500, | |
| "global_step": 3162, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 6.212067127227783, | |
| "learning_rate": 1.7783082729779372e-05, | |
| "loss": 1.8496, | |
| "step": 527 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7717889908256881, | |
| "eval_loss": 1.3957021236419678, | |
| "eval_runtime": 2.8127, | |
| "eval_samples_per_second": 310.024, | |
| "eval_steps_per_second": 2.489, | |
| "step": 527 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": NaN, | |
| "learning_rate": 1.4233214981823416e-05, | |
| "loss": 1.2959, | |
| "step": 1054 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7924311926605505, | |
| "eval_loss": 1.2182828187942505, | |
| "eval_runtime": 2.8191, | |
| "eval_samples_per_second": 309.323, | |
| "eval_steps_per_second": 2.483, | |
| "step": 1054 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 32.567195892333984, | |
| "learning_rate": 1.067659843586754e-05, | |
| "loss": 1.0858, | |
| "step": 1581 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7958715596330275, | |
| "eval_loss": 1.1549702882766724, | |
| "eval_runtime": 2.8186, | |
| "eval_samples_per_second": 309.375, | |
| "eval_steps_per_second": 2.484, | |
| "step": 1581 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 19.192340850830078, | |
| "learning_rate": 7.119981889911666e-06, | |
| "loss": 0.9641, | |
| "step": 2108 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.801605504587156, | |
| "eval_loss": 1.103049874305725, | |
| "eval_runtime": 2.81, | |
| "eval_samples_per_second": 310.323, | |
| "eval_steps_per_second": 2.491, | |
| "step": 2108 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 11.936226844787598, | |
| "learning_rate": 3.563365343955791e-06, | |
| "loss": 0.9032, | |
| "step": 2635 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.8038990825688074, | |
| "eval_loss": 1.0780220031738281, | |
| "eval_runtime": 2.8278, | |
| "eval_samples_per_second": 308.366, | |
| "eval_steps_per_second": 2.475, | |
| "step": 2635 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 28.104488372802734, | |
| "learning_rate": 6.748797999916271e-09, | |
| "loss": 0.8596, | |
| "step": 3162 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.8038990825688074, | |
| "eval_loss": 1.0743507146835327, | |
| "eval_runtime": 2.8183, | |
| "eval_samples_per_second": 309.402, | |
| "eval_steps_per_second": 2.484, | |
| "step": 3162 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 3162, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 6, | |
| "save_steps": 500, | |
| "total_flos": 48527917525620.0, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.6704103438400755, | |
| "learning_rate": 2.1339699275735247e-05, | |
| "num_train_epochs": 6, | |
| "temperature": 15 | |
| } | |
| } | |