| { | |
| "best_metric": 0.9282445542742083, | |
| "best_model_checkpoint": "output/fine_tuned/t5-base/QNLI/checkpoint-6548", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 8185, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.30543677458766033, | |
| "grad_norm": 3.3223073482513428, | |
| "learning_rate": 4.6945632254123404e-05, | |
| "loss": 0.4673, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.6108735491753207, | |
| "grad_norm": 3.269711494445801, | |
| "learning_rate": 4.389126450824679e-05, | |
| "loss": 0.3224, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.916310323762981, | |
| "grad_norm": 1.466129183769226, | |
| "learning_rate": 4.083689676237019e-05, | |
| "loss": 0.2856, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.914881933003844, | |
| "eval_loss": 0.22163903713226318, | |
| "eval_runtime": 10.4314, | |
| "eval_samples_per_second": 523.707, | |
| "eval_steps_per_second": 8.244, | |
| "step": 1637 | |
| }, | |
| { | |
| "epoch": 1.2217470983506413, | |
| "grad_norm": 2.509833335876465, | |
| "learning_rate": 3.778252901649359e-05, | |
| "loss": 0.2448, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.5271838729383018, | |
| "grad_norm": 3.2846474647521973, | |
| "learning_rate": 3.472816127061698e-05, | |
| "loss": 0.2259, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.832620647525962, | |
| "grad_norm": 2.8232882022857666, | |
| "learning_rate": 3.167379352474038e-05, | |
| "loss": 0.2258, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9220208676551346, | |
| "eval_loss": 0.20598259568214417, | |
| "eval_runtime": 9.5458, | |
| "eval_samples_per_second": 572.292, | |
| "eval_steps_per_second": 9.009, | |
| "step": 3274 | |
| }, | |
| { | |
| "epoch": 2.1380574221136226, | |
| "grad_norm": 2.700801372528076, | |
| "learning_rate": 2.8619425778863777e-05, | |
| "loss": 0.1986, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.4434941967012827, | |
| "grad_norm": 1.884992241859436, | |
| "learning_rate": 2.556505803298717e-05, | |
| "loss": 0.1777, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.748930971288943, | |
| "grad_norm": 1.391169548034668, | |
| "learning_rate": 2.251069028711057e-05, | |
| "loss": 0.1791, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.9276954054548783, | |
| "eval_loss": 0.20381644368171692, | |
| "eval_runtime": 9.2262, | |
| "eval_samples_per_second": 592.117, | |
| "eval_steps_per_second": 9.321, | |
| "step": 4911 | |
| }, | |
| { | |
| "epoch": 3.0543677458766036, | |
| "grad_norm": 2.7919557094573975, | |
| "learning_rate": 1.9456322541233964e-05, | |
| "loss": 0.1724, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 3.359804520464264, | |
| "grad_norm": 2.4957516193389893, | |
| "learning_rate": 1.6401954795357362e-05, | |
| "loss": 0.1461, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 3.665241295051924, | |
| "grad_norm": 3.270733118057251, | |
| "learning_rate": 1.3347587049480758e-05, | |
| "loss": 0.1472, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.9706780696395847, | |
| "grad_norm": 4.12385368347168, | |
| "learning_rate": 1.0293219303604154e-05, | |
| "loss": 0.1476, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9282445542742083, | |
| "eval_loss": 0.2214784473180771, | |
| "eval_runtime": 9.3481, | |
| "eval_samples_per_second": 584.394, | |
| "eval_steps_per_second": 9.2, | |
| "step": 6548 | |
| }, | |
| { | |
| "epoch": 4.276114844227245, | |
| "grad_norm": 3.3767876625061035, | |
| "learning_rate": 7.238851557727551e-06, | |
| "loss": 0.1305, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 4.581551618814905, | |
| "grad_norm": 2.924355983734131, | |
| "learning_rate": 4.184483811850947e-06, | |
| "loss": 0.1278, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 4.886988393402565, | |
| "grad_norm": 5.803096294403076, | |
| "learning_rate": 1.1301160659743434e-06, | |
| "loss": 0.1263, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9278784550613216, | |
| "eval_loss": 0.2300529181957245, | |
| "eval_runtime": 9.2683, | |
| "eval_samples_per_second": 589.431, | |
| "eval_steps_per_second": 9.279, | |
| "step": 8185 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "step": 8185, | |
| "total_flos": 7.99682578684032e+16, | |
| "train_loss": 0.2058652666915679, | |
| "train_runtime": 2769.836, | |
| "train_samples_per_second": 189.078, | |
| "train_steps_per_second": 2.955 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 8185, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 7.99682578684032e+16, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |