| { | |
| "best_metric": 0.3189895749092102, | |
| "best_model_checkpoint": "my_awesome_model_truncaded_split_36k\\checkpoint-1837", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 7348, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2721829069134458, | |
| "grad_norm": 4.77004337310791, | |
| "learning_rate": 9.319542732716386e-06, | |
| "loss": 0.5661, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5443658138268916, | |
| "grad_norm": 13.415882110595703, | |
| "learning_rate": 8.639085465432772e-06, | |
| "loss": 0.457, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.8165487207403375, | |
| "grad_norm": 13.942790985107422, | |
| "learning_rate": 7.958628198149157e-06, | |
| "loss": 0.3841, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8633455832312509, | |
| "eval_loss": 0.3189895749092102, | |
| "eval_runtime": 2115.733, | |
| "eval_samples_per_second": 3.473, | |
| "eval_steps_per_second": 0.217, | |
| "step": 1837 | |
| }, | |
| { | |
| "epoch": 1.0887316276537833, | |
| "grad_norm": 21.668664932250977, | |
| "learning_rate": 7.2781709308655426e-06, | |
| "loss": 0.343, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.360914534567229, | |
| "grad_norm": 14.608024597167969, | |
| "learning_rate": 6.597713663581927e-06, | |
| "loss": 0.3067, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.633097441480675, | |
| "grad_norm": 11.27497673034668, | |
| "learning_rate": 5.917256396298313e-06, | |
| "loss": 0.2714, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.905280348394121, | |
| "grad_norm": 8.441925048828125, | |
| "learning_rate": 5.236799129014698e-06, | |
| "loss": 0.2698, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8595345038791343, | |
| "eval_loss": 0.3298398554325104, | |
| "eval_runtime": 2168.9042, | |
| "eval_samples_per_second": 3.387, | |
| "eval_steps_per_second": 0.212, | |
| "step": 3674 | |
| }, | |
| { | |
| "epoch": 2.1774632553075666, | |
| "grad_norm": 35.317108154296875, | |
| "learning_rate": 4.5563418617310835e-06, | |
| "loss": 0.2298, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.4496461622210126, | |
| "grad_norm": 9.94012451171875, | |
| "learning_rate": 3.875884594447469e-06, | |
| "loss": 0.2087, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.721829069134458, | |
| "grad_norm": 24.299028396606445, | |
| "learning_rate": 3.1954273271638544e-06, | |
| "loss": 0.2119, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.9940119760479043, | |
| "grad_norm": 5.556156158447266, | |
| "learning_rate": 2.5149700598802396e-06, | |
| "loss": 0.2042, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.881312100176943, | |
| "eval_loss": 0.3292213976383209, | |
| "eval_runtime": 2182.1168, | |
| "eval_samples_per_second": 3.367, | |
| "eval_steps_per_second": 0.211, | |
| "step": 5511 | |
| }, | |
| { | |
| "epoch": 3.26619488296135, | |
| "grad_norm": 30.425508499145508, | |
| "learning_rate": 1.834512792596625e-06, | |
| "loss": 0.1642, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 3.538377789874796, | |
| "grad_norm": 10.257048606872559, | |
| "learning_rate": 1.1540555253130105e-06, | |
| "loss": 0.1731, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.810560696788242, | |
| "grad_norm": 37.602664947509766, | |
| "learning_rate": 4.735982580293958e-07, | |
| "loss": 0.1641, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8960119776779638, | |
| "eval_loss": 0.34636014699935913, | |
| "eval_runtime": 2178.6484, | |
| "eval_samples_per_second": 3.372, | |
| "eval_steps_per_second": 0.211, | |
| "step": 7348 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 7348, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "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": 2.907596283903792e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |