| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 20.0, | |
| "eval_steps": 500, | |
| "global_step": 20320, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.984251968503937, | |
| "grad_norm": 9809.25390625, | |
| "learning_rate": 0.00029969999999999997, | |
| "loss": 1.3222, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.2888088452521089, | |
| "eval_loss": 5.052667617797852, | |
| "eval_runtime": 867.6167, | |
| "eval_samples_per_second": 38.736, | |
| "eval_steps_per_second": 1.211, | |
| "step": 1016 | |
| }, | |
| { | |
| "epoch": 1.968503937007874, | |
| "grad_norm": 7169.68408203125, | |
| "learning_rate": 0.0005997, | |
| "loss": 1.1829, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.29311490927283645, | |
| "eval_loss": 4.973076343536377, | |
| "eval_runtime": 866.7489, | |
| "eval_samples_per_second": 38.775, | |
| "eval_steps_per_second": 1.213, | |
| "step": 2032 | |
| }, | |
| { | |
| "epoch": 2.952755905511811, | |
| "grad_norm": 5481.02392578125, | |
| "learning_rate": 0.0005672816593886462, | |
| "loss": 1.1528, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.29985522038131823, | |
| "eval_loss": 4.848499298095703, | |
| "eval_runtime": 870.5632, | |
| "eval_samples_per_second": 38.605, | |
| "eval_steps_per_second": 1.207, | |
| "step": 3048 | |
| }, | |
| { | |
| "epoch": 3.937007874015748, | |
| "grad_norm": 5024.47705078125, | |
| "learning_rate": 0.0005345305676855895, | |
| "loss": 1.128, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.3019899977220967, | |
| "eval_loss": 4.790306568145752, | |
| "eval_runtime": 870.3918, | |
| "eval_samples_per_second": 38.612, | |
| "eval_steps_per_second": 1.208, | |
| "step": 4064 | |
| }, | |
| { | |
| "epoch": 4.921259842519685, | |
| "grad_norm": 4731.38330078125, | |
| "learning_rate": 0.0005017794759825327, | |
| "loss": 1.1098, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.3039251673839655, | |
| "eval_loss": 4.753741264343262, | |
| "eval_runtime": 866.2087, | |
| "eval_samples_per_second": 38.799, | |
| "eval_steps_per_second": 1.213, | |
| "step": 5080 | |
| }, | |
| { | |
| "epoch": 5.905511811023622, | |
| "grad_norm": 4894.84375, | |
| "learning_rate": 0.00046902838427947594, | |
| "loss": 1.0951, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.30580944523971787, | |
| "eval_loss": 4.716702938079834, | |
| "eval_runtime": 865.218, | |
| "eval_samples_per_second": 38.843, | |
| "eval_steps_per_second": 1.215, | |
| "step": 6096 | |
| }, | |
| { | |
| "epoch": 6.889763779527559, | |
| "grad_norm": 5185.75244140625, | |
| "learning_rate": 0.00043627729257641914, | |
| "loss": 1.0824, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.30677394395426305, | |
| "eval_loss": 4.699747562408447, | |
| "eval_runtime": 865.4872, | |
| "eval_samples_per_second": 38.831, | |
| "eval_steps_per_second": 1.214, | |
| "step": 7112 | |
| }, | |
| { | |
| "epoch": 7.874015748031496, | |
| "grad_norm": 5380.1005859375, | |
| "learning_rate": 0.0004035262008733624, | |
| "loss": 1.0724, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.3080556721421747, | |
| "eval_loss": 4.674106597900391, | |
| "eval_runtime": 865.0937, | |
| "eval_samples_per_second": 38.849, | |
| "eval_steps_per_second": 1.215, | |
| "step": 8128 | |
| }, | |
| { | |
| "epoch": 8.858267716535433, | |
| "grad_norm": 5194.1533203125, | |
| "learning_rate": 0.00037077510917030566, | |
| "loss": 1.0626, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.30815867855524104, | |
| "eval_loss": 4.669048309326172, | |
| "eval_runtime": 865.3252, | |
| "eval_samples_per_second": 38.839, | |
| "eval_steps_per_second": 1.215, | |
| "step": 9144 | |
| }, | |
| { | |
| "epoch": 9.84251968503937, | |
| "grad_norm": 5698.17236328125, | |
| "learning_rate": 0.00033802401746724887, | |
| "loss": 1.0532, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.3097009797779021, | |
| "eval_loss": 4.659518241882324, | |
| "eval_runtime": 864.3515, | |
| "eval_samples_per_second": 38.882, | |
| "eval_steps_per_second": 1.216, | |
| "step": 10160 | |
| }, | |
| { | |
| "epoch": 10.826771653543307, | |
| "grad_norm": 5876.86181640625, | |
| "learning_rate": 0.0003052729257641921, | |
| "loss": 1.0428, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.3101439248226706, | |
| "eval_loss": 4.65607213973999, | |
| "eval_runtime": 865.0882, | |
| "eval_samples_per_second": 38.849, | |
| "eval_steps_per_second": 1.215, | |
| "step": 11176 | |
| }, | |
| { | |
| "epoch": 11.811023622047244, | |
| "grad_norm": 6474.33642578125, | |
| "learning_rate": 0.00027252183406113533, | |
| "loss": 1.0351, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.31076691273301343, | |
| "eval_loss": 4.64418363571167, | |
| "eval_runtime": 867.9075, | |
| "eval_samples_per_second": 38.723, | |
| "eval_steps_per_second": 1.211, | |
| "step": 12192 | |
| }, | |
| { | |
| "epoch": 12.795275590551181, | |
| "grad_norm": 7132.021484375, | |
| "learning_rate": 0.0002397707423580786, | |
| "loss": 1.0274, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.3109399099366426, | |
| "eval_loss": 4.64435338973999, | |
| "eval_runtime": 867.2849, | |
| "eval_samples_per_second": 38.751, | |
| "eval_steps_per_second": 1.212, | |
| "step": 13208 | |
| }, | |
| { | |
| "epoch": 13.779527559055119, | |
| "grad_norm": 6837.5517578125, | |
| "learning_rate": 0.00020701965065502182, | |
| "loss": 1.0206, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.3116507648211613, | |
| "eval_loss": 4.642159461975098, | |
| "eval_runtime": 866.026, | |
| "eval_samples_per_second": 38.807, | |
| "eval_steps_per_second": 1.214, | |
| "step": 14224 | |
| }, | |
| { | |
| "epoch": 14.763779527559056, | |
| "grad_norm": 7605.75244140625, | |
| "learning_rate": 0.00017426855895196505, | |
| "loss": 1.0125, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_accuracy": 0.3118397748928477, | |
| "eval_loss": 4.639294624328613, | |
| "eval_runtime": 867.3792, | |
| "eval_samples_per_second": 38.747, | |
| "eval_steps_per_second": 1.212, | |
| "step": 15240 | |
| }, | |
| { | |
| "epoch": 15.748031496062993, | |
| "grad_norm": 7835.32275390625, | |
| "learning_rate": 0.0001415174672489083, | |
| "loss": 1.0029, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.3124415675887439, | |
| "eval_loss": 4.64095401763916, | |
| "eval_runtime": 867.1892, | |
| "eval_samples_per_second": 38.755, | |
| "eval_steps_per_second": 1.212, | |
| "step": 16256 | |
| }, | |
| { | |
| "epoch": 16.73228346456693, | |
| "grad_norm": 8777.490234375, | |
| "learning_rate": 0.00010876637554585152, | |
| "loss": 0.9949, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_accuracy": 0.3122564006053912, | |
| "eval_loss": 4.6424150466918945, | |
| "eval_runtime": 865.8304, | |
| "eval_samples_per_second": 38.816, | |
| "eval_steps_per_second": 1.214, | |
| "step": 17272 | |
| }, | |
| { | |
| "epoch": 17.716535433070867, | |
| "grad_norm": 7878.4384765625, | |
| "learning_rate": 7.601528384279475e-05, | |
| "loss": 0.9891, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.31283891963100763, | |
| "eval_loss": 4.6433796882629395, | |
| "eval_runtime": 865.5459, | |
| "eval_samples_per_second": 38.829, | |
| "eval_steps_per_second": 1.214, | |
| "step": 18288 | |
| }, | |
| { | |
| "epoch": 18.700787401574804, | |
| "grad_norm": 8385.39453125, | |
| "learning_rate": 4.326419213973799e-05, | |
| "loss": 0.9799, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_accuracy": 0.3129960786524129, | |
| "eval_loss": 4.641451358795166, | |
| "eval_runtime": 865.77, | |
| "eval_samples_per_second": 38.819, | |
| "eval_steps_per_second": 1.214, | |
| "step": 19304 | |
| }, | |
| { | |
| "epoch": 19.68503937007874, | |
| "grad_norm": 8282.5029296875, | |
| "learning_rate": 1.0513100436681222e-05, | |
| "loss": 0.977, | |
| "step": 20000 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_accuracy": 0.3132122232568799, | |
| "eval_loss": 4.643510341644287, | |
| "eval_runtime": 865.9594, | |
| "eval_samples_per_second": 38.81, | |
| "eval_steps_per_second": 1.214, | |
| "step": 20320 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "step": 20320, | |
| "total_flos": 1.6974575566848e+17, | |
| "train_loss": 1.065762345621905, | |
| "train_runtime": 40277.3058, | |
| "train_samples_per_second": 16.129, | |
| "train_steps_per_second": 0.505 | |
| } | |
| ], | |
| "logging_steps": 1000, | |
| "max_steps": 20320, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "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": 1.6974575566848e+17, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |