| { | |
| "best_global_step": 10, | |
| "best_metric": 5.472066879272461, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 10, | |
| "global_step": 300, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.1, | |
| "grad_norm": 0.0, | |
| "learning_rate": 0.0001, | |
| "loss": 5.4479, | |
| "mean_token_accuracy": 0.3418468192219734, | |
| "num_tokens": 6333.0, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 6333.0, | |
| "eval_runtime": 97.2579, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 0.0, | |
| "learning_rate": 9.970890689061622e-05, | |
| "loss": 5.4121, | |
| "mean_token_accuracy": 0.34842379465699197, | |
| "num_tokens": 12685.0, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 12685.0, | |
| "eval_runtime": 97.2929, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 0.0, | |
| "learning_rate": 9.883901697039808e-05, | |
| "loss": 5.3717, | |
| "mean_token_accuracy": 0.3537356600165367, | |
| "num_tokens": 19121.0, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 19121.0, | |
| "eval_runtime": 97.3018, | |
| "eval_samples_per_second": 2.055, | |
| "eval_steps_per_second": 1.028, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 0.0, | |
| "learning_rate": 9.740045899781352e-05, | |
| "loss": 5.3996, | |
| "mean_token_accuracy": 0.3528102159500122, | |
| "num_tokens": 25515.0, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 25515.0, | |
| "eval_runtime": 97.3146, | |
| "eval_samples_per_second": 2.055, | |
| "eval_steps_per_second": 1.028, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 0.0, | |
| "learning_rate": 9.540998314539328e-05, | |
| "loss": 5.4201, | |
| "mean_token_accuracy": 0.3465014174580574, | |
| "num_tokens": 31902.0, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 31902.0, | |
| "eval_runtime": 97.3368, | |
| "eval_samples_per_second": 2.055, | |
| "eval_steps_per_second": 1.027, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 0.0, | |
| "learning_rate": 9.289076596533872e-05, | |
| "loss": 5.4205, | |
| "mean_token_accuracy": 0.34779798686504365, | |
| "num_tokens": 38297.0, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 38297.0, | |
| "eval_runtime": 97.3361, | |
| "eval_samples_per_second": 2.055, | |
| "eval_steps_per_second": 1.027, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "grad_norm": 0.0, | |
| "learning_rate": 8.987214052813604e-05, | |
| "loss": 5.4029, | |
| "mean_token_accuracy": 0.3479448229074478, | |
| "num_tokens": 44694.0, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 44694.0, | |
| "eval_runtime": 97.3396, | |
| "eval_samples_per_second": 2.055, | |
| "eval_steps_per_second": 1.027, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 0.0, | |
| "learning_rate": 8.638925487636848e-05, | |
| "loss": 5.3906, | |
| "mean_token_accuracy": 0.35015087351202967, | |
| "num_tokens": 51069.0, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 51069.0, | |
| "eval_runtime": 97.3095, | |
| "eval_samples_per_second": 2.055, | |
| "eval_steps_per_second": 1.028, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 0.0, | |
| "learning_rate": 8.248266277059607e-05, | |
| "loss": 5.4128, | |
| "mean_token_accuracy": 0.35007362365722655, | |
| "num_tokens": 57445.0, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 57445.0, | |
| "eval_runtime": 97.2884, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.0, | |
| "learning_rate": 7.819785149254532e-05, | |
| "loss": 5.3656, | |
| "mean_token_accuracy": 0.3533430092036724, | |
| "num_tokens": 63853.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 63853.0, | |
| "eval_runtime": 97.266, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "grad_norm": 0.0, | |
| "learning_rate": 7.358471220373832e-05, | |
| "loss": 5.3805, | |
| "mean_token_accuracy": 0.3449681714177132, | |
| "num_tokens": 70269.0, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.1, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 70269.0, | |
| "eval_runtime": 97.2658, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "grad_norm": 0.0, | |
| "learning_rate": 6.869695902655897e-05, | |
| "loss": 5.3858, | |
| "mean_token_accuracy": 0.35103842467069624, | |
| "num_tokens": 76691.0, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.2, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 76691.0, | |
| "eval_runtime": 97.2612, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "grad_norm": 0.0, | |
| "learning_rate": 6.359150361181715e-05, | |
| "loss": 5.365, | |
| "mean_token_accuracy": 0.3574948668479919, | |
| "num_tokens": 83083.0, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.3, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 83083.0, | |
| "eval_runtime": 97.2714, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "grad_norm": 0.0, | |
| "learning_rate": 5.832779247517273e-05, | |
| "loss": 5.4248, | |
| "mean_token_accuracy": 0.3450557515025139, | |
| "num_tokens": 89470.0, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 89470.0, | |
| "eval_runtime": 97.2785, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "grad_norm": 0.0, | |
| "learning_rate": 5.296711481829226e-05, | |
| "loss": 5.423, | |
| "mean_token_accuracy": 0.35387562289834024, | |
| "num_tokens": 95793.0, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 95793.0, | |
| "eval_runtime": 97.269, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.7571888894277604e-05, | |
| "loss": 5.3571, | |
| "mean_token_accuracy": 0.3545306995511055, | |
| "num_tokens": 102197.0, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.6, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 102197.0, | |
| "eval_runtime": 97.2172, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.029, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.220493522673067e-05, | |
| "loss": 5.4303, | |
| "mean_token_accuracy": 0.34326251968741417, | |
| "num_tokens": 108568.0, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.7, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 108568.0, | |
| "eval_runtime": 97.2546, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.692874514489173e-05, | |
| "loss": 5.4084, | |
| "mean_token_accuracy": 0.35154723301529883, | |
| "num_tokens": 114964.0, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.8, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 114964.0, | |
| "eval_runtime": 97.2112, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.029, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.180475315182563e-05, | |
| "loss": 5.4501, | |
| "mean_token_accuracy": 0.34329178035259245, | |
| "num_tokens": 121333.0, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 121333.0, | |
| "eval_runtime": 97.2249, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.029, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.0, | |
| "learning_rate": 2.6892621598001156e-05, | |
| "loss": 5.4197, | |
| "mean_token_accuracy": 0.3468141175806522, | |
| "num_tokens": 127706.0, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 127706.0, | |
| "eval_runtime": 97.2359, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.028, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "grad_norm": 0.0, | |
| "learning_rate": 2.2249545989326514e-05, | |
| "loss": 5.3449, | |
| "mean_token_accuracy": 0.3567413754761219, | |
| "num_tokens": 134095.0, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 2.1, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 134095.0, | |
| "eval_runtime": 97.2729, | |
| "eval_samples_per_second": 2.056, | |
| "eval_steps_per_second": 1.028, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "grad_norm": 0.0, | |
| "learning_rate": 1.7929589018443016e-05, | |
| "loss": 5.3965, | |
| "mean_token_accuracy": 0.35015432387590406, | |
| "num_tokens": 140527.0, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 2.2, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 140527.0, | |
| "eval_runtime": 97.2339, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.028, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "grad_norm": 0.0, | |
| "learning_rate": 1.3983051073632997e-05, | |
| "loss": 5.4028, | |
| "mean_token_accuracy": 0.3500514805316925, | |
| "num_tokens": 146939.0, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 2.3, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 146939.0, | |
| "eval_runtime": 97.3103, | |
| "eval_samples_per_second": 2.055, | |
| "eval_steps_per_second": 1.028, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "grad_norm": 0.0, | |
| "learning_rate": 1.0455884554962725e-05, | |
| "loss": 5.4175, | |
| "mean_token_accuracy": 0.34618459567427634, | |
| "num_tokens": 153360.0, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.4, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 153360.0, | |
| "eval_runtime": 97.2167, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.029, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 0.0, | |
| "learning_rate": 7.389158817201542e-06, | |
| "loss": 5.3845, | |
| "mean_token_accuracy": 0.3464725069701672, | |
| "num_tokens": 159731.0, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 159731.0, | |
| "eval_runtime": 97.2232, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.029, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "grad_norm": 0.0, | |
| "learning_rate": 4.818581969574742e-06, | |
| "loss": 5.4243, | |
| "mean_token_accuracy": 0.34798368737101554, | |
| "num_tokens": 166113.0, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.6, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 166113.0, | |
| "eval_runtime": 97.215, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.029, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.7, | |
| "grad_norm": 0.0, | |
| "learning_rate": 2.774085100381735e-06, | |
| "loss": 5.4282, | |
| "mean_token_accuracy": 0.34819247499108313, | |
| "num_tokens": 172481.0, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.7, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 172481.0, | |
| "eval_runtime": 97.2154, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.029, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "grad_norm": 0.0, | |
| "learning_rate": 1.2794737676536994e-06, | |
| "loss": 5.4238, | |
| "mean_token_accuracy": 0.3495753288269043, | |
| "num_tokens": 178813.0, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.8, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 178813.0, | |
| "eval_runtime": 97.2134, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.029, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "grad_norm": 0.0, | |
| "learning_rate": 3.5215081379718074e-07, | |
| "loss": 5.4196, | |
| "mean_token_accuracy": 0.34980408027768134, | |
| "num_tokens": 185220.0, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 185220.0, | |
| "eval_runtime": 97.2093, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.029, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.0, | |
| "learning_rate": 2.9137316938265825e-09, | |
| "loss": 5.404, | |
| "mean_token_accuracy": 0.34670546352863313, | |
| "num_tokens": 191559.0, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 5.472066879272461, | |
| "eval_mean_token_accuracy": 0.3502181801199913, | |
| "eval_num_tokens": 191559.0, | |
| "eval_runtime": 97.2389, | |
| "eval_samples_per_second": 2.057, | |
| "eval_steps_per_second": 1.028, | |
| "step": 300 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 300, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "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": 8882783138119680.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |