| { | |
| "best_metric": 1.2347044944763184, | |
| "best_model_checkpoint": "data/Llama-31-8B_task-1_120-samples_config-2/checkpoint-33", | |
| "epoch": 12.909090909090908, | |
| "eval_steps": 500, | |
| "global_step": 71, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18181818181818182, | |
| "grad_norm": 1.874598503112793, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 2.2898, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 1.8044602870941162, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 2.0811, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 1.7498782873153687, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 2.1077, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "eval_loss": 2.0274555683135986, | |
| "eval_runtime": 9.6401, | |
| "eval_samples_per_second": 2.49, | |
| "eval_steps_per_second": 2.49, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0909090909090908, | |
| "grad_norm": 1.5882542133331299, | |
| "learning_rate": 2.4e-05, | |
| "loss": 1.891, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.4545454545454546, | |
| "grad_norm": 1.762362003326416, | |
| "learning_rate": 3.2000000000000005e-05, | |
| "loss": 1.9895, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 1.5516672134399414, | |
| "learning_rate": 4e-05, | |
| "loss": 1.775, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.642242431640625, | |
| "eval_runtime": 9.6438, | |
| "eval_samples_per_second": 2.489, | |
| "eval_steps_per_second": 2.489, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 2.1818181818181817, | |
| "grad_norm": 0.7731907963752747, | |
| "learning_rate": 4.8e-05, | |
| "loss": 1.6804, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 2.5454545454545454, | |
| "grad_norm": 0.7053793668746948, | |
| "learning_rate": 5.6000000000000006e-05, | |
| "loss": 1.5485, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "grad_norm": 0.742195725440979, | |
| "learning_rate": 6.400000000000001e-05, | |
| "loss": 1.4756, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "eval_loss": 1.5183296203613281, | |
| "eval_runtime": 9.6289, | |
| "eval_samples_per_second": 2.492, | |
| "eval_steps_per_second": 2.492, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 3.2727272727272725, | |
| "grad_norm": 0.6049473881721497, | |
| "learning_rate": 7.2e-05, | |
| "loss": 1.4936, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 3.6363636363636362, | |
| "grad_norm": 0.6063958406448364, | |
| "learning_rate": 8e-05, | |
| "loss": 1.3992, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.5723155736923218, | |
| "learning_rate": 8.800000000000001e-05, | |
| "loss": 1.3188, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.3610658645629883, | |
| "eval_runtime": 9.6351, | |
| "eval_samples_per_second": 2.491, | |
| "eval_steps_per_second": 2.491, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.363636363636363, | |
| "grad_norm": 0.5451256036758423, | |
| "learning_rate": 9.6e-05, | |
| "loss": 1.2429, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 4.7272727272727275, | |
| "grad_norm": 0.5676644444465637, | |
| "learning_rate": 9.999512620046522e-05, | |
| "loss": 1.2525, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 4.909090909090909, | |
| "eval_loss": 1.2826378345489502, | |
| "eval_runtime": 9.6405, | |
| "eval_samples_per_second": 2.489, | |
| "eval_steps_per_second": 2.489, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 5.090909090909091, | |
| "grad_norm": 0.6644198298454285, | |
| "learning_rate": 9.995614150494293e-05, | |
| "loss": 1.1568, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 5.454545454545454, | |
| "grad_norm": 0.6349135041236877, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 1.0659, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.818181818181818, | |
| "grad_norm": 0.6323952078819275, | |
| "learning_rate": 9.976136999909156e-05, | |
| "loss": 1.0827, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 1.2347044944763184, | |
| "eval_runtime": 9.6312, | |
| "eval_samples_per_second": 2.492, | |
| "eval_steps_per_second": 2.492, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 6.181818181818182, | |
| "grad_norm": 0.6223853826522827, | |
| "learning_rate": 9.96057350657239e-05, | |
| "loss": 1.0485, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 6.545454545454545, | |
| "grad_norm": 0.828182578086853, | |
| "learning_rate": 9.941141907232765e-05, | |
| "loss": 0.9365, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "grad_norm": 0.7417443990707397, | |
| "learning_rate": 9.917857354066931e-05, | |
| "loss": 0.886, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "eval_loss": 1.2388437986373901, | |
| "eval_runtime": 9.6409, | |
| "eval_samples_per_second": 2.489, | |
| "eval_steps_per_second": 2.489, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 7.2727272727272725, | |
| "grad_norm": 0.7539990544319153, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.8094, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 7.636363636363637, | |
| "grad_norm": 0.8493594527244568, | |
| "learning_rate": 9.859805002892732e-05, | |
| "loss": 0.7218, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 1.0422478914260864, | |
| "learning_rate": 9.825082472361557e-05, | |
| "loss": 0.7475, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 1.3558769226074219, | |
| "eval_runtime": 9.638, | |
| "eval_samples_per_second": 2.49, | |
| "eval_steps_per_second": 2.49, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.363636363636363, | |
| "grad_norm": 1.0563033819198608, | |
| "learning_rate": 9.786597487660337e-05, | |
| "loss": 0.6106, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 8.727272727272727, | |
| "grad_norm": 1.1736115217208862, | |
| "learning_rate": 9.744380058222483e-05, | |
| "loss": 0.5298, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 8.909090909090908, | |
| "eval_loss": 1.4767857789993286, | |
| "eval_runtime": 9.6386, | |
| "eval_samples_per_second": 2.49, | |
| "eval_steps_per_second": 2.49, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 1.3753471374511719, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.5113, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 9.454545454545455, | |
| "grad_norm": 1.2556344270706177, | |
| "learning_rate": 9.648882429441257e-05, | |
| "loss": 0.3796, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 9.818181818181818, | |
| "grad_norm": 2.3060855865478516, | |
| "learning_rate": 9.595676696276172e-05, | |
| "loss": 0.313, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 1.7835429906845093, | |
| "eval_runtime": 9.6254, | |
| "eval_samples_per_second": 2.493, | |
| "eval_steps_per_second": 2.493, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 10.181818181818182, | |
| "grad_norm": 1.2391914129257202, | |
| "learning_rate": 9.538887392664544e-05, | |
| "loss": 0.2808, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 10.545454545454545, | |
| "grad_norm": 1.3251339197158813, | |
| "learning_rate": 9.478558801197065e-05, | |
| "loss": 0.1961, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "grad_norm": 2.499495267868042, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.183, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "eval_loss": 2.217900037765503, | |
| "eval_runtime": 9.659, | |
| "eval_samples_per_second": 2.485, | |
| "eval_steps_per_second": 2.485, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 11.272727272727273, | |
| "grad_norm": 1.0954416990280151, | |
| "learning_rate": 9.347474647526095e-05, | |
| "loss": 0.1037, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 11.636363636363637, | |
| "grad_norm": 1.1775236129760742, | |
| "learning_rate": 9.276821300802534e-05, | |
| "loss": 0.0899, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 1.319805383682251, | |
| "learning_rate": 9.202833017478422e-05, | |
| "loss": 0.0808, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 2.5561976432800293, | |
| "eval_runtime": 9.6365, | |
| "eval_samples_per_second": 2.491, | |
| "eval_steps_per_second": 2.491, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.363636363636363, | |
| "grad_norm": 1.1726994514465332, | |
| "learning_rate": 9.125567491391476e-05, | |
| "loss": 0.0411, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 12.727272727272727, | |
| "grad_norm": 1.364807367324829, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.0522, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "eval_loss": 2.762176752090454, | |
| "eval_runtime": 9.6334, | |
| "eval_samples_per_second": 2.491, | |
| "eval_steps_per_second": 2.491, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "step": 71, | |
| "total_flos": 1.7560187623702528e+16, | |
| "train_loss": 0.9357028481825976, | |
| "train_runtime": 1340.1572, | |
| "train_samples_per_second": 3.283, | |
| "train_steps_per_second": 0.187 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 250, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 25, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 7, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.7560187623702528e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |