| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.0, | |
| "eval_steps": 500, | |
| "global_step": 5022, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.8960573476702509, | |
| "grad_norm": 14.137969970703125, | |
| "learning_rate": 1.3440860215053763e-05, | |
| "loss": 2.3274, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 1.3252462148666382, | |
| "eval_rouge1": 0.5630271149940076, | |
| "eval_rouge2": 0.37605870482604453, | |
| "eval_rougeL": 0.4984598485550909, | |
| "eval_rougeLsum": 0.571321002640571, | |
| "eval_runtime": 49.4969, | |
| "eval_samples_per_second": 2.384, | |
| "eval_steps_per_second": 0.606, | |
| "step": 558 | |
| }, | |
| { | |
| "epoch": 1.7921146953405018, | |
| "grad_norm": 7.504210472106934, | |
| "learning_rate": 1.3679808841099164e-05, | |
| "loss": 1.17, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.0907840728759766, | |
| "eval_rouge1": 0.5562967978764126, | |
| "eval_rouge2": 0.3744307754882274, | |
| "eval_rougeL": 0.5008576418920765, | |
| "eval_rougeLsum": 0.5639195444626521, | |
| "eval_runtime": 49.3033, | |
| "eval_samples_per_second": 2.393, | |
| "eval_steps_per_second": 0.608, | |
| "step": 1116 | |
| }, | |
| { | |
| "epoch": 2.688172043010753, | |
| "grad_norm": 4.531522750854492, | |
| "learning_rate": 1.2186379928315411e-05, | |
| "loss": 0.8956, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 1.0720860958099365, | |
| "eval_rouge1": 0.5636947893309734, | |
| "eval_rouge2": 0.3744948097288884, | |
| "eval_rougeL": 0.5089233474314114, | |
| "eval_rougeLsum": 0.5687286653850421, | |
| "eval_runtime": 50.0187, | |
| "eval_samples_per_second": 2.359, | |
| "eval_steps_per_second": 0.6, | |
| "step": 1674 | |
| }, | |
| { | |
| "epoch": 3.5842293906810037, | |
| "grad_norm": 5.381931304931641, | |
| "learning_rate": 1.0692951015531662e-05, | |
| "loss": 0.7209, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.0957852602005005, | |
| "eval_rouge1": 0.5540030400056567, | |
| "eval_rouge2": 0.3692226432826772, | |
| "eval_rougeL": 0.49758037286998147, | |
| "eval_rougeLsum": 0.5561693313473722, | |
| "eval_runtime": 47.7686, | |
| "eval_samples_per_second": 2.47, | |
| "eval_steps_per_second": 0.628, | |
| "step": 2232 | |
| }, | |
| { | |
| "epoch": 4.480286738351254, | |
| "grad_norm": 4.125611305236816, | |
| "learning_rate": 9.19952210274791e-06, | |
| "loss": 0.5647, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_loss": 1.1148895025253296, | |
| "eval_rouge1": 0.5806893682446067, | |
| "eval_rouge2": 0.3943863308109349, | |
| "eval_rougeL": 0.5236896549899279, | |
| "eval_rougeLsum": 0.5875367327889313, | |
| "eval_runtime": 48.2153, | |
| "eval_samples_per_second": 2.447, | |
| "eval_steps_per_second": 0.622, | |
| "step": 2790 | |
| }, | |
| { | |
| "epoch": 5.376344086021505, | |
| "grad_norm": 5.281921863555908, | |
| "learning_rate": 7.706093189964159e-06, | |
| "loss": 0.4652, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 1.1711769104003906, | |
| "eval_rouge1": 0.567309904415172, | |
| "eval_rouge2": 0.37849761310553864, | |
| "eval_rougeL": 0.5092459232917496, | |
| "eval_rougeLsum": 0.5729114339966073, | |
| "eval_runtime": 47.1162, | |
| "eval_samples_per_second": 2.504, | |
| "eval_steps_per_second": 0.637, | |
| "step": 3348 | |
| }, | |
| { | |
| "epoch": 6.272401433691757, | |
| "grad_norm": 5.2352166175842285, | |
| "learning_rate": 6.212664277180406e-06, | |
| "loss": 0.3682, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_loss": 1.2149335145950317, | |
| "eval_rouge1": 0.5779989309920852, | |
| "eval_rouge2": 0.3932839170198733, | |
| "eval_rougeL": 0.519612547289248, | |
| "eval_rougeLsum": 0.5844791754998846, | |
| "eval_runtime": 49.3993, | |
| "eval_samples_per_second": 2.389, | |
| "eval_steps_per_second": 0.607, | |
| "step": 3906 | |
| }, | |
| { | |
| "epoch": 7.168458781362007, | |
| "grad_norm": 5.129788875579834, | |
| "learning_rate": 4.719235364396655e-06, | |
| "loss": 0.2964, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 1.2541477680206299, | |
| "eval_rouge1": 0.5741091721906807, | |
| "eval_rouge2": 0.3913854513407571, | |
| "eval_rougeL": 0.5175964085703821, | |
| "eval_rougeLsum": 0.5780595369378698, | |
| "eval_runtime": 51.0845, | |
| "eval_samples_per_second": 2.31, | |
| "eval_steps_per_second": 0.587, | |
| "step": 4464 | |
| }, | |
| { | |
| "epoch": 8.064516129032258, | |
| "grad_norm": 5.963613033294678, | |
| "learning_rate": 3.225806451612903e-06, | |
| "loss": 0.2434, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 8.960573476702509, | |
| "grad_norm": 5.311295032501221, | |
| "learning_rate": 1.7323775388291518e-06, | |
| "loss": 0.2005, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_loss": 1.283581256866455, | |
| "eval_rouge1": 0.5951372356242162, | |
| "eval_rouge2": 0.40327134125204, | |
| "eval_rougeL": 0.5352477504226778, | |
| "eval_rougeLsum": 0.6005008499350015, | |
| "eval_runtime": 90.0692, | |
| "eval_samples_per_second": 1.31, | |
| "eval_steps_per_second": 0.333, | |
| "step": 5022 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5580, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4941750029647872.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |