| { | |
| "best_metric": 13.1948, | |
| "best_model_checkpoint": "/content/tst-translation/checkpoint-4000", | |
| "epoch": 20.0, | |
| "eval_steps": 400, | |
| "global_step": 4020, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.9900497512437811, | |
| "grad_norm": 2.333070993423462, | |
| "learning_rate": 0.00045024875621890546, | |
| "loss": 3.4257, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.9900497512437811, | |
| "eval_bleu": 4.1008, | |
| "eval_gen_len": 77.8284, | |
| "eval_loss": 2.1086840629577637, | |
| "eval_runtime": 201.8226, | |
| "eval_samples_per_second": 1.992, | |
| "eval_steps_per_second": 0.129, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 3.9800995024875623, | |
| "grad_norm": 2.2885234355926514, | |
| "learning_rate": 0.00040049751243781097, | |
| "loss": 1.8571, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 3.9800995024875623, | |
| "eval_bleu": 8.6198, | |
| "eval_gen_len": 61.1418, | |
| "eval_loss": 1.9292024374008179, | |
| "eval_runtime": 149.5469, | |
| "eval_samples_per_second": 2.688, | |
| "eval_steps_per_second": 0.174, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 5.970149253731344, | |
| "grad_norm": 1.9040518999099731, | |
| "learning_rate": 0.0003507462686567164, | |
| "loss": 1.2467, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 5.970149253731344, | |
| "eval_bleu": 10.7074, | |
| "eval_gen_len": 48.3184, | |
| "eval_loss": 1.9778931140899658, | |
| "eval_runtime": 82.3257, | |
| "eval_samples_per_second": 4.883, | |
| "eval_steps_per_second": 0.316, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 7.960199004975125, | |
| "grad_norm": 1.6420375108718872, | |
| "learning_rate": 0.00030099502487562194, | |
| "loss": 0.8749, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 7.960199004975125, | |
| "eval_bleu": 11.8538, | |
| "eval_gen_len": 49.3483, | |
| "eval_loss": 2.0539379119873047, | |
| "eval_runtime": 80.462, | |
| "eval_samples_per_second": 4.996, | |
| "eval_steps_per_second": 0.323, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 9.950248756218905, | |
| "grad_norm": 1.6268000602722168, | |
| "learning_rate": 0.0002512437810945274, | |
| "loss": 0.6141, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 9.950248756218905, | |
| "eval_bleu": 12.4452, | |
| "eval_gen_len": 51.1269, | |
| "eval_loss": 2.1948225498199463, | |
| "eval_runtime": 88.7102, | |
| "eval_samples_per_second": 4.532, | |
| "eval_steps_per_second": 0.293, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 11.940298507462687, | |
| "grad_norm": 1.412307858467102, | |
| "learning_rate": 0.00020149253731343284, | |
| "loss": 0.4446, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 11.940298507462687, | |
| "eval_bleu": 12.3052, | |
| "eval_gen_len": 48.0995, | |
| "eval_loss": 2.3901803493499756, | |
| "eval_runtime": 67.4373, | |
| "eval_samples_per_second": 5.961, | |
| "eval_steps_per_second": 0.386, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 13.930348258706468, | |
| "grad_norm": 1.7401809692382812, | |
| "learning_rate": 0.00015174129353233832, | |
| "loss": 0.3251, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 13.930348258706468, | |
| "eval_bleu": 12.5824, | |
| "eval_gen_len": 49.1244, | |
| "eval_loss": 2.5697591304779053, | |
| "eval_runtime": 74.1886, | |
| "eval_samples_per_second": 5.419, | |
| "eval_steps_per_second": 0.35, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 15.92039800995025, | |
| "grad_norm": 1.2750484943389893, | |
| "learning_rate": 0.00010199004975124378, | |
| "loss": 0.2501, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 15.92039800995025, | |
| "eval_bleu": 13.0619, | |
| "eval_gen_len": 50.6095, | |
| "eval_loss": 2.663144588470459, | |
| "eval_runtime": 80.5868, | |
| "eval_samples_per_second": 4.988, | |
| "eval_steps_per_second": 0.323, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 17.91044776119403, | |
| "grad_norm": 0.941927433013916, | |
| "learning_rate": 5.223880597014925e-05, | |
| "loss": 0.1986, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 17.91044776119403, | |
| "eval_bleu": 13.0557, | |
| "eval_gen_len": 51.1443, | |
| "eval_loss": 2.787724494934082, | |
| "eval_runtime": 83.6803, | |
| "eval_samples_per_second": 4.804, | |
| "eval_steps_per_second": 0.311, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 19.90049751243781, | |
| "grad_norm": 1.0183725357055664, | |
| "learning_rate": 2.4875621890547264e-06, | |
| "loss": 0.1692, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 19.90049751243781, | |
| "eval_bleu": 13.1948, | |
| "eval_gen_len": 49.9179, | |
| "eval_loss": 2.8421177864074707, | |
| "eval_runtime": 82.7516, | |
| "eval_samples_per_second": 4.858, | |
| "eval_steps_per_second": 0.314, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "step": 4020, | |
| "total_flos": 4364211764689920.0, | |
| "train_loss": 0.9366569099141591, | |
| "train_runtime": 4542.9296, | |
| "train_samples_per_second": 7.075, | |
| "train_steps_per_second": 0.885 | |
| } | |
| ], | |
| "logging_steps": 400, | |
| "max_steps": 4020, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 400, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4364211764689920.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |