| { | |
| "best_global_step": 4500, | |
| "best_metric": 1.4357532262802124, | |
| "best_model_checkpoint": "./llama2-backward-output/checkpoint-4500", | |
| "epoch": 8.608666887198147, | |
| "eval_steps": 500, | |
| "global_step": 6500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.26463777704267283, | |
| "grad_norm": 1.4625507593154907, | |
| "learning_rate": 1.9600000000000002e-05, | |
| "loss": 2.2395, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5292755540853457, | |
| "grad_norm": 1.2510501146316528, | |
| "learning_rate": 1.9066666666666668e-05, | |
| "loss": 1.6278, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.6615944426066821, | |
| "eval_loss": NaN, | |
| "eval_runtime": 145.2798, | |
| "eval_samples_per_second": 9.251, | |
| "eval_steps_per_second": 4.626, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.7939133311280185, | |
| "grad_norm": 1.5188781023025513, | |
| "learning_rate": 1.8533333333333334e-05, | |
| "loss": 1.5642, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 1.0595434998346014, | |
| "grad_norm": 1.30088472366333, | |
| "learning_rate": 1.8e-05, | |
| "loss": 1.556, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 1.3241812768772743, | |
| "grad_norm": 1.7489525079727173, | |
| "learning_rate": 1.7466666666666667e-05, | |
| "loss": 1.5175, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.3241812768772743, | |
| "eval_loss": NaN, | |
| "eval_runtime": 145.4858, | |
| "eval_samples_per_second": 9.238, | |
| "eval_steps_per_second": 4.619, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.5888190539199472, | |
| "grad_norm": 2.2068583965301514, | |
| "learning_rate": 1.6933333333333336e-05, | |
| "loss": 1.5088, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.85345683096262, | |
| "grad_norm": 1.7514328956604004, | |
| "learning_rate": 1.64e-05, | |
| "loss": 1.5206, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.9857757194839563, | |
| "eval_loss": NaN, | |
| "eval_runtime": 145.5169, | |
| "eval_samples_per_second": 9.236, | |
| "eval_steps_per_second": 4.618, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.1190869996692028, | |
| "grad_norm": 1.1773896217346191, | |
| "learning_rate": 1.586666666666667e-05, | |
| "loss": 1.5358, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 2.3837247767118757, | |
| "grad_norm": 1.3273141384124756, | |
| "learning_rate": 1.5333333333333334e-05, | |
| "loss": 1.4694, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.6483625537545485, | |
| "grad_norm": 1.2176439762115479, | |
| "learning_rate": 1.48e-05, | |
| "loss": 1.4883, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.6483625537545485, | |
| "eval_loss": NaN, | |
| "eval_runtime": 145.5181, | |
| "eval_samples_per_second": 9.236, | |
| "eval_steps_per_second": 4.618, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.9130003307972214, | |
| "grad_norm": 1.7919120788574219, | |
| "learning_rate": 1.4266666666666668e-05, | |
| "loss": 1.5084, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 3.178630499503804, | |
| "grad_norm": 2.2114169597625732, | |
| "learning_rate": 1.3733333333333335e-05, | |
| "loss": 1.457, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 3.310949388025141, | |
| "eval_loss": NaN, | |
| "eval_runtime": 145.7705, | |
| "eval_samples_per_second": 9.22, | |
| "eval_steps_per_second": 4.61, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.443268276546477, | |
| "grad_norm": 2.1541590690612793, | |
| "learning_rate": 1.3200000000000002e-05, | |
| "loss": 1.4664, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 3.70790605358915, | |
| "grad_norm": 2.045276641845703, | |
| "learning_rate": 1.2666666666666667e-05, | |
| "loss": 1.4706, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 3.972543830631823, | |
| "grad_norm": 1.3809666633605957, | |
| "learning_rate": 1.2133333333333335e-05, | |
| "loss": 1.4671, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.972543830631823, | |
| "eval_loss": NaN, | |
| "eval_runtime": 145.8274, | |
| "eval_samples_per_second": 9.216, | |
| "eval_steps_per_second": 4.608, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 4.2381739993384056, | |
| "grad_norm": 1.9172261953353882, | |
| "learning_rate": 1.16e-05, | |
| "loss": 1.4704, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 4.502811776381078, | |
| "grad_norm": 2.1002511978149414, | |
| "learning_rate": 1.1066666666666669e-05, | |
| "loss": 1.4386, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 4.635130664902415, | |
| "eval_loss": 1.4397192001342773, | |
| "eval_runtime": 130.1701, | |
| "eval_samples_per_second": 9.211, | |
| "eval_steps_per_second": 4.609, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 4.767449553423751, | |
| "grad_norm": 2.430173873901367, | |
| "learning_rate": 1.0533333333333333e-05, | |
| "loss": 1.4629, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 5.033079722130334, | |
| "grad_norm": 1.9935014247894287, | |
| "learning_rate": 1e-05, | |
| "loss": 1.4462, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 5.297717499173007, | |
| "grad_norm": 1.5598417520523071, | |
| "learning_rate": 9.466666666666667e-06, | |
| "loss": 1.445, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 5.297717499173007, | |
| "eval_loss": 1.4383418560028076, | |
| "eval_runtime": 130.1186, | |
| "eval_samples_per_second": 9.215, | |
| "eval_steps_per_second": 4.611, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 5.56235527621568, | |
| "grad_norm": 2.314366340637207, | |
| "learning_rate": 8.933333333333333e-06, | |
| "loss": 1.4218, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 5.826993053258352, | |
| "grad_norm": 2.1189815998077393, | |
| "learning_rate": 8.400000000000001e-06, | |
| "loss": 1.4281, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 5.959311941779689, | |
| "eval_loss": 1.4357532262802124, | |
| "eval_runtime": 130.0198, | |
| "eval_samples_per_second": 9.222, | |
| "eval_steps_per_second": 4.615, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 6.092623221964936, | |
| "grad_norm": 2.3731560707092285, | |
| "learning_rate": 7.866666666666667e-06, | |
| "loss": 1.4449, | |
| "step": 4600 | |
| }, | |
| { | |
| "epoch": 6.357260999007608, | |
| "grad_norm": 2.974883556365967, | |
| "learning_rate": 7.333333333333333e-06, | |
| "loss": 1.4458, | |
| "step": 4800 | |
| }, | |
| { | |
| "epoch": 6.621898776050282, | |
| "grad_norm": 2.829678773880005, | |
| "learning_rate": 6.800000000000001e-06, | |
| "loss": 1.4312, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 6.621898776050282, | |
| "eval_loss": 1.435935616493225, | |
| "eval_runtime": 130.3688, | |
| "eval_samples_per_second": 9.197, | |
| "eval_steps_per_second": 4.602, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 6.886536553092954, | |
| "grad_norm": 2.288074493408203, | |
| "learning_rate": 6.266666666666668e-06, | |
| "loss": 1.392, | |
| "step": 5200 | |
| }, | |
| { | |
| "epoch": 7.152166721799537, | |
| "grad_norm": 1.9908281564712524, | |
| "learning_rate": 5.733333333333334e-06, | |
| "loss": 1.4258, | |
| "step": 5400 | |
| }, | |
| { | |
| "epoch": 7.284485610320873, | |
| "eval_loss": 1.4372740983963013, | |
| "eval_runtime": 287.6507, | |
| "eval_samples_per_second": 4.168, | |
| "eval_steps_per_second": 2.086, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 7.416804498842209, | |
| "grad_norm": 2.3958187103271484, | |
| "learning_rate": 5.2e-06, | |
| "loss": 1.4011, | |
| "step": 5600 | |
| }, | |
| { | |
| "epoch": 7.681442275884883, | |
| "grad_norm": 3.133821725845337, | |
| "learning_rate": 4.666666666666667e-06, | |
| "loss": 1.4119, | |
| "step": 5800 | |
| }, | |
| { | |
| "epoch": 7.946080052927555, | |
| "grad_norm": 2.6775546073913574, | |
| "learning_rate": 4.133333333333333e-06, | |
| "loss": 1.4287, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 7.946080052927555, | |
| "eval_loss": 1.43629789352417, | |
| "eval_runtime": 287.6708, | |
| "eval_samples_per_second": 4.168, | |
| "eval_steps_per_second": 2.086, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 8.211710221634139, | |
| "grad_norm": 2.524921417236328, | |
| "learning_rate": 3.6000000000000003e-06, | |
| "loss": 1.3975, | |
| "step": 6200 | |
| }, | |
| { | |
| "epoch": 8.476347998676811, | |
| "grad_norm": 2.997265100479126, | |
| "learning_rate": 3.066666666666667e-06, | |
| "loss": 1.3907, | |
| "step": 6400 | |
| }, | |
| { | |
| "epoch": 8.608666887198147, | |
| "eval_loss": 1.4365167617797852, | |
| "eval_runtime": 288.2825, | |
| "eval_samples_per_second": 4.159, | |
| "eval_steps_per_second": 2.081, | |
| "step": 6500 | |
| } | |
| ], | |
| "logging_steps": 200, | |
| "max_steps": 7550, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 3, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 3 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.1138506359662182e+18, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |