| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 1000, | |
| "global_step": 19818, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02522958926228681, | |
| "grad_norm": 1.6351510286331177, | |
| "learning_rate": 4.873852053688566e-05, | |
| "loss": 4.2934, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.05045917852457362, | |
| "grad_norm": 2.1481850147247314, | |
| "learning_rate": 4.747704107377132e-05, | |
| "loss": 2.8305, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.05045917852457362, | |
| "eval_accuracy": 0.46500522497769375, | |
| "eval_loss": 2.379011869430542, | |
| "eval_runtime": 56.6734, | |
| "eval_samples_per_second": 112.469, | |
| "eval_steps_per_second": 3.529, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.07568876778686043, | |
| "grad_norm": 1.5696136951446533, | |
| "learning_rate": 4.6215561610656984e-05, | |
| "loss": 2.2245, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.10091835704914724, | |
| "grad_norm": 1.5901012420654297, | |
| "learning_rate": 4.4954082147542644e-05, | |
| "loss": 1.9272, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.10091835704914724, | |
| "eval_accuracy": 0.5808557246708203, | |
| "eval_loss": 1.750130534172058, | |
| "eval_runtime": 55.5938, | |
| "eval_samples_per_second": 114.653, | |
| "eval_steps_per_second": 3.598, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.12614794631143406, | |
| "grad_norm": 1.5128012895584106, | |
| "learning_rate": 4.3692602684428305e-05, | |
| "loss": 1.7748, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 0.15137753557372086, | |
| "grad_norm": 1.318622350692749, | |
| "learning_rate": 4.243112322131396e-05, | |
| "loss": 1.6793, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.15137753557372086, | |
| "eval_accuracy": 0.6136108291841765, | |
| "eval_loss": 1.568946123123169, | |
| "eval_runtime": 54.4039, | |
| "eval_samples_per_second": 117.161, | |
| "eval_steps_per_second": 3.676, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 0.17660712483600766, | |
| "grad_norm": 1.4798802137374878, | |
| "learning_rate": 4.116964375819962e-05, | |
| "loss": 1.613, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 0.20183671409829448, | |
| "grad_norm": 1.2613641023635864, | |
| "learning_rate": 3.990816429508528e-05, | |
| "loss": 1.5605, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.20183671409829448, | |
| "eval_accuracy": 0.6316407595495017, | |
| "eval_loss": 1.4698580503463745, | |
| "eval_runtime": 55.0048, | |
| "eval_samples_per_second": 115.881, | |
| "eval_steps_per_second": 3.636, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 0.22706630336058128, | |
| "grad_norm": 1.1746481657028198, | |
| "learning_rate": 3.864668483197093e-05, | |
| "loss": 1.5228, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 0.2522958926228681, | |
| "grad_norm": 1.2291498184204102, | |
| "learning_rate": 3.738520536885659e-05, | |
| "loss": 1.4891, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.2522958926228681, | |
| "eval_accuracy": 0.642200060669245, | |
| "eval_loss": 1.4112207889556885, | |
| "eval_runtime": 54.5565, | |
| "eval_samples_per_second": 116.833, | |
| "eval_steps_per_second": 3.666, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 0.2775254818851549, | |
| "grad_norm": 1.257660150527954, | |
| "learning_rate": 3.6123725905742254e-05, | |
| "loss": 1.4607, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 0.3027550711474417, | |
| "grad_norm": 1.1654913425445557, | |
| "learning_rate": 3.4862246442627914e-05, | |
| "loss": 1.4391, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.3027550711474417, | |
| "eval_accuracy": 0.6513591843207115, | |
| "eval_loss": 1.364630103111267, | |
| "eval_runtime": 54.7084, | |
| "eval_samples_per_second": 116.509, | |
| "eval_steps_per_second": 3.656, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 0.3279846604097285, | |
| "grad_norm": 1.112001657485962, | |
| "learning_rate": 3.3600766979513575e-05, | |
| "loss": 1.4193, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 0.3532142496720153, | |
| "grad_norm": 1.1872960329055786, | |
| "learning_rate": 3.2339287516399235e-05, | |
| "loss": 1.3995, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.3532142496720153, | |
| "eval_accuracy": 0.6575222655822269, | |
| "eval_loss": 1.3316864967346191, | |
| "eval_runtime": 54.512, | |
| "eval_samples_per_second": 116.928, | |
| "eval_steps_per_second": 3.669, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 0.37844383893430217, | |
| "grad_norm": 1.129939079284668, | |
| "learning_rate": 3.1077808053284896e-05, | |
| "loss": 1.3852, | |
| "step": 7500 | |
| }, | |
| { | |
| "epoch": 0.40367342819658897, | |
| "grad_norm": 1.1093947887420654, | |
| "learning_rate": 2.9816328590170556e-05, | |
| "loss": 1.3707, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.40367342819658897, | |
| "eval_accuracy": 0.663146746266679, | |
| "eval_loss": 1.3020933866500854, | |
| "eval_runtime": 54.3588, | |
| "eval_samples_per_second": 117.258, | |
| "eval_steps_per_second": 3.679, | |
| "step": 8000 | |
| }, | |
| { | |
| "epoch": 0.42890301745887577, | |
| "grad_norm": 1.1245774030685425, | |
| "learning_rate": 2.855484912705621e-05, | |
| "loss": 1.3574, | |
| "step": 8500 | |
| }, | |
| { | |
| "epoch": 0.45413260672116257, | |
| "grad_norm": 1.1140567064285278, | |
| "learning_rate": 2.729336966394187e-05, | |
| "loss": 1.3424, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.45413260672116257, | |
| "eval_accuracy": 0.6674909770600935, | |
| "eval_loss": 1.2805943489074707, | |
| "eval_runtime": 54.0957, | |
| "eval_samples_per_second": 117.828, | |
| "eval_steps_per_second": 3.697, | |
| "step": 9000 | |
| }, | |
| { | |
| "epoch": 0.47936219598344937, | |
| "grad_norm": 1.1032965183258057, | |
| "learning_rate": 2.603189020082753e-05, | |
| "loss": 1.3331, | |
| "step": 9500 | |
| }, | |
| { | |
| "epoch": 0.5045917852457362, | |
| "grad_norm": 1.1406043767929077, | |
| "learning_rate": 2.477041073771319e-05, | |
| "loss": 1.3242, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.5045917852457362, | |
| "eval_accuracy": 0.6713588714661621, | |
| "eval_loss": 1.2613232135772705, | |
| "eval_runtime": 54.8004, | |
| "eval_samples_per_second": 116.313, | |
| "eval_steps_per_second": 3.65, | |
| "step": 10000 | |
| }, | |
| { | |
| "epoch": 0.529821374508023, | |
| "grad_norm": 1.0637890100479126, | |
| "learning_rate": 2.350893127459885e-05, | |
| "loss": 1.316, | |
| "step": 10500 | |
| }, | |
| { | |
| "epoch": 0.5550509637703098, | |
| "grad_norm": 1.0851575136184692, | |
| "learning_rate": 2.224745181148451e-05, | |
| "loss": 1.3058, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.5550509637703098, | |
| "eval_accuracy": 0.6747553370072272, | |
| "eval_loss": 1.2435107231140137, | |
| "eval_runtime": 54.5981, | |
| "eval_samples_per_second": 116.744, | |
| "eval_steps_per_second": 3.663, | |
| "step": 11000 | |
| }, | |
| { | |
| "epoch": 0.5802805530325966, | |
| "grad_norm": 1.1059494018554688, | |
| "learning_rate": 2.098597234837017e-05, | |
| "loss": 1.298, | |
| "step": 11500 | |
| }, | |
| { | |
| "epoch": 0.6055101422948834, | |
| "grad_norm": 1.1078968048095703, | |
| "learning_rate": 1.972449288525583e-05, | |
| "loss": 1.2888, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.6055101422948834, | |
| "eval_accuracy": 0.6777041444946341, | |
| "eval_loss": 1.229060173034668, | |
| "eval_runtime": 55.4356, | |
| "eval_samples_per_second": 114.98, | |
| "eval_steps_per_second": 3.608, | |
| "step": 12000 | |
| }, | |
| { | |
| "epoch": 0.6307397315571702, | |
| "grad_norm": 1.1068435907363892, | |
| "learning_rate": 1.846301342214149e-05, | |
| "loss": 1.2855, | |
| "step": 12500 | |
| }, | |
| { | |
| "epoch": 0.655969320819457, | |
| "grad_norm": 1.0919195413589478, | |
| "learning_rate": 1.7201533959027147e-05, | |
| "loss": 1.2748, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.655969320819457, | |
| "eval_accuracy": 0.6801306075727364, | |
| "eval_loss": 1.2177754640579224, | |
| "eval_runtime": 55.3404, | |
| "eval_samples_per_second": 115.178, | |
| "eval_steps_per_second": 3.614, | |
| "step": 13000 | |
| }, | |
| { | |
| "epoch": 0.6811989100817438, | |
| "grad_norm": 1.0631417036056519, | |
| "learning_rate": 1.5940054495912807e-05, | |
| "loss": 1.2717, | |
| "step": 13500 | |
| }, | |
| { | |
| "epoch": 0.7064284993440306, | |
| "grad_norm": 1.1203536987304688, | |
| "learning_rate": 1.4678575032798466e-05, | |
| "loss": 1.2654, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.7064284993440306, | |
| "eval_accuracy": 0.6820543563309032, | |
| "eval_loss": 1.2076771259307861, | |
| "eval_runtime": 54.4302, | |
| "eval_samples_per_second": 117.104, | |
| "eval_steps_per_second": 3.674, | |
| "step": 14000 | |
| }, | |
| { | |
| "epoch": 0.7316580886063175, | |
| "grad_norm": 1.0601987838745117, | |
| "learning_rate": 1.3417095569684127e-05, | |
| "loss": 1.2592, | |
| "step": 14500 | |
| }, | |
| { | |
| "epoch": 0.7568876778686043, | |
| "grad_norm": 1.0845381021499634, | |
| "learning_rate": 1.2155616106569785e-05, | |
| "loss": 1.2549, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.7568876778686043, | |
| "eval_accuracy": 0.6842664220266779, | |
| "eval_loss": 1.196437120437622, | |
| "eval_runtime": 54.7305, | |
| "eval_samples_per_second": 116.462, | |
| "eval_steps_per_second": 3.654, | |
| "step": 15000 | |
| }, | |
| { | |
| "epoch": 0.7821172671308911, | |
| "grad_norm": 1.1175463199615479, | |
| "learning_rate": 1.0894136643455446e-05, | |
| "loss": 1.2499, | |
| "step": 15500 | |
| }, | |
| { | |
| "epoch": 0.8073468563931779, | |
| "grad_norm": 1.1143847703933716, | |
| "learning_rate": 9.632657180341105e-06, | |
| "loss": 1.2459, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.8073468563931779, | |
| "eval_accuracy": 0.6860334367900387, | |
| "eval_loss": 1.1877895593643188, | |
| "eval_runtime": 54.8073, | |
| "eval_samples_per_second": 116.298, | |
| "eval_steps_per_second": 3.649, | |
| "step": 16000 | |
| }, | |
| { | |
| "epoch": 0.8325764456554647, | |
| "grad_norm": 1.0714515447616577, | |
| "learning_rate": 8.371177717226763e-06, | |
| "loss": 1.2401, | |
| "step": 16500 | |
| }, | |
| { | |
| "epoch": 0.8578060349177515, | |
| "grad_norm": 1.0937442779541016, | |
| "learning_rate": 7.109698254112424e-06, | |
| "loss": 1.2385, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 0.8578060349177515, | |
| "eval_accuracy": 0.6873077056382217, | |
| "eval_loss": 1.1818801164627075, | |
| "eval_runtime": 54.4299, | |
| "eval_samples_per_second": 117.105, | |
| "eval_steps_per_second": 3.674, | |
| "step": 17000 | |
| }, | |
| { | |
| "epoch": 0.8830356241800383, | |
| "grad_norm": 1.125985026359558, | |
| "learning_rate": 5.848218790998083e-06, | |
| "loss": 1.2335, | |
| "step": 17500 | |
| }, | |
| { | |
| "epoch": 0.9082652134423251, | |
| "grad_norm": 1.0988380908966064, | |
| "learning_rate": 4.586739327883742e-06, | |
| "loss": 1.2286, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 0.9082652134423251, | |
| "eval_accuracy": 0.6885514558318389, | |
| "eval_loss": 1.1755918264389038, | |
| "eval_runtime": 54.168, | |
| "eval_samples_per_second": 117.671, | |
| "eval_steps_per_second": 3.692, | |
| "step": 18000 | |
| }, | |
| { | |
| "epoch": 0.9334948027046119, | |
| "grad_norm": 1.0741068124771118, | |
| "learning_rate": 3.325259864769402e-06, | |
| "loss": 1.2284, | |
| "step": 18500 | |
| }, | |
| { | |
| "epoch": 0.9587243919668987, | |
| "grad_norm": 1.1670308113098145, | |
| "learning_rate": 2.063780401655061e-06, | |
| "loss": 1.2246, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 0.9587243919668987, | |
| "eval_accuracy": 0.6896297305064778, | |
| "eval_loss": 1.1708202362060547, | |
| "eval_runtime": 54.2124, | |
| "eval_samples_per_second": 117.575, | |
| "eval_steps_per_second": 3.689, | |
| "step": 19000 | |
| }, | |
| { | |
| "epoch": 0.9839539812291855, | |
| "grad_norm": 1.1070556640625, | |
| "learning_rate": 8.023009385407206e-07, | |
| "loss": 1.2213, | |
| "step": 19500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 19818, | |
| "total_flos": 3.31395116433408e+17, | |
| "train_loss": 1.5000046812714452, | |
| "train_runtime": 7707.6508, | |
| "train_samples_per_second": 82.275, | |
| "train_steps_per_second": 2.571 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 19818, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 1000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.31395116433408e+17, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |