| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.9884169884169884, | |
| "eval_steps": 500, | |
| "global_step": 144, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06864006864006864, | |
| "grad_norm": 117.0, | |
| "learning_rate": 1.6666666666666667e-05, | |
| "log_odds_chosen": 0.04570715129375458, | |
| "log_odds_ratio": -0.7358157634735107, | |
| "logps/chosen": -0.9796704053878784, | |
| "logps/rejected": -1.0139487981796265, | |
| "loss": 48.3858, | |
| "nll_loss": 1.5120567083358765, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.13728013728013727, | |
| "grad_norm": 43.0, | |
| "learning_rate": 3.3333333333333335e-05, | |
| "log_odds_chosen": 0.12318801879882812, | |
| "log_odds_ratio": -0.7144898176193237, | |
| "logps/chosen": -0.7773754000663757, | |
| "logps/rejected": -0.8364827036857605, | |
| "loss": 40.1791, | |
| "nll_loss": 1.2555955648422241, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.2059202059202059, | |
| "grad_norm": 42.5, | |
| "learning_rate": 5e-05, | |
| "log_odds_chosen": 0.1416315734386444, | |
| "log_odds_ratio": -0.7184887528419495, | |
| "logps/chosen": -0.7499162554740906, | |
| "logps/rejected": -0.8260188102722168, | |
| "loss": 37.8788, | |
| "nll_loss": 1.183712363243103, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.27456027456027454, | |
| "grad_norm": 36.0, | |
| "learning_rate": 4.981488838730066e-05, | |
| "log_odds_chosen": 0.1828576922416687, | |
| "log_odds_ratio": -0.7147494554519653, | |
| "logps/chosen": -0.7645934820175171, | |
| "logps/rejected": -0.8450287580490112, | |
| "loss": 36.3979, | |
| "nll_loss": 1.1374342441558838, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.3432003432003432, | |
| "grad_norm": 40.5, | |
| "learning_rate": 4.926229485393513e-05, | |
| "log_odds_chosen": 0.1581808477640152, | |
| "log_odds_ratio": -0.7021191716194153, | |
| "logps/chosen": -0.7081555724143982, | |
| "logps/rejected": -0.7702702283859253, | |
| "loss": 35.2689, | |
| "nll_loss": 1.1021528244018555, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.4118404118404118, | |
| "grad_norm": 31.875, | |
| "learning_rate": 4.83504027183137e-05, | |
| "log_odds_chosen": 0.09541453421115875, | |
| "log_odds_ratio": -0.7330835461616516, | |
| "logps/chosen": -0.7472664713859558, | |
| "logps/rejected": -0.7989091277122498, | |
| "loss": 34.0556, | |
| "nll_loss": 1.0642387866973877, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.4804804804804805, | |
| "grad_norm": 31.5, | |
| "learning_rate": 4.709271612634298e-05, | |
| "log_odds_chosen": 0.09454293549060822, | |
| "log_odds_ratio": -0.7360396981239319, | |
| "logps/chosen": -0.7235242128372192, | |
| "logps/rejected": -0.7580782771110535, | |
| "loss": 34.7533, | |
| "nll_loss": 1.0860421657562256, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.5491205491205491, | |
| "grad_norm": 36.5, | |
| "learning_rate": 4.550786006948777e-05, | |
| "log_odds_chosen": 0.18448153138160706, | |
| "log_odds_ratio": -0.7088391184806824, | |
| "logps/chosen": -0.6769216060638428, | |
| "logps/rejected": -0.7544242143630981, | |
| "loss": 33.5798, | |
| "nll_loss": 1.0493675470352173, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.6177606177606177, | |
| "grad_norm": 30.375, | |
| "learning_rate": 4.361930456859455e-05, | |
| "log_odds_chosen": 0.2148957997560501, | |
| "log_odds_ratio": -0.7018598318099976, | |
| "logps/chosen": -0.6819745898246765, | |
| "logps/rejected": -0.7681012153625488, | |
| "loss": 32.7666, | |
| "nll_loss": 1.0239557027816772, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.6864006864006864, | |
| "grad_norm": 29.375, | |
| "learning_rate": 4.14550171080187e-05, | |
| "log_odds_chosen": 0.1911373734474182, | |
| "log_odds_ratio": -0.7087734937667847, | |
| "logps/chosen": -0.699134349822998, | |
| "logps/rejected": -0.7818752527236938, | |
| "loss": 33.5611, | |
| "nll_loss": 1.0487844944000244, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.7550407550407551, | |
| "grad_norm": 32.5, | |
| "learning_rate": 3.9047048467134014e-05, | |
| "log_odds_chosen": 0.1783633977174759, | |
| "log_odds_ratio": -0.7113261818885803, | |
| "logps/chosen": -0.7090396881103516, | |
| "logps/rejected": -0.7853409051895142, | |
| "loss": 33.1202, | |
| "nll_loss": 1.035004734992981, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.8236808236808236, | |
| "grad_norm": 30.0, | |
| "learning_rate": 3.6431058082615964e-05, | |
| "log_odds_chosen": 0.18033447861671448, | |
| "log_odds_ratio": -0.7068595290184021, | |
| "logps/chosen": -0.7181381583213806, | |
| "logps/rejected": -0.7968712449073792, | |
| "loss": 33.0624, | |
| "nll_loss": 1.0332015752792358, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.8923208923208923, | |
| "grad_norm": 33.75, | |
| "learning_rate": 3.3645785970375284e-05, | |
| "log_odds_chosen": 0.15356490015983582, | |
| "log_odds_ratio": -0.7268341779708862, | |
| "logps/chosen": -0.6993820071220398, | |
| "logps/rejected": -0.779438853263855, | |
| "loss": 33.6264, | |
| "nll_loss": 1.050825834274292, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.960960960960961, | |
| "grad_norm": 29.0, | |
| "learning_rate": 3.0732479027412256e-05, | |
| "log_odds_chosen": 0.181875079870224, | |
| "log_odds_ratio": -0.7036765813827515, | |
| "logps/chosen": -0.6884555220603943, | |
| "logps/rejected": -0.763231635093689, | |
| "loss": 33.2219, | |
| "nll_loss": 1.0381855964660645, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.0411840411840412, | |
| "grad_norm": 24.5, | |
| "learning_rate": 2.7734280209446865e-05, | |
| "log_odds_chosen": 0.3063170909881592, | |
| "log_odds_ratio": -0.6562883853912354, | |
| "logps/chosen": -0.6133887767791748, | |
| "logps/rejected": -0.7328479290008545, | |
| "loss": 33.71, | |
| "nll_loss": 0.9260991811752319, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 1.1098241098241097, | |
| "grad_norm": 30.5, | |
| "learning_rate": 2.4695589629950063e-05, | |
| "log_odds_chosen": 0.6145213842391968, | |
| "log_odds_ratio": -0.5515514612197876, | |
| "logps/chosen": -0.5365720391273499, | |
| "logps/rejected": -0.7887281179428101, | |
| "loss": 26.4267, | |
| "nll_loss": 0.8258360028266907, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.1784641784641785, | |
| "grad_norm": 27.75, | |
| "learning_rate": 2.1661407042015055e-05, | |
| "log_odds_chosen": 0.6053102016448975, | |
| "log_odds_ratio": -0.5604658126831055, | |
| "logps/chosen": -0.5426899790763855, | |
| "logps/rejected": -0.7706879377365112, | |
| "loss": 26.4214, | |
| "nll_loss": 0.8256683349609375, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 1.247104247104247, | |
| "grad_norm": 28.125, | |
| "learning_rate": 1.867666544020798e-05, | |
| "log_odds_chosen": 0.6666785478591919, | |
| "log_odds_ratio": -0.5315438508987427, | |
| "logps/chosen": -0.49580830335617065, | |
| "logps/rejected": -0.7630711793899536, | |
| "loss": 25.8852, | |
| "nll_loss": 0.808911919593811, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.3157443157443156, | |
| "grad_norm": 30.375, | |
| "learning_rate": 1.5785565651040956e-05, | |
| "log_odds_chosen": 0.6142023801803589, | |
| "log_odds_ratio": -0.5481556057929993, | |
| "logps/chosen": -0.5331859588623047, | |
| "logps/rejected": -0.7694956660270691, | |
| "loss": 25.3047, | |
| "nll_loss": 0.7907715439796448, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 1.3843843843843844, | |
| "grad_norm": 27.5, | |
| "learning_rate": 1.303092176606978e-05, | |
| "log_odds_chosen": 0.7389415502548218, | |
| "log_odds_ratio": -0.49965333938598633, | |
| "logps/chosen": -0.5179293751716614, | |
| "logps/rejected": -0.8284605145454407, | |
| "loss": 26.6145, | |
| "nll_loss": 0.8317017555236816, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.4530244530244532, | |
| "grad_norm": 27.625, | |
| "learning_rate": 1.0453527111051184e-05, | |
| "log_odds_chosen": 0.7348083853721619, | |
| "log_odds_ratio": -0.517388641834259, | |
| "logps/chosen": -0.5103100538253784, | |
| "logps/rejected": -0.7935900688171387, | |
| "loss": 25.6598, | |
| "nll_loss": 0.8018671870231628, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 1.5216645216645217, | |
| "grad_norm": 28.5, | |
| "learning_rate": 8.09155014047743e-06, | |
| "log_odds_chosen": 0.6674805879592896, | |
| "log_odds_ratio": -0.5333995819091797, | |
| "logps/chosen": -0.5412867665290833, | |
| "logps/rejected": -0.8199647068977356, | |
| "loss": 26.3129, | |
| "nll_loss": 0.8222776651382446, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.5903045903045903, | |
| "grad_norm": 27.625, | |
| "learning_rate": 5.9799692036430485e-06, | |
| "log_odds_chosen": 0.5439990758895874, | |
| "log_odds_ratio": -0.5787786841392517, | |
| "logps/chosen": -0.5497544407844543, | |
| "logps/rejected": -0.7601868510246277, | |
| "loss": 26.8215, | |
| "nll_loss": 0.8381711840629578, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 1.6589446589446588, | |
| "grad_norm": 28.875, | |
| "learning_rate": 4.150054552753055e-06, | |
| "log_odds_chosen": 0.6353312730789185, | |
| "log_odds_ratio": -0.5571193695068359, | |
| "logps/chosen": -0.5275944471359253, | |
| "logps/rejected": -0.7502118349075317, | |
| "loss": 26.0757, | |
| "nll_loss": 0.8148647546768188, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.7275847275847276, | |
| "grad_norm": 28.75, | |
| "learning_rate": 2.6289052639777044e-06, | |
| "log_odds_chosen": 0.5769211053848267, | |
| "log_odds_ratio": -0.550868809223175, | |
| "logps/chosen": -0.5242232084274292, | |
| "logps/rejected": -0.7588512301445007, | |
| "loss": 25.6315, | |
| "nll_loss": 0.8009853363037109, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.7962247962247964, | |
| "grad_norm": 33.75, | |
| "learning_rate": 1.439047929157139e-06, | |
| "log_odds_chosen": 0.689731776714325, | |
| "log_odds_ratio": -0.5165315866470337, | |
| "logps/chosen": -0.5240885019302368, | |
| "logps/rejected": -0.8086684942245483, | |
| "loss": 25.9948, | |
| "nll_loss": 0.812338650226593, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.864864864864865, | |
| "grad_norm": 27.0, | |
| "learning_rate": 5.981030611018234e-07, | |
| "log_odds_chosen": 0.6484060883522034, | |
| "log_odds_ratio": -0.5438445806503296, | |
| "logps/chosen": -0.551119863986969, | |
| "logps/rejected": -0.7980315089225769, | |
| "loss": 27.2419, | |
| "nll_loss": 0.8513097763061523, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 1.9335049335049335, | |
| "grad_norm": 28.625, | |
| "learning_rate": 1.1852415266910554e-07, | |
| "log_odds_chosen": 0.7377719879150391, | |
| "log_odds_ratio": -0.502840518951416, | |
| "logps/chosen": -0.5086525082588196, | |
| "logps/rejected": -0.8028281331062317, | |
| "loss": 24.634, | |
| "nll_loss": 0.7698137760162354, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.9884169884169884, | |
| "step": 144, | |
| "total_flos": 0.0, | |
| "train_loss": 30.958674271901447, | |
| "train_runtime": 2313.4467, | |
| "train_samples_per_second": 4.03, | |
| "train_steps_per_second": 0.062 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 144, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |