| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9879931389365352, | |
| "eval_steps": 500, | |
| "global_step": 72, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0686106346483705, | |
| "grad_norm": 19.232778549194336, | |
| "learning_rate": 2.5e-07, | |
| "log_odds_chosen": 1.2014477252960205, | |
| "log_odds_ratio": -0.3934864103794098, | |
| "logps/chosen": -0.419245183467865, | |
| "logps/rejected": -0.8377097845077515, | |
| "loss": 11.1878, | |
| "nll_loss": 0.6684826612472534, | |
| "rewards/accuracies": 0.828125, | |
| "rewards/chosen": -0.02096226066350937, | |
| "rewards/margins": 0.020923227071762085, | |
| "rewards/rejected": -0.041885487735271454, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.137221269296741, | |
| "grad_norm": 16.76073455810547, | |
| "learning_rate": 4.996988640512931e-07, | |
| "log_odds_chosen": 1.282283067703247, | |
| "log_odds_ratio": -0.3863990008831024, | |
| "logps/chosen": -0.3940752148628235, | |
| "logps/rejected": -0.8234782218933105, | |
| "loss": 10.6398, | |
| "nll_loss": 0.6343874335289001, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.019703760743141174, | |
| "rewards/margins": 0.021470148116350174, | |
| "rewards/rejected": -0.04117390885949135, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.2058319039451115, | |
| "grad_norm": 18.169845581054688, | |
| "learning_rate": 4.892350839330522e-07, | |
| "log_odds_chosen": 1.4661200046539307, | |
| "log_odds_ratio": -0.3449063301086426, | |
| "logps/chosen": -0.35424643754959106, | |
| "logps/rejected": -0.8525441884994507, | |
| "loss": 9.6632, | |
| "nll_loss": 0.5748022794723511, | |
| "rewards/accuracies": 0.878125011920929, | |
| "rewards/chosen": -0.017712321132421494, | |
| "rewards/margins": 0.02491488680243492, | |
| "rewards/rejected": -0.042627207934856415, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.274442538593482, | |
| "grad_norm": 15.6996488571167, | |
| "learning_rate": 4.64432152500068e-07, | |
| "log_odds_chosen": 1.7303870916366577, | |
| "log_odds_ratio": -0.30832380056381226, | |
| "logps/chosen": -0.3413855731487274, | |
| "logps/rejected": -0.8977711796760559, | |
| "loss": 8.8048, | |
| "nll_loss": 0.5222700238227844, | |
| "rewards/accuracies": 0.8843749761581421, | |
| "rewards/chosen": -0.01706927828490734, | |
| "rewards/margins": 0.027819275856018066, | |
| "rewards/rejected": -0.04488855600357056, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.34305317324185247, | |
| "grad_norm": 18.215797424316406, | |
| "learning_rate": 4.2677669529663686e-07, | |
| "log_odds_chosen": 1.483076810836792, | |
| "log_odds_ratio": -0.32862988114356995, | |
| "logps/chosen": -0.3441302180290222, | |
| "logps/rejected": -0.8243424296379089, | |
| "loss": 9.7759, | |
| "nll_loss": 0.5818092226982117, | |
| "rewards/accuracies": 0.8968750238418579, | |
| "rewards/chosen": -0.01720651052892208, | |
| "rewards/margins": 0.024010611698031425, | |
| "rewards/rejected": -0.041217125952243805, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.411663807890223, | |
| "grad_norm": 16.51038360595703, | |
| "learning_rate": 3.7852568604830535e-07, | |
| "log_odds_chosen": 1.546562910079956, | |
| "log_odds_ratio": -0.30873748660087585, | |
| "logps/chosen": -0.34601905941963196, | |
| "logps/rejected": -0.8720181584358215, | |
| "loss": 9.2217, | |
| "nll_loss": 0.5480064153671265, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.017300952225923538, | |
| "rewards/margins": 0.026299959048628807, | |
| "rewards/rejected": -0.043600909411907196, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.48027444253859347, | |
| "grad_norm": 16.503437042236328, | |
| "learning_rate": 3.2257116931361555e-07, | |
| "log_odds_chosen": 1.6262096166610718, | |
| "log_odds_ratio": -0.29989489912986755, | |
| "logps/chosen": -0.3171854317188263, | |
| "logps/rejected": -0.8487747311592102, | |
| "loss": 8.9613, | |
| "nll_loss": 0.5319884419441223, | |
| "rewards/accuracies": 0.903124988079071, | |
| "rewards/chosen": -0.015859272330999374, | |
| "rewards/margins": 0.026579465717077255, | |
| "rewards/rejected": -0.04243873432278633, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.548885077186964, | |
| "grad_norm": 16.347063064575195, | |
| "learning_rate": 2.6226691858185454e-07, | |
| "log_odds_chosen": 1.7782882452011108, | |
| "log_odds_ratio": -0.28193312883377075, | |
| "logps/chosen": -0.29190748929977417, | |
| "logps/rejected": -0.8516559600830078, | |
| "loss": 8.6856, | |
| "nll_loss": 0.5151814222335815, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.01459537260234356, | |
| "rewards/margins": 0.027987420558929443, | |
| "rewards/rejected": -0.04258279502391815, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.6174957118353345, | |
| "grad_norm": 18.475563049316406, | |
| "learning_rate": 2.0122741949596793e-07, | |
| "log_odds_chosen": 1.6845076084136963, | |
| "log_odds_ratio": -0.30887991189956665, | |
| "logps/chosen": -0.31784507632255554, | |
| "logps/rejected": -0.852371096611023, | |
| "loss": 8.6938, | |
| "nll_loss": 0.5149728059768677, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": -0.015892256051301956, | |
| "rewards/margins": 0.02672630548477173, | |
| "rewards/rejected": -0.042618557810783386, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.6861063464837049, | |
| "grad_norm": 17.385778427124023, | |
| "learning_rate": 1.4311122664242953e-07, | |
| "log_odds_chosen": 1.7722632884979248, | |
| "log_odds_ratio": -0.2871997356414795, | |
| "logps/chosen": -0.3112233579158783, | |
| "logps/rejected": -0.8770394325256348, | |
| "loss": 8.9052, | |
| "nll_loss": 0.5289064049720764, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.01556116808205843, | |
| "rewards/margins": 0.028290802612900734, | |
| "rewards/rejected": -0.04385197162628174, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.7547169811320755, | |
| "grad_norm": 17.232837677001953, | |
| "learning_rate": 9.140167895908865e-08, | |
| "log_odds_chosen": 1.7312713861465454, | |
| "log_odds_ratio": -0.28886836767196655, | |
| "logps/chosen": -0.323580265045166, | |
| "logps/rejected": -0.8828792572021484, | |
| "loss": 8.9703, | |
| "nll_loss": 0.5326738357543945, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.01617901399731636, | |
| "rewards/margins": 0.02796494960784912, | |
| "rewards/rejected": -0.04414396360516548, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.823327615780446, | |
| "grad_norm": 16.895586013793945, | |
| "learning_rate": 4.919811712983879e-08, | |
| "log_odds_chosen": 1.7012507915496826, | |
| "log_odds_ratio": -0.29369616508483887, | |
| "logps/chosen": -0.33875319361686707, | |
| "logps/rejected": -0.8981720805168152, | |
| "loss": 9.0854, | |
| "nll_loss": 0.5401446223258972, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -0.016937660053372383, | |
| "rewards/margins": 0.027970943599939346, | |
| "rewards/rejected": -0.04490860179066658, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.8919382504288165, | |
| "grad_norm": 18.454177856445312, | |
| "learning_rate": 1.9030116872178314e-08, | |
| "log_odds_chosen": 1.8286237716674805, | |
| "log_odds_ratio": -0.2700421214103699, | |
| "logps/chosen": -0.3031145930290222, | |
| "logps/rejected": -0.8966620564460754, | |
| "loss": 8.785, | |
| "nll_loss": 0.5219055414199829, | |
| "rewards/accuracies": 0.9156249761581421, | |
| "rewards/chosen": -0.0151557307690382, | |
| "rewards/margins": 0.0296773761510849, | |
| "rewards/rejected": -0.04483310505747795, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.9605488850771869, | |
| "grad_norm": 17.21074104309082, | |
| "learning_rate": 2.7058725088047464e-09, | |
| "log_odds_chosen": 1.7211929559707642, | |
| "log_odds_ratio": -0.28593194484710693, | |
| "logps/chosen": -0.3278438448905945, | |
| "logps/rejected": -0.8622439503669739, | |
| "loss": 9.2475, | |
| "nll_loss": 0.5499622821807861, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -0.016392191872000694, | |
| "rewards/margins": 0.02672000229358673, | |
| "rewards/rejected": -0.043112196028232574, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.9879931389365352, | |
| "step": 72, | |
| "total_flos": 0.0, | |
| "train_loss": 9.334032826953464, | |
| "train_runtime": 762.9212, | |
| "train_samples_per_second": 6.111, | |
| "train_steps_per_second": 0.094 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 72, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "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 | |
| } | |