| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9221902017291066, | |
| "eval_steps": 500, | |
| "global_step": 10, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.09221902017291066, | |
| "grad_norm": 0.37403297424316406, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": 0.10453951358795166, | |
| "logits/rejected": 0.2759632468223572, | |
| "logps/chosen": -132.67144775390625, | |
| "logps/ref_chosen": -133.03013610839844, | |
| "logps/ref_rejected": -164.177734375, | |
| "logps/rejected": -163.5892333984375, | |
| "loss": 0.6943, | |
| "rewards/accuracies": 0.3984375, | |
| "rewards/chosen": 0.003586653620004654, | |
| "rewards/margins": -0.002298325300216675, | |
| "rewards/rejected": 0.005884978454560041, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.1844380403458213, | |
| "grad_norm": 0.358889102935791, | |
| "learning_rate": 4.849231551964771e-07, | |
| "logits/chosen": 0.08250421285629272, | |
| "logits/rejected": 0.33024948835372925, | |
| "logps/chosen": -134.9761962890625, | |
| "logps/ref_chosen": -135.27749633789062, | |
| "logps/ref_rejected": -188.41795349121094, | |
| "logps/rejected": -187.88470458984375, | |
| "loss": 0.6943, | |
| "rewards/accuracies": 0.3515625, | |
| "rewards/chosen": 0.003012962406501174, | |
| "rewards/margins": -0.0023194574750959873, | |
| "rewards/rejected": 0.005332420114427805, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.276657060518732, | |
| "grad_norm": 0.3239206075668335, | |
| "learning_rate": 4.415111107797445e-07, | |
| "logits/chosen": 0.08983182162046432, | |
| "logits/rejected": 0.2510722875595093, | |
| "logps/chosen": -134.5572052001953, | |
| "logps/ref_chosen": -134.77491760253906, | |
| "logps/ref_rejected": -161.10980224609375, | |
| "logps/rejected": -160.8002471923828, | |
| "loss": 0.6936, | |
| "rewards/accuracies": 0.453125, | |
| "rewards/chosen": 0.002177180489525199, | |
| "rewards/margins": -0.0009184688096866012, | |
| "rewards/rejected": 0.003095649415627122, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.3688760806916426, | |
| "grad_norm": 0.3113608658313751, | |
| "learning_rate": 3.75e-07, | |
| "logits/chosen": 0.18616041541099548, | |
| "logits/rejected": 0.3378028869628906, | |
| "logps/chosen": -141.66685485839844, | |
| "logps/ref_chosen": -142.0138702392578, | |
| "logps/ref_rejected": -173.76629638671875, | |
| "logps/rejected": -173.24481201171875, | |
| "loss": 0.694, | |
| "rewards/accuracies": 0.390625, | |
| "rewards/chosen": 0.0034699777606874704, | |
| "rewards/margins": -0.0017446475103497505, | |
| "rewards/rejected": 0.005214625503867865, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.4610951008645533, | |
| "grad_norm": 0.3619195818901062, | |
| "learning_rate": 2.934120444167326e-07, | |
| "logits/chosen": 0.1221667155623436, | |
| "logits/rejected": 0.268534779548645, | |
| "logps/chosen": -134.57679748535156, | |
| "logps/ref_chosen": -134.8294677734375, | |
| "logps/ref_rejected": -177.42715454101562, | |
| "logps/rejected": -177.0855712890625, | |
| "loss": 0.6936, | |
| "rewards/accuracies": 0.4609375, | |
| "rewards/chosen": 0.002526558004319668, | |
| "rewards/margins": -0.0008892094483599067, | |
| "rewards/rejected": 0.00341576780192554, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.553314121037464, | |
| "grad_norm": 0.3865343928337097, | |
| "learning_rate": 2.065879555832674e-07, | |
| "logits/chosen": 0.19299811124801636, | |
| "logits/rejected": 0.30019116401672363, | |
| "logps/chosen": -135.14425659179688, | |
| "logps/ref_chosen": -135.45623779296875, | |
| "logps/ref_rejected": -159.72341918945312, | |
| "logps/rejected": -159.44705200195312, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.515625, | |
| "rewards/chosen": 0.003119847271591425, | |
| "rewards/margins": 0.00035619616392068565, | |
| "rewards/rejected": 0.002763650845736265, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.6455331412103746, | |
| "grad_norm": 0.40479081869125366, | |
| "learning_rate": 1.2500000000000005e-07, | |
| "logits/chosen": 0.15088775753974915, | |
| "logits/rejected": 0.3000352084636688, | |
| "logps/chosen": -139.46456909179688, | |
| "logps/ref_chosen": -139.45156860351562, | |
| "logps/ref_rejected": -172.6890869140625, | |
| "logps/rejected": -172.62013244628906, | |
| "loss": 0.6936, | |
| "rewards/accuracies": 0.4765625, | |
| "rewards/chosen": -0.0001300960429944098, | |
| "rewards/margins": -0.000819505425170064, | |
| "rewards/rejected": 0.0006894093239679933, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.7377521613832853, | |
| "grad_norm": 0.3541257679462433, | |
| "learning_rate": 5.848888922025552e-08, | |
| "logits/chosen": 0.15152569115161896, | |
| "logits/rejected": 0.2909863293170929, | |
| "logps/chosen": -133.0579376220703, | |
| "logps/ref_chosen": -133.19911193847656, | |
| "logps/ref_rejected": -167.17926025390625, | |
| "logps/rejected": -167.0704803466797, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4609375, | |
| "rewards/chosen": 0.0014116661623120308, | |
| "rewards/margins": 0.00032381698838435113, | |
| "rewards/rejected": 0.0010878491448238492, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.829971181556196, | |
| "grad_norm": 0.36632558703422546, | |
| "learning_rate": 1.507684480352292e-08, | |
| "logits/chosen": 0.0733647495508194, | |
| "logits/rejected": 0.21200355887413025, | |
| "logps/chosen": -137.79318237304688, | |
| "logps/ref_chosen": -137.95736694335938, | |
| "logps/ref_rejected": -172.7656707763672, | |
| "logps/rejected": -172.64659118652344, | |
| "loss": 0.6929, | |
| "rewards/accuracies": 0.53125, | |
| "rewards/chosen": 0.0016417349688708782, | |
| "rewards/margins": 0.00045086207683198154, | |
| "rewards/rejected": 0.0011908727465197444, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.9221902017291066, | |
| "grad_norm": 0.33898451924324036, | |
| "learning_rate": 0.0, | |
| "logits/chosen": 0.10225574672222137, | |
| "logits/rejected": 0.29343536496162415, | |
| "logps/chosen": -135.6978759765625, | |
| "logps/ref_chosen": -135.66470336914062, | |
| "logps/ref_rejected": -178.8431854248047, | |
| "logps/rejected": -178.8609619140625, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.4453125, | |
| "rewards/chosen": -0.0003318000235594809, | |
| "rewards/margins": -0.00015407620230689645, | |
| "rewards/rejected": -0.00017772376304492354, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.9221902017291066, | |
| "step": 10, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6935525000095367, | |
| "train_runtime": 687.8249, | |
| "train_samples_per_second": 2.017, | |
| "train_steps_per_second": 0.015 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 10, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |