| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9997382884061764, | |
| "eval_steps": 100, | |
| "global_step": 955, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -2.980285167694092, | |
| "logits/rejected": -2.87275767326355, | |
| "logps/chosen": -313.4390563964844, | |
| "logps/rejected": -236.1754150390625, | |
| "loss": 0.6931, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.9767171129220025e-06, | |
| "logits/chosen": -2.8189077377319336, | |
| "logits/rejected": -2.828260660171509, | |
| "logps/chosen": -285.1784362792969, | |
| "logps/rejected": -271.1445617675781, | |
| "loss": 0.6803, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.5738636255264282, | |
| "rewards/chosen": 0.01021653600037098, | |
| "rewards/margins": 0.029710056260228157, | |
| "rewards/rejected": -0.019493522122502327, | |
| "step": 100, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.394644935972061e-06, | |
| "logits/chosen": -2.8255398273468018, | |
| "logits/rejected": -2.813203811645508, | |
| "logps/chosen": -278.2582092285156, | |
| "logps/rejected": -262.1144714355469, | |
| "loss": 0.6236, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.6868749856948853, | |
| "rewards/chosen": 0.05237884446978569, | |
| "rewards/margins": 0.20995216071605682, | |
| "rewards/rejected": -0.15757331252098083, | |
| "step": 200, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 3.812572759022119e-06, | |
| "logits/chosen": -2.809633255004883, | |
| "logits/rejected": -2.80940580368042, | |
| "logps/chosen": -283.6641540527344, | |
| "logps/rejected": -257.23779296875, | |
| "loss": 0.5901, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.6956250071525574, | |
| "rewards/chosen": 0.04034877195954323, | |
| "rewards/margins": 0.3768764138221741, | |
| "rewards/rejected": -0.33652764558792114, | |
| "step": 300, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.2305005820721774e-06, | |
| "logits/chosen": -2.821267604827881, | |
| "logits/rejected": -2.8011868000030518, | |
| "logps/chosen": -284.2022399902344, | |
| "logps/rejected": -267.48358154296875, | |
| "loss": 0.5757, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7081249952316284, | |
| "rewards/chosen": 0.022827474400401115, | |
| "rewards/margins": 0.4984941780567169, | |
| "rewards/rejected": -0.47566673159599304, | |
| "step": 400, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.6484284051222353e-06, | |
| "logits/chosen": -2.7993886470794678, | |
| "logits/rejected": -2.7845959663391113, | |
| "logps/chosen": -270.09637451171875, | |
| "logps/rejected": -256.13458251953125, | |
| "loss": 0.5798, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.6612499952316284, | |
| "rewards/chosen": -0.004339172504842281, | |
| "rewards/margins": 0.4106566905975342, | |
| "rewards/rejected": -0.41499578952789307, | |
| "step": 500, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 2.0663562281722936e-06, | |
| "logits/chosen": -2.824444532394409, | |
| "logits/rejected": -2.808506965637207, | |
| "logps/chosen": -284.5413513183594, | |
| "logps/rejected": -270.1356201171875, | |
| "loss": 0.5643, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.06895674020051956, | |
| "rewards/margins": 0.5562920570373535, | |
| "rewards/rejected": -0.4873352348804474, | |
| "step": 600, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.4842840512223516e-06, | |
| "logits/chosen": -2.8160746097564697, | |
| "logits/rejected": -2.795208692550659, | |
| "logps/chosen": -280.79766845703125, | |
| "logps/rejected": -253.41058349609375, | |
| "loss": 0.5631, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7231249809265137, | |
| "rewards/chosen": 0.07620371133089066, | |
| "rewards/margins": 0.5511507391929626, | |
| "rewards/rejected": -0.474947065114975, | |
| "step": 700, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 9.022118742724098e-07, | |
| "logits/chosen": -2.8168938159942627, | |
| "logits/rejected": -2.8180816173553467, | |
| "logps/chosen": -287.77679443359375, | |
| "logps/rejected": -266.3511657714844, | |
| "loss": 0.5548, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7112500071525574, | |
| "rewards/chosen": 0.09802371263504028, | |
| "rewards/margins": 0.5919383764266968, | |
| "rewards/rejected": -0.49391472339630127, | |
| "step": 800, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 3.2013969732246806e-07, | |
| "logits/chosen": -2.808311700820923, | |
| "logits/rejected": -2.812298536300659, | |
| "logps/chosen": -276.976806640625, | |
| "logps/rejected": -261.3187561035156, | |
| "loss": 0.5689, | |
| "pred_label": 0.0, | |
| "rewards/accuracies": 0.7093750238418579, | |
| "rewards/chosen": 0.05313897505402565, | |
| "rewards/margins": 0.551435649394989, | |
| "rewards/rejected": -0.4982966482639313, | |
| "step": 900, | |
| "use_label": 0.0 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -2.8302695751190186, | |
| "eval_logits/rejected": -2.8291618824005127, | |
| "eval_logps/chosen": -283.5841369628906, | |
| "eval_logps/rejected": -264.64422607421875, | |
| "eval_loss": 0.5655555129051208, | |
| "eval_pred_label": 0.0, | |
| "eval_rewards/accuracies": 0.699999988079071, | |
| "eval_rewards/chosen": 0.06732505559921265, | |
| "eval_rewards/margins": 0.6017746329307556, | |
| "eval_rewards/rejected": -0.534449577331543, | |
| "eval_runtime": 457.335, | |
| "eval_samples_per_second": 4.373, | |
| "eval_steps_per_second": 0.273, | |
| "eval_use_label": 0.0, | |
| "step": 955 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5870625535855118, | |
| "train_runtime": 25384.4334, | |
| "train_samples_per_second": 2.408, | |
| "train_steps_per_second": 0.038 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 955, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |