| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9945, | |
| "eval_steps": 500, | |
| "global_step": 153, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 17.99154727967293, | |
| "learning_rate": 3.125e-08, | |
| "logits/chosen": -2.152977705001831, | |
| "logits/rejected": -2.3121213912963867, | |
| "logps/chosen": -254.60496520996094, | |
| "logps/rejected": -224.12643432617188, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 15.888543656019479, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.4332120418548584, | |
| "logits/rejected": -2.349087953567505, | |
| "logps/chosen": -218.75491333007812, | |
| "logps/rejected": -224.596923828125, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.43589743971824646, | |
| "rewards/chosen": -0.0051579843275249004, | |
| "rewards/margins": 0.0010460736230015755, | |
| "rewards/rejected": -0.006204057950526476, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 18.00599863061672, | |
| "learning_rate": 4.989490450759331e-07, | |
| "logits/chosen": -2.44968318939209, | |
| "logits/rejected": -2.3714213371276855, | |
| "logps/chosen": -227.60235595703125, | |
| "logps/rejected": -245.01002502441406, | |
| "loss": 0.6881, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.2174525260925293, | |
| "rewards/margins": 0.025719773024320602, | |
| "rewards/rejected": -0.243172287940979, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 17.361907335413566, | |
| "learning_rate": 4.872270441827174e-07, | |
| "logits/chosen": -2.3445615768432617, | |
| "logits/rejected": -2.285738706588745, | |
| "logps/chosen": -264.4424743652344, | |
| "logps/rejected": -267.0367431640625, | |
| "loss": 0.6952, | |
| "rewards/accuracies": 0.4923076927661896, | |
| "rewards/chosen": -0.6213539838790894, | |
| "rewards/margins": 0.022094279527664185, | |
| "rewards/rejected": -0.6434482336044312, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 17.40493821515386, | |
| "learning_rate": 4.6308512113530063e-07, | |
| "logits/chosen": -2.5398178100585938, | |
| "logits/rejected": -2.436666488647461, | |
| "logps/chosen": -252.19931030273438, | |
| "logps/rejected": -264.8144226074219, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.5730769038200378, | |
| "rewards/chosen": -0.43709850311279297, | |
| "rewards/margins": 0.043844155967235565, | |
| "rewards/rejected": -0.4809426963329315, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 14.207926284792292, | |
| "learning_rate": 4.277872161641681e-07, | |
| "logits/chosen": -2.579521894454956, | |
| "logits/rejected": -2.5611038208007812, | |
| "logps/chosen": -251.29273986816406, | |
| "logps/rejected": -248.70640563964844, | |
| "loss": 0.6889, | |
| "rewards/accuracies": 0.5269230604171753, | |
| "rewards/chosen": -0.18656322360038757, | |
| "rewards/margins": 0.01753987930715084, | |
| "rewards/rejected": -0.20410311222076416, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 16.71668002097283, | |
| "learning_rate": 3.8318133624280046e-07, | |
| "logits/chosen": -2.5447630882263184, | |
| "logits/rejected": -2.5085957050323486, | |
| "logps/chosen": -254.7252960205078, | |
| "logps/rejected": -250.2134552001953, | |
| "loss": 0.6877, | |
| "rewards/accuracies": 0.4769230782985687, | |
| "rewards/chosen": -0.20926551520824432, | |
| "rewards/margins": 0.024503052234649658, | |
| "rewards/rejected": -0.23376856744289398, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 15.128093177880084, | |
| "learning_rate": 3.316028034595861e-07, | |
| "logits/chosen": -2.4117591381073, | |
| "logits/rejected": -2.3856163024902344, | |
| "logps/chosen": -240.95860290527344, | |
| "logps/rejected": -251.49269104003906, | |
| "loss": 0.6945, | |
| "rewards/accuracies": 0.5692307949066162, | |
| "rewards/chosen": -0.2835744321346283, | |
| "rewards/margins": 0.03665730357170105, | |
| "rewards/rejected": -0.32023176550865173, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 15.984185802661708, | |
| "learning_rate": 2.7575199021178855e-07, | |
| "logits/chosen": -2.502197742462158, | |
| "logits/rejected": -2.4183943271636963, | |
| "logps/chosen": -259.74285888671875, | |
| "logps/rejected": -259.47991943359375, | |
| "loss": 0.6796, | |
| "rewards/accuracies": 0.557692289352417, | |
| "rewards/chosen": -0.22889916598796844, | |
| "rewards/margins": 0.03880741447210312, | |
| "rewards/rejected": -0.26770660281181335, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 16.944533463136676, | |
| "learning_rate": 2.1855294234408068e-07, | |
| "logits/chosen": -2.4329493045806885, | |
| "logits/rejected": -2.377147674560547, | |
| "logps/chosen": -238.77903747558594, | |
| "logps/rejected": -275.1826171875, | |
| "loss": 0.6783, | |
| "rewards/accuracies": 0.5692307949066162, | |
| "rewards/chosen": -0.33898892998695374, | |
| "rewards/margins": 0.07301792502403259, | |
| "rewards/rejected": -0.41200685501098633, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 20.171571457821667, | |
| "learning_rate": 1.6300029195778453e-07, | |
| "logits/chosen": -2.404921770095825, | |
| "logits/rejected": -2.3588438034057617, | |
| "logps/chosen": -249.01068115234375, | |
| "logps/rejected": -261.3023376464844, | |
| "loss": 0.671, | |
| "rewards/accuracies": 0.5807692408561707, | |
| "rewards/chosen": -0.362224817276001, | |
| "rewards/margins": 0.0861731767654419, | |
| "rewards/rejected": -0.4483979642391205, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 22.208608685833642, | |
| "learning_rate": 1.1200247470632392e-07, | |
| "logits/chosen": -2.384084939956665, | |
| "logits/rejected": -2.4500632286071777, | |
| "logps/chosen": -260.6557312011719, | |
| "logps/rejected": -259.0586853027344, | |
| "loss": 0.6855, | |
| "rewards/accuracies": 0.4653846025466919, | |
| "rewards/chosen": -0.40450161695480347, | |
| "rewards/margins": 0.021371768787503242, | |
| "rewards/rejected": -0.42587342858314514, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 21.198358455810045, | |
| "learning_rate": 6.822945986946385e-08, | |
| "logits/chosen": -2.3326570987701416, | |
| "logits/rejected": -2.2644848823547363, | |
| "logps/chosen": -269.0829772949219, | |
| "logps/rejected": -277.46099853515625, | |
| "loss": 0.6829, | |
| "rewards/accuracies": 0.5692307949066162, | |
| "rewards/chosen": -0.45251235365867615, | |
| "rewards/margins": 0.05605296790599823, | |
| "rewards/rejected": -0.508565366268158, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 24.487908321616903, | |
| "learning_rate": 3.397296523427806e-08, | |
| "logits/chosen": -2.3290350437164307, | |
| "logits/rejected": -2.368476629257202, | |
| "logps/chosen": -264.3924255371094, | |
| "logps/rejected": -283.0903625488281, | |
| "loss": 0.6724, | |
| "rewards/accuracies": 0.5346153974533081, | |
| "rewards/chosen": -0.5441482663154602, | |
| "rewards/margins": 0.08576709777116776, | |
| "rewards/rejected": -0.6299152970314026, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 24.363492054663, | |
| "learning_rate": 1.1026475173977978e-08, | |
| "logits/chosen": -2.415417194366455, | |
| "logits/rejected": -2.2751691341400146, | |
| "logps/chosen": -280.5184326171875, | |
| "logps/rejected": -270.88372802734375, | |
| "loss": 0.683, | |
| "rewards/accuracies": 0.5807692408561707, | |
| "rewards/chosen": -0.5419987440109253, | |
| "rewards/margins": 0.07570147514343262, | |
| "rewards/rejected": -0.6177002787590027, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 24.646600457502558, | |
| "learning_rate": 5.913435276374834e-10, | |
| "logits/chosen": -2.4706640243530273, | |
| "logits/rejected": -2.419950485229492, | |
| "logps/chosen": -264.6922302246094, | |
| "logps/rejected": -289.4066467285156, | |
| "loss": 0.6716, | |
| "rewards/accuracies": 0.6038461327552795, | |
| "rewards/chosen": -0.4906882047653198, | |
| "rewards/margins": 0.13459746539592743, | |
| "rewards/rejected": -0.6252856850624084, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "step": 153, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6840333373718013, | |
| "train_runtime": 39783.2449, | |
| "train_samples_per_second": 0.503, | |
| "train_steps_per_second": 0.004 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 153, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |