| { | |
| "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, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 18.026980953188083, | |
| "learning_rate": 3.125e-08, | |
| "logits/chosen": -1.4277913570404053, | |
| "logits/rejected": -1.5556963682174683, | |
| "logps/chosen": -139.25473022460938, | |
| "logps/pi_response": -223.89479064941406, | |
| "logps/ref_response": -223.89479064941406, | |
| "logps/rejected": -126.51285552978516, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "eta": 0.0009999999310821295, | |
| "grad_norm": 16.443972323199848, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -1.7425695657730103, | |
| "logits/rejected": -1.763687014579773, | |
| "logps/chosen": -179.83900451660156, | |
| "logps/pi_response": -274.66912841796875, | |
| "logps/ref_response": -272.7638854980469, | |
| "logps/rejected": -183.200927734375, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.47863247990608215, | |
| "rewards/chosen": -0.009404909797012806, | |
| "rewards/margins": 0.0002608599897939712, | |
| "rewards/rejected": -0.009665770456194878, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 20.67365935289489, | |
| "learning_rate": 4.989490450759331e-07, | |
| "logits/chosen": -1.660647988319397, | |
| "logits/rejected": -1.5326403379440308, | |
| "logps/chosen": -199.7965087890625, | |
| "logps/pi_response": -299.9808654785156, | |
| "logps/ref_response": -270.13507080078125, | |
| "logps/rejected": -201.1602783203125, | |
| "loss": 0.6905, | |
| "rewards/accuracies": 0.5730769038200378, | |
| "rewards/chosen": -0.12711657583713531, | |
| "rewards/margins": 0.027275390923023224, | |
| "rewards/rejected": -0.15439198911190033, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 16.68985237393775, | |
| "learning_rate": 4.872270441827174e-07, | |
| "logits/chosen": -1.4958996772766113, | |
| "logits/rejected": -1.5024709701538086, | |
| "logps/chosen": -192.6185302734375, | |
| "logps/pi_response": -322.8236389160156, | |
| "logps/ref_response": -270.1427917480469, | |
| "logps/rejected": -197.94105529785156, | |
| "loss": 0.6956, | |
| "rewards/accuracies": 0.5192307829856873, | |
| "rewards/chosen": -0.22080166637897491, | |
| "rewards/margins": 0.01529843732714653, | |
| "rewards/rejected": -0.23610009253025055, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 18.609452469800672, | |
| "learning_rate": 4.6308512113530063e-07, | |
| "logits/chosen": -1.47615385055542, | |
| "logits/rejected": -1.4204249382019043, | |
| "logps/chosen": -221.22259521484375, | |
| "logps/pi_response": -322.4052429199219, | |
| "logps/ref_response": -275.0693359375, | |
| "logps/rejected": -216.37770080566406, | |
| "loss": 0.6913, | |
| "rewards/accuracies": 0.5192307829856873, | |
| "rewards/chosen": -0.20735466480255127, | |
| "rewards/margins": -0.0017751154955476522, | |
| "rewards/rejected": -0.20557956397533417, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 16.469711313950462, | |
| "learning_rate": 4.277872161641681e-07, | |
| "logits/chosen": -1.3499563932418823, | |
| "logits/rejected": -1.2568999528884888, | |
| "logps/chosen": -196.1409149169922, | |
| "logps/pi_response": -306.9660949707031, | |
| "logps/ref_response": -266.49676513671875, | |
| "logps/rejected": -198.00282287597656, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.5269230604171753, | |
| "rewards/chosen": -0.19123147428035736, | |
| "rewards/margins": 0.02352394536137581, | |
| "rewards/rejected": -0.21475543081760406, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 14.511218616556945, | |
| "learning_rate": 3.8318133624280046e-07, | |
| "logits/chosen": -1.4744631052017212, | |
| "logits/rejected": -1.430633306503296, | |
| "logps/chosen": -225.62750244140625, | |
| "logps/pi_response": -312.2039489746094, | |
| "logps/ref_response": -265.6466064453125, | |
| "logps/rejected": -226.8577880859375, | |
| "loss": 0.6857, | |
| "rewards/accuracies": 0.5307692289352417, | |
| "rewards/chosen": -0.27648842334747314, | |
| "rewards/margins": 0.02260260097682476, | |
| "rewards/rejected": -0.29909104108810425, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 16.09750795949007, | |
| "learning_rate": 3.316028034595861e-07, | |
| "logits/chosen": -1.326416015625, | |
| "logits/rejected": -1.3301128149032593, | |
| "logps/chosen": -239.2198028564453, | |
| "logps/pi_response": -351.46148681640625, | |
| "logps/ref_response": -276.6808166503906, | |
| "logps/rejected": -242.2137908935547, | |
| "loss": 0.6913, | |
| "rewards/accuracies": 0.48076921701431274, | |
| "rewards/chosen": -0.5429643988609314, | |
| "rewards/margins": -0.012777900323271751, | |
| "rewards/rejected": -0.5301865339279175, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 17.876259342764257, | |
| "learning_rate": 2.7575199021178855e-07, | |
| "logits/chosen": -1.2329574823379517, | |
| "logits/rejected": -1.1573874950408936, | |
| "logps/chosen": -236.77549743652344, | |
| "logps/pi_response": -343.8861389160156, | |
| "logps/ref_response": -271.09393310546875, | |
| "logps/rejected": -233.1557159423828, | |
| "loss": 0.6869, | |
| "rewards/accuracies": 0.5115384459495544, | |
| "rewards/chosen": -0.48039481043815613, | |
| "rewards/margins": 0.0135263130068779, | |
| "rewards/rejected": -0.4939211308956146, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 16.681056784028165, | |
| "learning_rate": 2.1855294234408068e-07, | |
| "logits/chosen": -1.3373737335205078, | |
| "logits/rejected": -1.2385435104370117, | |
| "logps/chosen": -223.1448211669922, | |
| "logps/pi_response": -334.9044494628906, | |
| "logps/ref_response": -266.6827697753906, | |
| "logps/rejected": -225.2432403564453, | |
| "loss": 0.6904, | |
| "rewards/accuracies": 0.5769230723381042, | |
| "rewards/chosen": -0.39324137568473816, | |
| "rewards/margins": 0.031876321882009506, | |
| "rewards/rejected": -0.42511770129203796, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 69.55993387837812, | |
| "learning_rate": 1.6300029195778453e-07, | |
| "logits/chosen": -1.1172312498092651, | |
| "logits/rejected": -1.0658454895019531, | |
| "logps/chosen": -242.79571533203125, | |
| "logps/pi_response": -359.3255310058594, | |
| "logps/ref_response": -268.1200256347656, | |
| "logps/rejected": -244.661865234375, | |
| "loss": 0.6854, | |
| "rewards/accuracies": 0.5769230723381042, | |
| "rewards/chosen": -0.6469340920448303, | |
| "rewards/margins": 0.03381125256419182, | |
| "rewards/rejected": -0.6807453632354736, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 17.847360893927068, | |
| "learning_rate": 1.1200247470632392e-07, | |
| "logits/chosen": -1.0057731866836548, | |
| "logits/rejected": -1.0688775777816772, | |
| "logps/chosen": -267.8451232910156, | |
| "logps/pi_response": -402.1688537597656, | |
| "logps/ref_response": -288.839599609375, | |
| "logps/rejected": -276.8705139160156, | |
| "loss": 0.68, | |
| "rewards/accuracies": 0.5346153974533081, | |
| "rewards/chosen": -0.8569359183311462, | |
| "rewards/margins": 0.06452393531799316, | |
| "rewards/rejected": -0.9214598536491394, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 15.803343162533332, | |
| "learning_rate": 6.822945986946385e-08, | |
| "logits/chosen": -0.9977442622184753, | |
| "logits/rejected": -1.0390866994857788, | |
| "logps/chosen": -259.6402587890625, | |
| "logps/pi_response": -377.9400939941406, | |
| "logps/ref_response": -266.2973937988281, | |
| "logps/rejected": -268.2873840332031, | |
| "loss": 0.6756, | |
| "rewards/accuracies": 0.5269230604171753, | |
| "rewards/chosen": -0.7336382269859314, | |
| "rewards/margins": 0.044992994517087936, | |
| "rewards/rejected": -0.7786312103271484, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 22.692522756797395, | |
| "learning_rate": 3.397296523427806e-08, | |
| "logits/chosen": -0.895865797996521, | |
| "logits/rejected": -0.7990739345550537, | |
| "logps/chosen": -237.3929443359375, | |
| "logps/pi_response": -364.76470947265625, | |
| "logps/ref_response": -257.66021728515625, | |
| "logps/rejected": -240.6884307861328, | |
| "loss": 0.6773, | |
| "rewards/accuracies": 0.5653846263885498, | |
| "rewards/chosen": -0.688443124294281, | |
| "rewards/margins": 0.04434271529316902, | |
| "rewards/rejected": -0.7327858209609985, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 19.97244033218429, | |
| "learning_rate": 1.1026475173977978e-08, | |
| "logits/chosen": -1.052285075187683, | |
| "logits/rejected": -1.1165075302124023, | |
| "logps/chosen": -247.59332275390625, | |
| "logps/pi_response": -372.91534423828125, | |
| "logps/ref_response": -265.1993408203125, | |
| "logps/rejected": -252.52793884277344, | |
| "loss": 0.6715, | |
| "rewards/accuracies": 0.5923076868057251, | |
| "rewards/chosen": -0.640332043170929, | |
| "rewards/margins": 0.03581571578979492, | |
| "rewards/rejected": -0.6761477589607239, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "eta": 0.0010000000474974513, | |
| "grad_norm": 17.943905253705193, | |
| "learning_rate": 5.913435276374834e-10, | |
| "logits/chosen": -1.0871530771255493, | |
| "logits/rejected": -1.1508903503417969, | |
| "logps/chosen": -246.49159240722656, | |
| "logps/pi_response": -384.3344421386719, | |
| "logps/ref_response": -275.81292724609375, | |
| "logps/rejected": -250.1259307861328, | |
| "loss": 0.6818, | |
| "rewards/accuracies": 0.5692307949066162, | |
| "rewards/chosen": -0.6599874496459961, | |
| "rewards/margins": 0.061169885098934174, | |
| "rewards/rejected": -0.721157431602478, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "step": 153, | |
| "total_flos": 0.0, | |
| "train_loss": 0.685623891213361, | |
| "train_runtime": 23340.5225, | |
| "train_samples_per_second": 0.857, | |
| "train_steps_per_second": 0.007 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 153, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 200, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |