| { | |
| "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": 17.363519218688417, | |
| "learning_rate": 3.125e-08, | |
| "logits/chosen": -2.205641031265259, | |
| "logits/rejected": -2.2929024696350098, | |
| "logps/chosen": -215.50050354003906, | |
| "logps/rejected": -237.99966430664062, | |
| "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": 18.34878510832685, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.382091522216797, | |
| "logits/rejected": -2.295259952545166, | |
| "logps/chosen": -222.111328125, | |
| "logps/rejected": -210.6314697265625, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4615384638309479, | |
| "rewards/chosen": 0.0012403662549331784, | |
| "rewards/margins": 0.0005746442475356162, | |
| "rewards/rejected": 0.0006657222402282059, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 18.337158741008302, | |
| "learning_rate": 4.989490450759331e-07, | |
| "logits/chosen": -2.383176803588867, | |
| "logits/rejected": -2.422689199447632, | |
| "logps/chosen": -212.94821166992188, | |
| "logps/rejected": -232.4334259033203, | |
| "loss": 0.6895, | |
| "rewards/accuracies": 0.5384615659713745, | |
| "rewards/chosen": -0.08582816272974014, | |
| "rewards/margins": 0.0013778842985630035, | |
| "rewards/rejected": -0.08720605075359344, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 22.542054212888594, | |
| "learning_rate": 4.872270441827174e-07, | |
| "logits/chosen": -2.3914377689361572, | |
| "logits/rejected": -2.3520281314849854, | |
| "logps/chosen": -215.9875030517578, | |
| "logps/rejected": -227.70399475097656, | |
| "loss": 0.6861, | |
| "rewards/accuracies": 0.5384615659713745, | |
| "rewards/chosen": -0.25440388917922974, | |
| "rewards/margins": 0.03694874048233032, | |
| "rewards/rejected": -0.29135259985923767, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "grad_norm": 18.6076942310058, | |
| "learning_rate": 4.6308512113530063e-07, | |
| "logits/chosen": -2.4768216609954834, | |
| "logits/rejected": -2.4233009815216064, | |
| "logps/chosen": -229.41555786132812, | |
| "logps/rejected": -242.26214599609375, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.5384615659713745, | |
| "rewards/chosen": -0.19061818718910217, | |
| "rewards/margins": 0.01797662116587162, | |
| "rewards/rejected": -0.20859479904174805, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 15.540618809394221, | |
| "learning_rate": 4.277872161641681e-07, | |
| "logits/chosen": -2.5903377532958984, | |
| "logits/rejected": -2.5592682361602783, | |
| "logps/chosen": -225.72836303710938, | |
| "logps/rejected": -240.36595153808594, | |
| "loss": 0.6897, | |
| "rewards/accuracies": 0.5307692289352417, | |
| "rewards/chosen": -0.033847782760858536, | |
| "rewards/margins": 0.001607205718755722, | |
| "rewards/rejected": -0.03545498102903366, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 24.02478451571149, | |
| "learning_rate": 3.8318133624280046e-07, | |
| "logits/chosen": -2.6142632961273193, | |
| "logits/rejected": -2.6310534477233887, | |
| "logps/chosen": -223.6808624267578, | |
| "logps/rejected": -255.2255401611328, | |
| "loss": 0.6903, | |
| "rewards/accuracies": 0.5423076748847961, | |
| "rewards/chosen": -0.10960451513528824, | |
| "rewards/margins": 0.02808019518852234, | |
| "rewards/rejected": -0.13768470287322998, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 20.288409772533704, | |
| "learning_rate": 3.316028034595861e-07, | |
| "logits/chosen": -2.5127460956573486, | |
| "logits/rejected": -2.5037307739257812, | |
| "logps/chosen": -230.75833129882812, | |
| "logps/rejected": -256.0094909667969, | |
| "loss": 0.6832, | |
| "rewards/accuracies": 0.4961538314819336, | |
| "rewards/chosen": -0.2965443730354309, | |
| "rewards/margins": 0.03154058754444122, | |
| "rewards/rejected": -0.3280849754810333, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 22.680428500041128, | |
| "learning_rate": 2.7575199021178855e-07, | |
| "logits/chosen": -2.5435211658477783, | |
| "logits/rejected": -2.5064070224761963, | |
| "logps/chosen": -251.8400421142578, | |
| "logps/rejected": -273.2138671875, | |
| "loss": 0.6778, | |
| "rewards/accuracies": 0.5538461804389954, | |
| "rewards/chosen": -0.28813430666923523, | |
| "rewards/margins": 0.06685086339712143, | |
| "rewards/rejected": -0.3549851179122925, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 18.86897478499211, | |
| "learning_rate": 2.1855294234408068e-07, | |
| "logits/chosen": -2.5077600479125977, | |
| "logits/rejected": -2.503957986831665, | |
| "logps/chosen": -251.26556396484375, | |
| "logps/rejected": -242.37310791015625, | |
| "loss": 0.6842, | |
| "rewards/accuracies": 0.5307692289352417, | |
| "rewards/chosen": -0.24721869826316833, | |
| "rewards/margins": 0.009251880459487438, | |
| "rewards/rejected": -0.25647059082984924, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 20.56846597020704, | |
| "learning_rate": 1.6300029195778453e-07, | |
| "logits/chosen": -2.492692708969116, | |
| "logits/rejected": -2.3162038326263428, | |
| "logps/chosen": -253.9385223388672, | |
| "logps/rejected": -268.18414306640625, | |
| "loss": 0.6821, | |
| "rewards/accuracies": 0.5538461804389954, | |
| "rewards/chosen": -0.4339679181575775, | |
| "rewards/margins": 0.06273461133241653, | |
| "rewards/rejected": -0.49670252203941345, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 29.852940409061166, | |
| "learning_rate": 1.1200247470632392e-07, | |
| "logits/chosen": -2.409569501876831, | |
| "logits/rejected": -2.3867456912994385, | |
| "logps/chosen": -297.3172607421875, | |
| "logps/rejected": -285.6565246582031, | |
| "loss": 0.6823, | |
| "rewards/accuracies": 0.4923076927661896, | |
| "rewards/chosen": -0.73952317237854, | |
| "rewards/margins": 0.020068956539034843, | |
| "rewards/rejected": -0.7595921754837036, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "grad_norm": 20.239291104683563, | |
| "learning_rate": 6.822945986946385e-08, | |
| "logits/chosen": -2.3316874504089355, | |
| "logits/rejected": -2.2675819396972656, | |
| "logps/chosen": -275.67767333984375, | |
| "logps/rejected": -291.9703674316406, | |
| "loss": 0.6858, | |
| "rewards/accuracies": 0.5384615659713745, | |
| "rewards/chosen": -0.5500468611717224, | |
| "rewards/margins": 0.06261468678712845, | |
| "rewards/rejected": -0.6126615405082703, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 24.27835409910822, | |
| "learning_rate": 3.397296523427806e-08, | |
| "logits/chosen": -2.4436075687408447, | |
| "logits/rejected": -2.309699296951294, | |
| "logps/chosen": -233.41680908203125, | |
| "logps/rejected": -262.0289611816406, | |
| "loss": 0.6782, | |
| "rewards/accuracies": 0.5807692408561707, | |
| "rewards/chosen": -0.36947229504585266, | |
| "rewards/margins": 0.0935312956571579, | |
| "rewards/rejected": -0.46300360560417175, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 24.790097756991855, | |
| "learning_rate": 1.1026475173977978e-08, | |
| "logits/chosen": -2.3928282260894775, | |
| "logits/rejected": -2.382559299468994, | |
| "logps/chosen": -244.34410095214844, | |
| "logps/rejected": -248.87876892089844, | |
| "loss": 0.6769, | |
| "rewards/accuracies": 0.5615384578704834, | |
| "rewards/chosen": -0.27959996461868286, | |
| "rewards/margins": 0.04778864234685898, | |
| "rewards/rejected": -0.32738858461380005, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 24.029709323211232, | |
| "learning_rate": 5.913435276374834e-10, | |
| "logits/chosen": -2.4940154552459717, | |
| "logits/rejected": -2.3921005725860596, | |
| "logps/chosen": -241.1715545654297, | |
| "logps/rejected": -262.944580078125, | |
| "loss": 0.6737, | |
| "rewards/accuracies": 0.6192307472229004, | |
| "rewards/chosen": -0.2436859905719757, | |
| "rewards/margins": 0.09733694791793823, | |
| "rewards/rejected": -0.34102290868759155, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "step": 153, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6839175197034101, | |
| "train_runtime": 39806.5992, | |
| "train_samples_per_second": 0.502, | |
| "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 | |
| } | |