| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9992429977289932, | |
| "eval_steps": 500, | |
| "global_step": 165, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 43.6239071005382, | |
| "learning_rate": 2.941176470588235e-09, | |
| "logits/chosen": -1.3522639274597168, | |
| "logits/rejected": -1.3693311214447021, | |
| "logps/chosen": -262.57476806640625, | |
| "logps/rejected": -283.94244384765625, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "grad_norm": 40.670158110610615, | |
| "learning_rate": 2.941176470588235e-08, | |
| "logits/chosen": -1.1757179498672485, | |
| "logits/rejected": -1.2358938455581665, | |
| "logps/chosen": -280.3355407714844, | |
| "logps/rejected": -300.9811706542969, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.4236111044883728, | |
| "rewards/chosen": 0.0006423706654459238, | |
| "rewards/margins": 0.00042565667536109686, | |
| "rewards/rejected": 0.00021671393187716603, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "grad_norm": 40.34301193149703, | |
| "learning_rate": 4.994932636402031e-08, | |
| "logits/chosen": -1.1265027523040771, | |
| "logits/rejected": -1.3426095247268677, | |
| "logps/chosen": -277.8979187011719, | |
| "logps/rejected": -299.1261291503906, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.5406249761581421, | |
| "rewards/chosen": 0.0021728514693677425, | |
| "rewards/margins": 0.000990995205938816, | |
| "rewards/rejected": 0.0011818561470136046, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 38.47792188182457, | |
| "learning_rate": 4.905416503522123e-08, | |
| "logits/chosen": -1.0218889713287354, | |
| "logits/rejected": -1.151049256324768, | |
| "logps/chosen": -273.4291687011719, | |
| "logps/rejected": -301.57781982421875, | |
| "loss": 0.6924, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.016858745366334915, | |
| "rewards/margins": 0.0027274340391159058, | |
| "rewards/rejected": 0.014131310395896435, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 56.52700447561209, | |
| "learning_rate": 4.707922373336523e-08, | |
| "logits/chosen": -1.084263801574707, | |
| "logits/rejected": -1.2900816202163696, | |
| "logps/chosen": -292.4299011230469, | |
| "logps/rejected": -308.45062255859375, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.030295047909021378, | |
| "rewards/margins": 0.000461754942080006, | |
| "rewards/rejected": 0.029833292588591576, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 38.00149695178064, | |
| "learning_rate": 4.4113156629677314e-08, | |
| "logits/chosen": -1.167959451675415, | |
| "logits/rejected": -1.299862265586853, | |
| "logps/chosen": -296.1455383300781, | |
| "logps/rejected": -305.6954040527344, | |
| "loss": 0.6903, | |
| "rewards/accuracies": 0.4906249940395355, | |
| "rewards/chosen": 0.04823774844408035, | |
| "rewards/margins": 0.00479243416339159, | |
| "rewards/rejected": 0.043445318937301636, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 37.72422002493558, | |
| "learning_rate": 4.028910905897228e-08, | |
| "logits/chosen": -1.181056261062622, | |
| "logits/rejected": -1.0861554145812988, | |
| "logps/chosen": -292.48040771484375, | |
| "logps/rejected": -304.0435485839844, | |
| "loss": 0.6912, | |
| "rewards/accuracies": 0.528124988079071, | |
| "rewards/chosen": 0.06034231185913086, | |
| "rewards/margins": 0.006084255874156952, | |
| "rewards/rejected": 0.054258059710264206, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 42.79122392676645, | |
| "learning_rate": 3.577874068920445e-08, | |
| "logits/chosen": -1.210323691368103, | |
| "logits/rejected": -1.065538763999939, | |
| "logps/chosen": -286.93572998046875, | |
| "logps/rejected": -306.0190124511719, | |
| "loss": 0.6918, | |
| "rewards/accuracies": 0.5093749761581421, | |
| "rewards/chosen": 0.06624683737754822, | |
| "rewards/margins": 0.004461642820388079, | |
| "rewards/rejected": 0.06178520247340202, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 39.75881017842744, | |
| "learning_rate": 3.078451980100854e-08, | |
| "logits/chosen": -1.1516613960266113, | |
| "logits/rejected": -1.3043029308319092, | |
| "logps/chosen": -270.6875, | |
| "logps/rejected": -290.72998046875, | |
| "loss": 0.6911, | |
| "rewards/accuracies": 0.515625, | |
| "rewards/chosen": 0.04866773635149002, | |
| "rewards/margins": 0.0037067097146064043, | |
| "rewards/rejected": 0.04496103152632713, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 43.01085877492651, | |
| "learning_rate": 2.5530634583340587e-08, | |
| "logits/chosen": -1.2572039365768433, | |
| "logits/rejected": -1.0870755910873413, | |
| "logps/chosen": -273.9654235839844, | |
| "logps/rejected": -290.50836181640625, | |
| "loss": 0.6898, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": 0.03104880452156067, | |
| "rewards/margins": 0.00489948783069849, | |
| "rewards/rejected": 0.026149321347475052, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 47.166985943498034, | |
| "learning_rate": 2.0252929432814285e-08, | |
| "logits/chosen": -1.1381770372390747, | |
| "logits/rejected": -1.3748772144317627, | |
| "logps/chosen": -282.6134338378906, | |
| "logps/rejected": -304.66790771484375, | |
| "loss": 0.6897, | |
| "rewards/accuracies": 0.5843750238418579, | |
| "rewards/chosen": 0.03918559476733208, | |
| "rewards/margins": 0.01519505213946104, | |
| "rewards/rejected": 0.023990539833903313, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 44.673514372021515, | |
| "learning_rate": 1.5188318011445905e-08, | |
| "logits/chosen": -1.065263271331787, | |
| "logits/rejected": -1.2649091482162476, | |
| "logps/chosen": -277.50775146484375, | |
| "logps/rejected": -300.609619140625, | |
| "loss": 0.6902, | |
| "rewards/accuracies": 0.5531250238418579, | |
| "rewards/chosen": 0.028311368077993393, | |
| "rewards/margins": 0.008604733273386955, | |
| "rewards/rejected": 0.01970663294196129, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 46.58773437073759, | |
| "learning_rate": 1.0564148305586295e-08, | |
| "logits/chosen": -1.1271841526031494, | |
| "logits/rejected": -1.1778924465179443, | |
| "logps/chosen": -279.55084228515625, | |
| "logps/rejected": -298.75030517578125, | |
| "loss": 0.6902, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 0.028242077678442, | |
| "rewards/margins": 0.003035143483430147, | |
| "rewards/rejected": 0.02520693466067314, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 41.9191117851892, | |
| "learning_rate": 6.587997083462196e-09, | |
| "logits/chosen": -1.0855623483657837, | |
| "logits/rejected": -1.1804945468902588, | |
| "logps/chosen": -283.80682373046875, | |
| "logps/rejected": -294.71844482421875, | |
| "loss": 0.6889, | |
| "rewards/accuracies": 0.590624988079071, | |
| "rewards/chosen": 0.033344708383083344, | |
| "rewards/margins": 0.013705698773264885, | |
| "rewards/rejected": 0.01963900588452816, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 44.355218289856595, | |
| "learning_rate": 3.438351873250492e-09, | |
| "logits/chosen": -1.092165470123291, | |
| "logits/rejected": -1.280500054359436, | |
| "logps/chosen": -278.0908508300781, | |
| "logps/rejected": -305.513427734375, | |
| "loss": 0.6901, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.033924926072359085, | |
| "rewards/margins": 0.007848087698221207, | |
| "rewards/rejected": 0.026076842099428177, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 44.46354128863562, | |
| "learning_rate": 1.256598743236703e-09, | |
| "logits/chosen": -1.0778554677963257, | |
| "logits/rejected": -1.2542009353637695, | |
| "logps/chosen": -265.0628967285156, | |
| "logps/rejected": -297.0721130371094, | |
| "loss": 0.6898, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 0.022747624665498734, | |
| "rewards/margins": 0.007828270085155964, | |
| "rewards/rejected": 0.014919353649020195, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 46.98138173894952, | |
| "learning_rate": 1.4067554877743859e-10, | |
| "logits/chosen": -1.170921802520752, | |
| "logits/rejected": -1.1549434661865234, | |
| "logps/chosen": -280.11676025390625, | |
| "logps/rejected": -299.28729248046875, | |
| "loss": 0.6895, | |
| "rewards/accuracies": 0.528124988079071, | |
| "rewards/chosen": 0.031538333743810654, | |
| "rewards/margins": 0.006610988173633814, | |
| "rewards/rejected": 0.024927344173192978, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 165, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6908076347726764, | |
| "train_runtime": 32496.9517, | |
| "train_samples_per_second": 0.65, | |
| "train_steps_per_second": 0.005 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 165, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 200, | |
| "total_flos": 0.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |