| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984, | |
| "eval_steps": 100, | |
| "global_step": 312, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.5625e-07, | |
| "logits/chosen": -2.7731900215148926, | |
| "logits/rejected": -2.6362287998199463, | |
| "logps/chosen": -356.1260070800781, | |
| "logps/rejected": -311.3892822265625, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.5625e-06, | |
| "logits/chosen": -2.3915464878082275, | |
| "logits/rejected": -2.3424172401428223, | |
| "logps/chosen": -243.08827209472656, | |
| "logps/rejected": -240.88124084472656, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.004824994597584009, | |
| "rewards/margins": 0.001562346238642931, | |
| "rewards/rejected": 0.003262649057433009, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.125e-06, | |
| "logits/chosen": -2.6855998039245605, | |
| "logits/rejected": -2.503112316131592, | |
| "logps/chosen": -276.1568908691406, | |
| "logps/rejected": -245.57150268554688, | |
| "loss": 0.687, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.04376252368092537, | |
| "rewards/margins": 0.011996113695204258, | |
| "rewards/rejected": 0.03176640719175339, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.6875000000000004e-06, | |
| "logits/chosen": -2.5015687942504883, | |
| "logits/rejected": -2.448686122894287, | |
| "logps/chosen": -244.99642944335938, | |
| "logps/rejected": -249.10916137695312, | |
| "loss": 0.6832, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.05081823468208313, | |
| "rewards/margins": 0.018069546669721603, | |
| "rewards/rejected": 0.032748688012361526, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.989935734988098e-06, | |
| "logits/chosen": -2.5158021450042725, | |
| "logits/rejected": -2.337573289871216, | |
| "logps/chosen": -272.7821350097656, | |
| "logps/rejected": -227.36007690429688, | |
| "loss": 0.6668, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.0395994558930397, | |
| "rewards/margins": 0.04740050435066223, | |
| "rewards/rejected": -0.00780104985460639, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.949188496058089e-06, | |
| "logits/chosen": -2.430145502090454, | |
| "logits/rejected": -2.4263150691986084, | |
| "logps/chosen": -249.273681640625, | |
| "logps/rejected": -266.5956726074219, | |
| "loss": 0.6528, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.02035255916416645, | |
| "rewards/margins": 0.059511054307222366, | |
| "rewards/rejected": -0.03915848955512047, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.8776412907378845e-06, | |
| "logits/chosen": -2.4893181324005127, | |
| "logits/rejected": -2.418604612350464, | |
| "logps/chosen": -301.8047790527344, | |
| "logps/rejected": -252.42892456054688, | |
| "loss": 0.6591, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.06781601160764694, | |
| "rewards/margins": 0.07723621279001236, | |
| "rewards/rejected": -0.1450522094964981, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.7761938666470405e-06, | |
| "logits/chosen": -2.4578957557678223, | |
| "logits/rejected": -2.4078097343444824, | |
| "logps/chosen": -259.1146545410156, | |
| "logps/rejected": -255.2762908935547, | |
| "loss": 0.6412, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.09445185959339142, | |
| "rewards/margins": 0.15250881016254425, | |
| "rewards/rejected": -0.24696068465709686, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.646121984004666e-06, | |
| "logits/chosen": -2.5219717025756836, | |
| "logits/rejected": -2.3697924613952637, | |
| "logps/chosen": -289.8721618652344, | |
| "logps/rejected": -306.9769287109375, | |
| "loss": 0.6357, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.21299012005329132, | |
| "rewards/margins": 0.16702046990394592, | |
| "rewards/rejected": -0.38001060485839844, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.4890613722044526e-06, | |
| "logits/chosen": -2.427962064743042, | |
| "logits/rejected": -2.326305866241455, | |
| "logps/chosen": -265.3756408691406, | |
| "logps/rejected": -262.7252197265625, | |
| "loss": 0.6272, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.16604574024677277, | |
| "rewards/margins": 0.19639183580875397, | |
| "rewards/rejected": -0.36243754625320435, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.3069871595684795e-06, | |
| "logits/chosen": -2.213723659515381, | |
| "logits/rejected": -2.217102527618408, | |
| "logps/chosen": -245.6179962158203, | |
| "logps/rejected": -293.12518310546875, | |
| "loss": 0.6269, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.38440248370170593, | |
| "rewards/margins": 0.126637801527977, | |
| "rewards/rejected": -0.5110402703285217, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "eval_logits/chosen": -2.3665878772735596, | |
| "eval_logits/rejected": -2.256598949432373, | |
| "eval_logps/chosen": -287.2168884277344, | |
| "eval_logps/rejected": -285.49847412109375, | |
| "eval_loss": 0.6268974542617798, | |
| "eval_rewards/accuracies": 0.6819999814033508, | |
| "eval_rewards/chosen": -0.23765824735164642, | |
| "eval_rewards/margins": 0.2054254114627838, | |
| "eval_rewards/rejected": -0.44308364391326904, | |
| "eval_runtime": 543.2957, | |
| "eval_samples_per_second": 3.681, | |
| "eval_steps_per_second": 0.46, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.102189034962561e-06, | |
| "logits/chosen": -2.338050603866577, | |
| "logits/rejected": -2.2199347019195557, | |
| "logps/chosen": -304.7019958496094, | |
| "logps/rejected": -283.5575256347656, | |
| "loss": 0.6232, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.18706437945365906, | |
| "rewards/margins": 0.24619019031524658, | |
| "rewards/rejected": -0.43325456976890564, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.8772424536302565e-06, | |
| "logits/chosen": -2.199939250946045, | |
| "logits/rejected": -2.1462173461914062, | |
| "logps/chosen": -280.5738525390625, | |
| "logps/rejected": -272.75537109375, | |
| "loss": 0.6256, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.23016035556793213, | |
| "rewards/margins": 0.2529276907444, | |
| "rewards/rejected": -0.48308807611465454, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.634976249348867e-06, | |
| "logits/chosen": -2.4285922050476074, | |
| "logits/rejected": -2.252119541168213, | |
| "logps/chosen": -337.8984375, | |
| "logps/rejected": -329.4248962402344, | |
| "loss": 0.6299, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.27586501836776733, | |
| "rewards/margins": 0.285078763961792, | |
| "rewards/rejected": -0.5609437823295593, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.3784370602033572e-06, | |
| "logits/chosen": -2.072373628616333, | |
| "logits/rejected": -1.9053455591201782, | |
| "logps/chosen": -251.76571655273438, | |
| "logps/rejected": -285.0694885253906, | |
| "loss": 0.6067, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.5010747909545898, | |
| "rewards/margins": 0.26867786049842834, | |
| "rewards/rejected": -0.7697526216506958, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.1108510153447352e-06, | |
| "logits/chosen": -2.21221661567688, | |
| "logits/rejected": -2.136280059814453, | |
| "logps/chosen": -338.2016296386719, | |
| "logps/rejected": -331.0526428222656, | |
| "loss": 0.608, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.639680027961731, | |
| "rewards/margins": 0.2550516426563263, | |
| "rewards/rejected": -0.8947317004203796, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 2.835583164544139e-06, | |
| "logits/chosen": -2.2209646701812744, | |
| "logits/rejected": -2.022948980331421, | |
| "logps/chosen": -377.3534851074219, | |
| "logps/rejected": -344.77252197265625, | |
| "loss": 0.5937, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.7194479703903198, | |
| "rewards/margins": 0.39620086550712585, | |
| "rewards/rejected": -1.115648865699768, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.556095160739513e-06, | |
| "logits/chosen": -2.1350314617156982, | |
| "logits/rejected": -1.85476553440094, | |
| "logps/chosen": -351.29638671875, | |
| "logps/rejected": -354.8650817871094, | |
| "loss": 0.6069, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.7903974652290344, | |
| "rewards/margins": 0.24958536028862, | |
| "rewards/rejected": -1.039982795715332, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.2759017277414165e-06, | |
| "logits/chosen": -2.0943400859832764, | |
| "logits/rejected": -1.8893616199493408, | |
| "logps/chosen": -322.147216796875, | |
| "logps/rejected": -327.81304931640625, | |
| "loss": 0.6252, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.6768954992294312, | |
| "rewards/margins": 0.20395174622535706, | |
| "rewards/rejected": -0.8808472752571106, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 1.9985264605418185e-06, | |
| "logits/chosen": -1.9419981241226196, | |
| "logits/rejected": -1.7324016094207764, | |
| "logps/chosen": -328.23760986328125, | |
| "logps/rejected": -314.13922119140625, | |
| "loss": 0.584, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.4452829360961914, | |
| "rewards/margins": 0.4189114570617676, | |
| "rewards/rejected": -0.864194393157959, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.7274575140626318e-06, | |
| "logits/chosen": -2.144902229309082, | |
| "logits/rejected": -1.7156444787979126, | |
| "logps/chosen": -362.327880859375, | |
| "logps/rejected": -322.9747619628906, | |
| "loss": 0.6332, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.4486660957336426, | |
| "rewards/margins": 0.3436250388622284, | |
| "rewards/rejected": -0.7922911047935486, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "eval_logits/chosen": -1.893760323524475, | |
| "eval_logits/rejected": -1.6871448755264282, | |
| "eval_logps/chosen": -322.544189453125, | |
| "eval_logps/rejected": -337.0687255859375, | |
| "eval_loss": 0.5820500254631042, | |
| "eval_rewards/accuracies": 0.7059999704360962, | |
| "eval_rewards/chosen": -0.5909315943717957, | |
| "eval_rewards/margins": 0.3678547739982605, | |
| "eval_rewards/rejected": -0.9587863683700562, | |
| "eval_runtime": 543.1459, | |
| "eval_samples_per_second": 3.682, | |
| "eval_steps_per_second": 0.46, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.466103737583699e-06, | |
| "logits/chosen": -1.8559290170669556, | |
| "logits/rejected": -1.7014697790145874, | |
| "logps/chosen": -324.19256591796875, | |
| "logps/rejected": -352.70697021484375, | |
| "loss": 0.552, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.6294658780097961, | |
| "rewards/margins": 0.45733365416526794, | |
| "rewards/rejected": -1.0867995023727417, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.217751806485235e-06, | |
| "logits/chosen": -1.8568174839019775, | |
| "logits/rejected": -1.6362855434417725, | |
| "logps/chosen": -356.0939636230469, | |
| "logps/rejected": -389.1434326171875, | |
| "loss": 0.5765, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.6206706762313843, | |
| "rewards/margins": 0.5877247452735901, | |
| "rewards/rejected": -1.2083956003189087, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 9.855248903979505e-07, | |
| "logits/chosen": -1.9677798748016357, | |
| "logits/rejected": -1.8001766204833984, | |
| "logps/chosen": -333.63409423828125, | |
| "logps/rejected": -372.82232666015625, | |
| "loss": 0.5961, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.6273213028907776, | |
| "rewards/margins": 0.4957484304904938, | |
| "rewards/rejected": -1.1230696439743042, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.723433775328385e-07, | |
| "logits/chosen": -1.6269299983978271, | |
| "logits/rejected": -1.5314247608184814, | |
| "logps/chosen": -343.7135314941406, | |
| "logps/rejected": -360.920166015625, | |
| "loss": 0.5733, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.6133186221122742, | |
| "rewards/margins": 0.3855026662349701, | |
| "rewards/rejected": -0.9988213777542114, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 5.808881491049723e-07, | |
| "logits/chosen": -1.8092960119247437, | |
| "logits/rejected": -1.4363586902618408, | |
| "logps/chosen": -302.46234130859375, | |
| "logps/rejected": -305.09393310546875, | |
| "loss": 0.5822, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.7107473611831665, | |
| "rewards/margins": 0.3086285889148712, | |
| "rewards/rejected": -1.0193760395050049, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.1356686569674344e-07, | |
| "logits/chosen": -2.0522053241729736, | |
| "logits/rejected": -1.6467043161392212, | |
| "logps/chosen": -370.14019775390625, | |
| "logps/rejected": -350.79364013671875, | |
| "loss": 0.5594, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.6428462266921997, | |
| "rewards/margins": 0.41562938690185547, | |
| "rewards/rejected": -1.0584756135940552, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 2.7248368952908055e-07, | |
| "logits/chosen": -1.736702561378479, | |
| "logits/rejected": -1.5174537897109985, | |
| "logps/chosen": -293.9969177246094, | |
| "logps/rejected": -317.34844970703125, | |
| "loss": 0.5803, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.5701481103897095, | |
| "rewards/margins": 0.500605583190918, | |
| "rewards/rejected": -1.0707536935806274, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.59412823400657e-07, | |
| "logits/chosen": -1.6159837245941162, | |
| "logits/rejected": -1.2289941310882568, | |
| "logps/chosen": -342.9421081542969, | |
| "logps/rejected": -372.6164855957031, | |
| "loss": 0.5578, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.7729519605636597, | |
| "rewards/margins": 0.49912238121032715, | |
| "rewards/rejected": -1.2720743417739868, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 7.577619905828281e-08, | |
| "logits/chosen": -1.502423644065857, | |
| "logits/rejected": -1.456081509590149, | |
| "logps/chosen": -329.8805236816406, | |
| "logps/rejected": -356.34417724609375, | |
| "loss": 0.5827, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.678324818611145, | |
| "rewards/margins": 0.42007485032081604, | |
| "rewards/rejected": -1.0983997583389282, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.262559558016325e-08, | |
| "logits/chosen": -1.6769917011260986, | |
| "logits/rejected": -1.406165361404419, | |
| "logps/chosen": -328.3318786621094, | |
| "logps/rejected": -353.73968505859375, | |
| "loss": 0.5648, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.6493935585021973, | |
| "rewards/margins": 0.43068727850914, | |
| "rewards/rejected": -1.0800807476043701, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "eval_logits/chosen": -1.7015434503555298, | |
| "eval_logits/rejected": -1.4598934650421143, | |
| "eval_logps/chosen": -331.1508483886719, | |
| "eval_logps/rejected": -351.8941955566406, | |
| "eval_loss": 0.5735270977020264, | |
| "eval_rewards/accuracies": 0.6940000057220459, | |
| "eval_rewards/chosen": -0.6769981980323792, | |
| "eval_rewards/margins": 0.4300425946712494, | |
| "eval_rewards/rejected": -1.1070406436920166, | |
| "eval_runtime": 542.9185, | |
| "eval_samples_per_second": 3.684, | |
| "eval_steps_per_second": 0.46, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 6.294126437336734e-10, | |
| "logits/chosen": -1.7649977207183838, | |
| "logits/rejected": -1.521240234375, | |
| "logps/chosen": -326.1722717285156, | |
| "logps/rejected": -356.2889709472656, | |
| "loss": 0.5603, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.6321894526481628, | |
| "rewards/margins": 0.4693359434604645, | |
| "rewards/rejected": -1.1015253067016602, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 312, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6116275028922619, | |
| "train_runtime": 6907.8509, | |
| "train_samples_per_second": 1.448, | |
| "train_steps_per_second": 0.045 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 312, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |