| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 100, | |
| "global_step": 385, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.282051282051282e-08, | |
| "logits/chosen": -1.7278180122375488, | |
| "logits/rejected": -1.7377450466156006, | |
| "logps/chosen": -29.553977966308594, | |
| "logps/rejected": -42.813133239746094, | |
| "loss": 1.0, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.2820512820512818e-07, | |
| "logits/chosen": -1.866065502166748, | |
| "logits/rejected": -1.8703795671463013, | |
| "logps/chosen": -36.988380432128906, | |
| "logps/rejected": -33.66728210449219, | |
| "loss": 0.9945, | |
| "rewards/accuracies": 0.5694444179534912, | |
| "rewards/chosen": 0.0018219809280708432, | |
| "rewards/margins": 0.00554280448704958, | |
| "rewards/rejected": -0.0037208227440714836, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.5641025641025636e-07, | |
| "logits/chosen": -1.997332215309143, | |
| "logits/rejected": -1.999983549118042, | |
| "logps/chosen": -29.625896453857422, | |
| "logps/rejected": -29.035802841186523, | |
| "loss": 1.0011, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": 0.0016297042602673173, | |
| "rewards/margins": -0.0011343134101480246, | |
| "rewards/rejected": 0.00276401755400002, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 3.8461538461538463e-07, | |
| "logits/chosen": -1.9199495315551758, | |
| "logits/rejected": -1.917249321937561, | |
| "logps/chosen": -31.421478271484375, | |
| "logps/rejected": -33.2115364074707, | |
| "loss": 1.0012, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": -0.0005385443801060319, | |
| "rewards/margins": -0.0012309944722801447, | |
| "rewards/rejected": 0.0006924502667970955, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.999896948438433e-07, | |
| "logits/chosen": -2.0169284343719482, | |
| "logits/rejected": -2.008178949356079, | |
| "logps/chosen": -32.59435272216797, | |
| "logps/rejected": -32.49193572998047, | |
| "loss": 1.0038, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.0017543068388476968, | |
| "rewards/margins": -0.0038279418367892504, | |
| "rewards/rejected": 0.0020736351143568754, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.987541037542186e-07, | |
| "logits/chosen": -1.86457097530365, | |
| "logits/rejected": -1.8537908792495728, | |
| "logps/chosen": -33.56566619873047, | |
| "logps/rejected": -35.423240661621094, | |
| "loss": 1.0037, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": -0.0008188748615793884, | |
| "rewards/margins": -0.003729355288669467, | |
| "rewards/rejected": 0.0029104803688824177, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.954691471941118e-07, | |
| "logits/chosen": -1.9449050426483154, | |
| "logits/rejected": -1.9468472003936768, | |
| "logps/chosen": -32.59955596923828, | |
| "logps/rejected": -33.1828498840332, | |
| "loss": 0.9988, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 6.278457294683903e-05, | |
| "rewards/margins": 0.0012094037374481559, | |
| "rewards/rejected": -0.0011466194409877062, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.901618883413548e-07, | |
| "logits/chosen": -2.079878330230713, | |
| "logits/rejected": -2.084862232208252, | |
| "logps/chosen": -33.98878860473633, | |
| "logps/rejected": -36.574462890625, | |
| "loss": 0.9995, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -3.0292372684925795e-05, | |
| "rewards/margins": 0.0005289571708999574, | |
| "rewards/rejected": -0.0005592495435848832, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.828760511501322e-07, | |
| "logits/chosen": -1.9425569772720337, | |
| "logits/rejected": -1.9457191228866577, | |
| "logps/chosen": -34.40068054199219, | |
| "logps/rejected": -34.5762939453125, | |
| "loss": 0.9989, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.002712726127356291, | |
| "rewards/margins": 0.0011012849863618612, | |
| "rewards/rejected": 0.0016114413738250732, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.736716601303429e-07, | |
| "logits/chosen": -1.9507396221160889, | |
| "logits/rejected": -1.9552500247955322, | |
| "logps/chosen": -32.460357666015625, | |
| "logps/rejected": -32.354434967041016, | |
| "loss": 0.9997, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.0007068099221214652, | |
| "rewards/margins": 0.00025928454124368727, | |
| "rewards/rejected": 0.0004475254681892693, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.62624545834521e-07, | |
| "logits/chosen": -2.0492873191833496, | |
| "logits/rejected": -2.0472888946533203, | |
| "logps/chosen": -32.23810958862305, | |
| "logps/rejected": -31.260278701782227, | |
| "loss": 1.0003, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": 0.00027285455144010484, | |
| "rewards/margins": -0.00032049461151473224, | |
| "rewards/rejected": 0.0005933489883318543, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "eval_logits/chosen": -2.243312358856201, | |
| "eval_logits/rejected": -2.238436222076416, | |
| "eval_logps/chosen": -34.023216247558594, | |
| "eval_logps/rejected": -37.49723434448242, | |
| "eval_loss": 1.0008214712142944, | |
| "eval_rewards/accuracies": 0.49833887815475464, | |
| "eval_rewards/chosen": 0.0011338649783283472, | |
| "eval_rewards/margins": -0.0008048001327551901, | |
| "eval_rewards/rejected": 0.0019386651692911983, | |
| "eval_runtime": 146.1271, | |
| "eval_samples_per_second": 2.347, | |
| "eval_steps_per_second": 0.294, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.4982572012636904e-07, | |
| "logits/chosen": -2.005169153213501, | |
| "logits/rejected": -2.0027499198913574, | |
| "logps/chosen": -33.2365837097168, | |
| "logps/rejected": -34.01953125, | |
| "loss": 1.0003, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": 0.0007230007322505116, | |
| "rewards/margins": -0.00026112591149285436, | |
| "rewards/rejected": 0.0009841264691203833, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.353806263777677e-07, | |
| "logits/chosen": -2.0166728496551514, | |
| "logits/rejected": -2.0083022117614746, | |
| "logps/chosen": -32.457847595214844, | |
| "logps/rejected": -32.18357467651367, | |
| "loss": 0.9991, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.0012942379107698798, | |
| "rewards/margins": 0.0009121716138906777, | |
| "rewards/rejected": -0.002206409815698862, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.194082707715275e-07, | |
| "logits/chosen": -2.0462448596954346, | |
| "logits/rejected": -2.038203001022339, | |
| "logps/chosen": -30.475027084350586, | |
| "logps/rejected": -32.046302795410156, | |
| "loss": 1.0006, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 6.882836896693334e-05, | |
| "rewards/margins": -0.0006015264661982656, | |
| "rewards/rejected": 0.0006703549297526479, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.020402418666621e-07, | |
| "logits/chosen": -1.976986289024353, | |
| "logits/rejected": -1.9872560501098633, | |
| "logps/chosen": -31.407278060913086, | |
| "logps/rejected": -32.543296813964844, | |
| "loss": 0.9979, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.0012384230503812432, | |
| "rewards/margins": 0.0020802102517336607, | |
| "rewards/rejected": -0.0008417873759754002, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 3.8341962650351185e-07, | |
| "logits/chosen": -1.8905508518218994, | |
| "logits/rejected": -1.891632080078125, | |
| "logps/chosen": -34.20501708984375, | |
| "logps/rejected": -34.77235412597656, | |
| "loss": 0.9997, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": -0.0012326312717050314, | |
| "rewards/margins": 0.00027702696388587356, | |
| "rewards/rejected": -0.001509658177383244, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.636998309800572e-07, | |
| "logits/chosen": -1.94281005859375, | |
| "logits/rejected": -1.939327597618103, | |
| "logps/chosen": -36.144107818603516, | |
| "logps/rejected": -32.72822570800781, | |
| "loss": 0.9967, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": 0.0027573786210268736, | |
| "rewards/margins": 0.0033046435564756393, | |
| "rewards/rejected": -0.0005472644697874784, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.430433172111807e-07, | |
| "logits/chosen": -2.0420708656311035, | |
| "logits/rejected": -2.0346803665161133, | |
| "logps/chosen": -33.771934509277344, | |
| "logps/rejected": -31.371145248413086, | |
| "loss": 0.9955, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": 0.002593145240098238, | |
| "rewards/margins": 0.004522555507719517, | |
| "rewards/rejected": -0.0019294100347906351, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.216202642830543e-07, | |
| "logits/chosen": -2.0475738048553467, | |
| "logits/rejected": -2.0528526306152344, | |
| "logps/chosen": -32.524593353271484, | |
| "logps/rejected": -32.510643005371094, | |
| "loss": 0.9954, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.0020192908123135567, | |
| "rewards/margins": 0.004569421522319317, | |
| "rewards/rejected": -0.002550130244344473, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 2.9960716642946403e-07, | |
| "logits/chosen": -2.048490524291992, | |
| "logits/rejected": -2.0457024574279785, | |
| "logps/chosen": -31.492746353149414, | |
| "logps/rejected": -31.319293975830078, | |
| "loss": 1.0003, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.0003792368806898594, | |
| "rewards/margins": -0.0002712804707698524, | |
| "rewards/rejected": -0.0001079567227861844, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.771853789806683e-07, | |
| "logits/chosen": -1.9185683727264404, | |
| "logits/rejected": -1.9232347011566162, | |
| "logps/chosen": -31.5926513671875, | |
| "logps/rejected": -32.78697204589844, | |
| "loss": 0.999, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.001363297225907445, | |
| "rewards/margins": 0.0009675474721007049, | |
| "rewards/rejected": 0.0003957498411182314, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "eval_logits/chosen": -2.243807315826416, | |
| "eval_logits/rejected": -2.2389235496520996, | |
| "eval_logps/chosen": -34.018714904785156, | |
| "eval_logps/rejected": -37.517478942871094, | |
| "eval_loss": 0.9983054399490356, | |
| "eval_rewards/accuracies": 0.5328072905540466, | |
| "eval_rewards/chosen": 0.0015839905245229602, | |
| "eval_rewards/margins": 0.0016701342537999153, | |
| "eval_rewards/rejected": -8.614380931248888e-05, | |
| "eval_runtime": 145.691, | |
| "eval_samples_per_second": 2.354, | |
| "eval_steps_per_second": 0.295, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2.5453962426402e-07, | |
| "logits/chosen": -2.0318503379821777, | |
| "logits/rejected": -2.042539596557617, | |
| "logps/chosen": -31.948400497436523, | |
| "logps/rejected": -33.86983871459961, | |
| "loss": 0.997, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.0022904814686626196, | |
| "rewards/margins": 0.0030044156592339277, | |
| "rewards/rejected": -0.0007139344816096127, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.318564697655179e-07, | |
| "logits/chosen": -1.9251388311386108, | |
| "logits/rejected": -1.9399843215942383, | |
| "logps/chosen": -30.099853515625, | |
| "logps/rejected": -31.55409812927246, | |
| "loss": 0.9992, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.0011685896897688508, | |
| "rewards/margins": 0.000760918774176389, | |
| "rewards/rejected": 0.00040767071186564863, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.093227910899832e-07, | |
| "logits/chosen": -1.9835479259490967, | |
| "logits/rejected": -1.9875112771987915, | |
| "logps/chosen": -33.387638092041016, | |
| "logps/rejected": -31.554845809936523, | |
| "loss": 0.9965, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.0029833675362169743, | |
| "rewards/margins": 0.003537458134815097, | |
| "rewards/rejected": -0.0005540908314287663, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.8712423238279356e-07, | |
| "logits/chosen": -1.9826898574829102, | |
| "logits/rejected": -1.9607274532318115, | |
| "logps/chosen": -34.158443450927734, | |
| "logps/rejected": -34.963783264160156, | |
| "loss": 0.9992, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.00046964577632024884, | |
| "rewards/margins": 0.0008290981058962643, | |
| "rewards/rejected": -0.0012987437658011913, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.654436768970182e-07, | |
| "logits/chosen": -2.024381160736084, | |
| "logits/rejected": -2.0210862159729004, | |
| "logps/chosen": -32.9254035949707, | |
| "logps/rejected": -36.251712799072266, | |
| "loss": 0.9974, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -3.328696038806811e-05, | |
| "rewards/margins": 0.0026453982573002577, | |
| "rewards/rejected": -0.0026786853559315205, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.444597403062196e-07, | |
| "logits/chosen": -1.8911311626434326, | |
| "logits/rejected": -1.8886839151382446, | |
| "logps/chosen": -34.194557189941406, | |
| "logps/rejected": -35.51445770263672, | |
| "loss": 0.9996, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.0006158108590170741, | |
| "rewards/margins": 0.0003673938917927444, | |
| "rewards/rejected": -0.0009832048090174794, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.2434529917578887e-07, | |
| "logits/chosen": -1.8759450912475586, | |
| "logits/rejected": -1.8734045028686523, | |
| "logps/chosen": -34.40558624267578, | |
| "logps/rejected": -31.752349853515625, | |
| "loss": 1.0027, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": -0.0007984804688021541, | |
| "rewards/margins": -0.0026967611629515886, | |
| "rewards/rejected": 0.0018982805777341127, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0526606671603521e-07, | |
| "logits/chosen": -1.980015754699707, | |
| "logits/rejected": -1.9693737030029297, | |
| "logps/chosen": -35.33230209350586, | |
| "logps/rejected": -31.845691680908203, | |
| "loss": 0.9971, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.002433922840282321, | |
| "rewards/margins": 0.0029038069769740105, | |
| "rewards/rejected": -0.0004698836710304022, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 8.737922755071453e-08, | |
| "logits/chosen": -2.0756278038024902, | |
| "logits/rejected": -2.060606002807617, | |
| "logps/chosen": -30.907390594482422, | |
| "logps/rejected": -32.64055252075195, | |
| "loss": 0.9995, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.001673347083851695, | |
| "rewards/margins": 0.0005450047319754958, | |
| "rewards/rejected": 0.0011283422354608774, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 7.08321427484816e-08, | |
| "logits/chosen": -1.946616768836975, | |
| "logits/rejected": -1.9440828561782837, | |
| "logps/chosen": -32.894561767578125, | |
| "logps/rejected": -30.812387466430664, | |
| "loss": 0.9972, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.003255209419876337, | |
| "rewards/margins": 0.002812772523611784, | |
| "rewards/rejected": 0.00044243651791475713, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "eval_logits/chosen": -2.2434821128845215, | |
| "eval_logits/rejected": -2.2386035919189453, | |
| "eval_logps/chosen": -34.017669677734375, | |
| "eval_logps/rejected": -37.50018310546875, | |
| "eval_loss": 0.9999711513519287, | |
| "eval_rewards/accuracies": 0.490448534488678, | |
| "eval_rewards/chosen": 0.0016884810756891966, | |
| "eval_rewards/margins": 4.470993735594675e-05, | |
| "eval_rewards/rejected": 0.001643771305680275, | |
| "eval_runtime": 145.8271, | |
| "eval_samples_per_second": 2.352, | |
| "eval_steps_per_second": 0.295, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 5.576113578589034e-08, | |
| "logits/chosen": -1.9287067651748657, | |
| "logits/rejected": -1.925451636314392, | |
| "logps/chosen": -31.603496551513672, | |
| "logps/rejected": -33.734046936035156, | |
| "loss": 1.0008, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": -1.4918483429937623e-05, | |
| "rewards/margins": -0.000825587019789964, | |
| "rewards/rejected": 0.0008106685127131641, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.229036944380912e-08, | |
| "logits/chosen": -1.9802377223968506, | |
| "logits/rejected": -1.9679291248321533, | |
| "logps/chosen": -34.585323333740234, | |
| "logps/rejected": -33.57084274291992, | |
| "loss": 0.9957, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": 0.00107504369225353, | |
| "rewards/margins": 0.0042753927409648895, | |
| "rewards/rejected": -0.0032003491651266813, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.053082288996112e-08, | |
| "logits/chosen": -2.015906810760498, | |
| "logits/rejected": -2.014427661895752, | |
| "logps/chosen": -33.49116516113281, | |
| "logps/rejected": -32.47978973388672, | |
| "loss": 1.0005, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": -0.0005370675935409963, | |
| "rewards/margins": -0.0005161463050171733, | |
| "rewards/rejected": -2.0921288523823023e-05, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.05793773749158e-08, | |
| "logits/chosen": -2.1030631065368652, | |
| "logits/rejected": -2.0872654914855957, | |
| "logps/chosen": -34.18492889404297, | |
| "logps/rejected": -33.08319854736328, | |
| "loss": 1.004, | |
| "rewards/accuracies": 0.42500001192092896, | |
| "rewards/chosen": -0.0005535687087103724, | |
| "rewards/margins": -0.00403733691200614, | |
| "rewards/rejected": 0.003483767854049802, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.251801807404168e-08, | |
| "logits/chosen": -1.9745471477508545, | |
| "logits/rejected": -1.9736032485961914, | |
| "logps/chosen": -33.23271942138672, | |
| "logps/rejected": -32.4765510559082, | |
| "loss": 0.9939, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.004937793128192425, | |
| "rewards/margins": 0.006078074686229229, | |
| "rewards/rejected": -0.001140281674452126, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 6.41315865106129e-09, | |
| "logits/chosen": -1.9305438995361328, | |
| "logits/rejected": -1.940913438796997, | |
| "logps/chosen": -32.22040939331055, | |
| "logps/rejected": -35.28728103637695, | |
| "loss": 1.0017, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.0015721082454547286, | |
| "rewards/margins": -0.001719816355034709, | |
| "rewards/rejected": 0.00014770813868381083, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.3150941078050324e-09, | |
| "logits/chosen": -2.069648265838623, | |
| "logits/rejected": -2.0630898475646973, | |
| "logps/chosen": -33.63695526123047, | |
| "logps/rejected": -29.226470947265625, | |
| "loss": 0.9985, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.0015011827927082777, | |
| "rewards/margins": 0.0015057541895657778, | |
| "rewards/rejected": -4.571699719235767e-06, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 2.575864278703266e-10, | |
| "logits/chosen": -1.928865671157837, | |
| "logits/rejected": -1.9310123920440674, | |
| "logps/chosen": -34.243560791015625, | |
| "logps/rejected": -30.892742156982422, | |
| "loss": 0.9984, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.0004508168494794518, | |
| "rewards/margins": 0.0015608124667778611, | |
| "rewards/rejected": -0.0020116292871534824, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 385, | |
| "total_flos": 0.0, | |
| "train_loss": 0.9991211135666092, | |
| "train_runtime": 3253.9902, | |
| "train_samples_per_second": 0.946, | |
| "train_steps_per_second": 0.118 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 385, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |