| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.898550724637681, | |
| "eval_steps": 40, | |
| "global_step": 400, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.036231884057971016, | |
| "grad_norm": 61.66789827681496, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": -2.7321553230285645, | |
| "logits/rejected": -2.7100937366485596, | |
| "logps/chosen": -182.5845489501953, | |
| "logps/rejected": -189.55001831054688, | |
| "loss": 0.689, | |
| "rewards/accuracies": 0.3187499940395355, | |
| "rewards/chosen": -0.0021577859297394753, | |
| "rewards/margins": 0.005646524019539356, | |
| "rewards/rejected": -0.007804309483617544, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.07246376811594203, | |
| "grad_norm": 44.96703657621111, | |
| "learning_rate": 1e-06, | |
| "logits/chosen": -2.753889799118042, | |
| "logits/rejected": -2.7519516944885254, | |
| "logps/chosen": -197.34320068359375, | |
| "logps/rejected": -184.00961303710938, | |
| "loss": 0.6275, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.030874451622366905, | |
| "rewards/margins": 0.18904080986976624, | |
| "rewards/rejected": -0.15816636383533478, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.10869565217391304, | |
| "grad_norm": 51.54402636298773, | |
| "learning_rate": 9.996221126793764e-07, | |
| "logits/chosen": -2.694437265396118, | |
| "logits/rejected": -2.691904067993164, | |
| "logps/chosen": -203.14883422851562, | |
| "logps/rejected": -204.52386474609375, | |
| "loss": 0.5851, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.6205412745475769, | |
| "rewards/margins": 0.9350436925888062, | |
| "rewards/rejected": -0.314502477645874, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.14492753623188406, | |
| "grad_norm": 35.11282014809116, | |
| "learning_rate": 9.984890219128145e-07, | |
| "logits/chosen": -2.609405517578125, | |
| "logits/rejected": -2.5795176029205322, | |
| "logps/chosen": -188.33395385742188, | |
| "logps/rejected": -192.52633666992188, | |
| "loss": 0.5144, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.8838651776313782, | |
| "rewards/margins": 1.275496244430542, | |
| "rewards/rejected": -0.39163118600845337, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.18115942028985507, | |
| "grad_norm": 37.05591291167376, | |
| "learning_rate": 9.966024404228493e-07, | |
| "logits/chosen": -2.4429798126220703, | |
| "logits/rejected": -2.4225034713745117, | |
| "logps/chosen": -179.79977416992188, | |
| "logps/rejected": -179.25279235839844, | |
| "loss": 0.5036, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.40532931685447693, | |
| "rewards/margins": 0.8947975039482117, | |
| "rewards/rejected": -0.4894680976867676, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.21739130434782608, | |
| "grad_norm": 33.27618771325657, | |
| "learning_rate": 9.939652198703783e-07, | |
| "logits/chosen": -2.319044589996338, | |
| "logits/rejected": -2.320253372192383, | |
| "logps/chosen": -188.62039184570312, | |
| "logps/rejected": -193.8806915283203, | |
| "loss": 0.4987, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.6481167078018188, | |
| "rewards/margins": 1.2183058261871338, | |
| "rewards/rejected": -0.5701891183853149, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.2536231884057971, | |
| "grad_norm": 38.645073192304736, | |
| "learning_rate": 9.905813465442354e-07, | |
| "logits/chosen": -2.242053747177124, | |
| "logits/rejected": -2.2155580520629883, | |
| "logps/chosen": -203.94546508789062, | |
| "logps/rejected": -194.8705291748047, | |
| "loss": 0.5091, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.8871867060661316, | |
| "rewards/margins": 1.2737131118774414, | |
| "rewards/rejected": -0.3865264356136322, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.2898550724637681, | |
| "grad_norm": 30.937642063399068, | |
| "learning_rate": 9.864559353357187e-07, | |
| "logits/chosen": -2.0998620986938477, | |
| "logits/rejected": -2.094836473464966, | |
| "logps/chosen": -182.6898956298828, | |
| "logps/rejected": -185.73983764648438, | |
| "loss": 0.4889, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 1.0410809516906738, | |
| "rewards/margins": 1.0222995281219482, | |
| "rewards/rejected": 0.018781563267111778, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.2898550724637681, | |
| "eval_logits/chosen": -2.035081624984741, | |
| "eval_logits/rejected": -1.9875677824020386, | |
| "eval_logps/chosen": -188.77523803710938, | |
| "eval_logps/rejected": -200.09498596191406, | |
| "eval_loss": 0.4642000198364258, | |
| "eval_rewards/accuracies": 0.7943548560142517, | |
| "eval_rewards/chosen": 1.1544201374053955, | |
| "eval_rewards/margins": 1.3431099653244019, | |
| "eval_rewards/rejected": -0.18868987262248993, | |
| "eval_runtime": 248.6586, | |
| "eval_samples_per_second": 15.773, | |
| "eval_steps_per_second": 0.249, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.32608695652173914, | |
| "grad_norm": 29.205255063362404, | |
| "learning_rate": 9.815952220071804e-07, | |
| "logits/chosen": -1.9733244180679321, | |
| "logits/rejected": -1.9242098331451416, | |
| "logps/chosen": -195.65652465820312, | |
| "logps/rejected": -221.744384765625, | |
| "loss": 0.4587, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 1.3803393840789795, | |
| "rewards/margins": 1.8610279560089111, | |
| "rewards/rejected": -0.4806883931159973, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.36231884057971014, | |
| "grad_norm": 29.678687938529745, | |
| "learning_rate": 9.76006553766365e-07, | |
| "logits/chosen": -1.8085625171661377, | |
| "logits/rejected": -1.772071123123169, | |
| "logps/chosen": -198.9405517578125, | |
| "logps/rejected": -203.29978942871094, | |
| "loss": 0.4531, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": 0.8211376070976257, | |
| "rewards/margins": 1.646414041519165, | |
| "rewards/rejected": -0.8252763748168945, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.39855072463768115, | |
| "grad_norm": 30.73732601293307, | |
| "learning_rate": 9.696983781607415e-07, | |
| "logits/chosen": -1.7908868789672852, | |
| "logits/rejected": -1.7599233388900757, | |
| "logps/chosen": -183.01809692382812, | |
| "logps/rejected": -172.02801513671875, | |
| "loss": 0.4595, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": 0.607209324836731, | |
| "rewards/margins": 1.649515151977539, | |
| "rewards/rejected": -1.042305827140808, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.43478260869565216, | |
| "grad_norm": 41.49881496846909, | |
| "learning_rate": 9.626802303086209e-07, | |
| "logits/chosen": -1.9050066471099854, | |
| "logits/rejected": -1.8625679016113281, | |
| "logps/chosen": -186.8179931640625, | |
| "logps/rejected": -194.26707458496094, | |
| "loss": 0.4276, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": 0.4140642285346985, | |
| "rewards/margins": 1.712418794631958, | |
| "rewards/rejected": -1.2983543872833252, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.47101449275362317, | |
| "grad_norm": 36.480881476794956, | |
| "learning_rate": 9.549627184863528e-07, | |
| "logits/chosen": -2.1638290882110596, | |
| "logits/rejected": -2.0793471336364746, | |
| "logps/chosen": -191.9010467529297, | |
| "logps/rejected": -193.20651245117188, | |
| "loss": 0.4293, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.045776158571243286, | |
| "rewards/margins": 1.6033704280853271, | |
| "rewards/rejected": -1.5575940608978271, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.5072463768115942, | |
| "grad_norm": 26.318712537847357, | |
| "learning_rate": 9.465575080933957e-07, | |
| "logits/chosen": -2.1214632987976074, | |
| "logits/rejected": -2.081162214279175, | |
| "logps/chosen": -172.66419982910156, | |
| "logps/rejected": -208.19149780273438, | |
| "loss": 0.3963, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": 0.17696644365787506, | |
| "rewards/margins": 1.8183233737945557, | |
| "rewards/rejected": -1.6413570642471313, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.5434782608695652, | |
| "grad_norm": 32.35315227999789, | |
| "learning_rate": 9.374773040194878e-07, | |
| "logits/chosen": -2.1271562576293945, | |
| "logits/rejected": -2.0680038928985596, | |
| "logps/chosen": -205.5594482421875, | |
| "logps/rejected": -210.7834014892578, | |
| "loss": 0.4353, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 0.16769471764564514, | |
| "rewards/margins": 1.786273717880249, | |
| "rewards/rejected": -1.6185792684555054, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.5797101449275363, | |
| "grad_norm": 29.96593011887677, | |
| "learning_rate": 9.277358314405818e-07, | |
| "logits/chosen": -2.046504497528076, | |
| "logits/rejected": -2.0122628211975098, | |
| "logps/chosen": -189.37771606445312, | |
| "logps/rejected": -206.10659790039062, | |
| "loss": 0.3941, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.2374907284975052, | |
| "rewards/margins": 1.910300612449646, | |
| "rewards/rejected": -2.1477913856506348, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.5797101449275363, | |
| "eval_logits/chosen": -1.9900643825531006, | |
| "eval_logits/rejected": -1.9449083805084229, | |
| "eval_logps/chosen": -202.59437561035156, | |
| "eval_logps/rejected": -221.12725830078125, | |
| "eval_loss": 0.4217630624771118, | |
| "eval_rewards/accuracies": 0.8044354915618896, | |
| "eval_rewards/chosen": -0.22749020159244537, | |
| "eval_rewards/margins": 2.0644266605377197, | |
| "eval_rewards/rejected": -2.291916847229004, | |
| "eval_runtime": 247.3121, | |
| "eval_samples_per_second": 15.859, | |
| "eval_steps_per_second": 0.251, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.6159420289855072, | |
| "grad_norm": 34.207898226903815, | |
| "learning_rate": 9.173478150725651e-07, | |
| "logits/chosen": -2.0169568061828613, | |
| "logits/rejected": -1.9391515254974365, | |
| "logps/chosen": -209.73440551757812, | |
| "logps/rejected": -215.14205932617188, | |
| "loss": 0.422, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": 0.1130056157708168, | |
| "rewards/margins": 2.3998470306396484, | |
| "rewards/rejected": -2.2868411540985107, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.6521739130434783, | |
| "grad_norm": 28.850805260840193, | |
| "learning_rate": 9.063289569141251e-07, | |
| "logits/chosen": -2.1180360317230225, | |
| "logits/rejected": -2.080235242843628, | |
| "logps/chosen": -214.2292022705078, | |
| "logps/rejected": -223.37564086914062, | |
| "loss": 0.4135, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 0.5250645279884338, | |
| "rewards/margins": 2.2779579162597656, | |
| "rewards/rejected": -1.7528936862945557, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.6884057971014492, | |
| "grad_norm": 34.73111091101309, | |
| "learning_rate": 8.946959125124051e-07, | |
| "logits/chosen": -2.2047152519226074, | |
| "logits/rejected": -2.1581788063049316, | |
| "logps/chosen": -209.3234405517578, | |
| "logps/rejected": -195.6946563720703, | |
| "loss": 0.4077, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 0.25443512201309204, | |
| "rewards/margins": 2.2029635906219482, | |
| "rewards/rejected": -1.948528528213501, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 0.7246376811594203, | |
| "grad_norm": 32.202226649445414, | |
| "learning_rate": 8.824662657873238e-07, | |
| "logits/chosen": -2.256134510040283, | |
| "logits/rejected": -2.2395310401916504, | |
| "logps/chosen": -176.19464111328125, | |
| "logps/rejected": -209.0999298095703, | |
| "loss": 0.3767, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.3457742929458618, | |
| "rewards/margins": 1.9383150339126587, | |
| "rewards/rejected": -2.2840893268585205, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.7608695652173914, | |
| "grad_norm": 31.35693490170537, | |
| "learning_rate": 8.696585024526135e-07, | |
| "logits/chosen": -2.3358142375946045, | |
| "logits/rejected": -2.307068347930908, | |
| "logps/chosen": -191.8298797607422, | |
| "logps/rejected": -227.55612182617188, | |
| "loss": 0.4017, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.2825666069984436, | |
| "rewards/margins": 2.5402400493621826, | |
| "rewards/rejected": -2.8228065967559814, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 0.7971014492753623, | |
| "grad_norm": 36.29919961993296, | |
| "learning_rate": 8.562919820737535e-07, | |
| "logits/chosen": -2.2931602001190186, | |
| "logits/rejected": -2.2568557262420654, | |
| "logps/chosen": -207.34158325195312, | |
| "logps/rejected": -208.9143524169922, | |
| "loss": 0.3767, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.3017815351486206, | |
| "rewards/margins": 2.3830084800720215, | |
| "rewards/rejected": -2.6847901344299316, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.8333333333333334, | |
| "grad_norm": 28.819478551304137, | |
| "learning_rate": 8.423869088050315e-07, | |
| "logits/chosen": -2.2734763622283936, | |
| "logits/rejected": -2.2522785663604736, | |
| "logps/chosen": -195.6924591064453, | |
| "logps/rejected": -221.85104370117188, | |
| "loss": 0.3892, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.12773282825946808, | |
| "rewards/margins": 2.3888843059539795, | |
| "rewards/rejected": -2.5166170597076416, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 0.8695652173913043, | |
| "grad_norm": 37.085951928654964, | |
| "learning_rate": 8.2796430084997e-07, | |
| "logits/chosen": -2.2070839405059814, | |
| "logits/rejected": -2.1736464500427246, | |
| "logps/chosen": -196.70889282226562, | |
| "logps/rejected": -207.93417358398438, | |
| "loss": 0.3717, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.12862932682037354, | |
| "rewards/margins": 2.698848009109497, | |
| "rewards/rejected": -2.570218563079834, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.8695652173913043, | |
| "eval_logits/chosen": -2.122933864593506, | |
| "eval_logits/rejected": -2.0901710987091064, | |
| "eval_logps/chosen": -202.4208221435547, | |
| "eval_logps/rejected": -223.09356689453125, | |
| "eval_loss": 0.438678115606308, | |
| "eval_rewards/accuracies": 0.8286290168762207, | |
| "eval_rewards/chosen": -0.2101391851902008, | |
| "eval_rewards/margins": 2.278407573699951, | |
| "eval_rewards/rejected": -2.488546848297119, | |
| "eval_runtime": 247.3754, | |
| "eval_samples_per_second": 15.854, | |
| "eval_steps_per_second": 0.251, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.9057971014492754, | |
| "grad_norm": 27.843666767086823, | |
| "learning_rate": 8.130459586912753e-07, | |
| "logits/chosen": -2.0930094718933105, | |
| "logits/rejected": -2.0791330337524414, | |
| "logps/chosen": -220.4320526123047, | |
| "logps/rejected": -218.3594512939453, | |
| "loss": 0.4575, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.9592973589897156, | |
| "rewards/margins": 1.66092050075531, | |
| "rewards/rejected": -2.620218276977539, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 0.9420289855072463, | |
| "grad_norm": 28.902451958691426, | |
| "learning_rate": 7.97654432138333e-07, | |
| "logits/chosen": -2.1393418312072754, | |
| "logits/rejected": -2.103482723236084, | |
| "logps/chosen": -214.5059051513672, | |
| "logps/rejected": -248.8673858642578, | |
| "loss": 0.388, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -0.1911260336637497, | |
| "rewards/margins": 2.8205361366271973, | |
| "rewards/rejected": -3.011662006378174, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.9782608695652174, | |
| "grad_norm": 26.45174989804514, | |
| "learning_rate": 7.81812986242061e-07, | |
| "logits/chosen": -2.2018837928771973, | |
| "logits/rejected": -2.1459240913391113, | |
| "logps/chosen": -193.5723419189453, | |
| "logps/rejected": -233.81838989257812, | |
| "loss": 0.365, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": 0.10008885711431503, | |
| "rewards/margins": 3.072911262512207, | |
| "rewards/rejected": -2.972822427749634, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 1.0144927536231885, | |
| "grad_norm": 16.755101670108846, | |
| "learning_rate": 7.655455661286375e-07, | |
| "logits/chosen": -2.190566301345825, | |
| "logits/rejected": -2.1852922439575195, | |
| "logps/chosen": -190.56027221679688, | |
| "logps/rejected": -234.6202392578125, | |
| "loss": 0.2524, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": 0.05487387627363205, | |
| "rewards/margins": 3.58606219291687, | |
| "rewards/rejected": -3.531188488006592, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.0507246376811594, | |
| "grad_norm": 16.022443175608785, | |
| "learning_rate": 7.488767608052628e-07, | |
| "logits/chosen": -2.2735958099365234, | |
| "logits/rejected": -2.199552059173584, | |
| "logps/chosen": -190.21511840820312, | |
| "logps/rejected": -237.59854125976562, | |
| "loss": 0.171, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 0.7184330821037292, | |
| "rewards/margins": 4.115548133850098, | |
| "rewards/rejected": -3.3971149921417236, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 1.0869565217391304, | |
| "grad_norm": 16.37925141930124, | |
| "learning_rate": 7.318317659926636e-07, | |
| "logits/chosen": -2.250054121017456, | |
| "logits/rejected": -2.221043109893799, | |
| "logps/chosen": -174.0218048095703, | |
| "logps/rejected": -237.14120483398438, | |
| "loss": 0.1436, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.0139648914337158, | |
| "rewards/margins": 4.169236660003662, | |
| "rewards/rejected": -3.1552722454071045, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.1231884057971016, | |
| "grad_norm": 13.693805739113818, | |
| "learning_rate": 7.144363460405189e-07, | |
| "logits/chosen": -2.3314082622528076, | |
| "logits/rejected": -2.2712883949279785, | |
| "logps/chosen": -192.2017059326172, | |
| "logps/rejected": -236.91586303710938, | |
| "loss": 0.1415, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 1.1952037811279297, | |
| "rewards/margins": 4.705626487731934, | |
| "rewards/rejected": -3.510422945022583, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 1.1594202898550725, | |
| "grad_norm": 13.203700040695288, | |
| "learning_rate": 6.967167949833762e-07, | |
| "logits/chosen": -2.324735164642334, | |
| "logits/rejected": -2.283231258392334, | |
| "logps/chosen": -194.0807342529297, | |
| "logps/rejected": -247.2218475341797, | |
| "loss": 0.1459, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.3761358857154846, | |
| "rewards/margins": 4.9711785316467285, | |
| "rewards/rejected": -4.595042705535889, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.1594202898550725, | |
| "eval_logits/chosen": -2.3007450103759766, | |
| "eval_logits/rejected": -2.273284673690796, | |
| "eval_logps/chosen": -204.34878540039062, | |
| "eval_logps/rejected": -232.13629150390625, | |
| "eval_loss": 0.42876046895980835, | |
| "eval_rewards/accuracies": 0.8286290168762207, | |
| "eval_rewards/chosen": -0.40293240547180176, | |
| "eval_rewards/margins": 2.9898877143859863, | |
| "eval_rewards/rejected": -3.392819881439209, | |
| "eval_runtime": 247.3207, | |
| "eval_samples_per_second": 15.858, | |
| "eval_steps_per_second": 0.251, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.1956521739130435, | |
| "grad_norm": 17.287170066985027, | |
| "learning_rate": 6.786998967959219e-07, | |
| "logits/chosen": -2.313910961151123, | |
| "logits/rejected": -2.274190902709961, | |
| "logps/chosen": -200.9847869873047, | |
| "logps/rejected": -228.14138793945312, | |
| "loss": 0.1497, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 0.39603787660598755, | |
| "rewards/margins": 4.507396697998047, | |
| "rewards/rejected": -4.111359119415283, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 1.2318840579710144, | |
| "grad_norm": 16.43946047429659, | |
| "learning_rate": 6.604128849076838e-07, | |
| "logits/chosen": -2.340721607208252, | |
| "logits/rejected": -2.305487871170044, | |
| "logps/chosen": -200.94406127929688, | |
| "logps/rejected": -238.47921752929688, | |
| "loss": 0.1533, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 0.9461654424667358, | |
| "rewards/margins": 4.6484904289245605, | |
| "rewards/rejected": -3.702324390411377, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.2681159420289856, | |
| "grad_norm": 13.545354366232296, | |
| "learning_rate": 6.418834010383609e-07, | |
| "logits/chosen": -2.400474786758423, | |
| "logits/rejected": -2.343231439590454, | |
| "logps/chosen": -172.61148071289062, | |
| "logps/rejected": -230.24197387695312, | |
| "loss": 0.1459, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 0.5564432144165039, | |
| "rewards/margins": 4.526711463928223, | |
| "rewards/rejected": -3.9702675342559814, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.3043478260869565, | |
| "grad_norm": 16.208760712520707, | |
| "learning_rate": 6.231394534160007e-07, | |
| "logits/chosen": -2.3880228996276855, | |
| "logits/rejected": -2.377924680709839, | |
| "logps/chosen": -186.1154327392578, | |
| "logps/rejected": -228.6842498779297, | |
| "loss": 0.1408, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": 1.147924542427063, | |
| "rewards/margins": 4.681893348693848, | |
| "rewards/rejected": -3.533968687057495, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.3405797101449275, | |
| "grad_norm": 16.391749704364813, | |
| "learning_rate": 6.042093744411828e-07, | |
| "logits/chosen": -2.3000502586364746, | |
| "logits/rejected": -2.280015468597412, | |
| "logps/chosen": -184.9687042236328, | |
| "logps/rejected": -230.6714630126953, | |
| "loss": 0.1426, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 1.1101362705230713, | |
| "rewards/margins": 4.539022922515869, | |
| "rewards/rejected": -3.4288864135742188, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 1.3768115942028984, | |
| "grad_norm": 15.103230138315158, | |
| "learning_rate": 5.851217778611993e-07, | |
| "logits/chosen": -2.2735018730163574, | |
| "logits/rejected": -2.272218942642212, | |
| "logps/chosen": -196.11097717285156, | |
| "logps/rejected": -217.5700225830078, | |
| "loss": 0.1351, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 0.8358832597732544, | |
| "rewards/margins": 4.623397350311279, | |
| "rewards/rejected": -3.7875142097473145, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 1.4130434782608696, | |
| "grad_norm": 23.621795316466446, | |
| "learning_rate": 5.659055155189651e-07, | |
| "logits/chosen": -2.325028896331787, | |
| "logits/rejected": -2.252711296081543, | |
| "logps/chosen": -193.0414276123047, | |
| "logps/rejected": -231.9776611328125, | |
| "loss": 0.1542, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 0.06639621406793594, | |
| "rewards/margins": 4.913178443908691, | |
| "rewards/rejected": -4.8467817306518555, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 1.4492753623188406, | |
| "grad_norm": 11.913033785498357, | |
| "learning_rate": 5.465896337420358e-07, | |
| "logits/chosen": -2.3189964294433594, | |
| "logits/rejected": -2.2462618350982666, | |
| "logps/chosen": -207.55685424804688, | |
| "logps/rejected": -269.5002746582031, | |
| "loss": 0.1455, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 0.3675020635128021, | |
| "rewards/margins": 5.308381080627441, | |
| "rewards/rejected": -4.940878868103027, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.4492753623188406, | |
| "eval_logits/chosen": -2.2696547508239746, | |
| "eval_logits/rejected": -2.246581554412842, | |
| "eval_logps/chosen": -205.6576690673828, | |
| "eval_logps/rejected": -234.53866577148438, | |
| "eval_loss": 0.42552247643470764, | |
| "eval_rewards/accuracies": 0.8165322542190552, | |
| "eval_rewards/chosen": -0.5338226556777954, | |
| "eval_rewards/margins": 3.099236011505127, | |
| "eval_rewards/rejected": -3.633058547973633, | |
| "eval_runtime": 247.2511, | |
| "eval_samples_per_second": 15.862, | |
| "eval_steps_per_second": 0.251, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.4855072463768115, | |
| "grad_norm": 21.298531736271027, | |
| "learning_rate": 5.272033294376521e-07, | |
| "logits/chosen": -2.2834343910217285, | |
| "logits/rejected": -2.2502143383026123, | |
| "logps/chosen": -191.12301635742188, | |
| "logps/rejected": -221.69924926757812, | |
| "loss": 0.1402, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 0.3291940987110138, | |
| "rewards/margins": 4.7104597091674805, | |
| "rewards/rejected": -4.381266117095947, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 1.5217391304347827, | |
| "grad_norm": 16.624769093346845, | |
| "learning_rate": 5.077759059601755e-07, | |
| "logits/chosen": -2.312807559967041, | |
| "logits/rejected": -2.29304838180542, | |
| "logps/chosen": -206.1201629638672, | |
| "logps/rejected": -224.8920135498047, | |
| "loss": 0.1564, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.7266442775726318, | |
| "rewards/margins": 5.301372528076172, | |
| "rewards/rejected": -4.574727535247803, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.5579710144927537, | |
| "grad_norm": 17.710918361893075, | |
| "learning_rate": 4.883367288176238e-07, | |
| "logits/chosen": -2.325953483581543, | |
| "logits/rejected": -2.355797290802002, | |
| "logps/chosen": -178.1147918701172, | |
| "logps/rejected": -228.8972930908203, | |
| "loss": 0.1462, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.0754575729370117, | |
| "rewards/margins": 5.179568290710449, | |
| "rewards/rejected": -4.104111194610596, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 1.5942028985507246, | |
| "grad_norm": 17.61185095540283, | |
| "learning_rate": 4.6891518128425974e-07, | |
| "logits/chosen": -2.3683056831359863, | |
| "logits/rejected": -2.327507495880127, | |
| "logps/chosen": -198.29257202148438, | |
| "logps/rejected": -239.8541717529297, | |
| "loss": 0.149, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.2280995845794678, | |
| "rewards/margins": 5.207882881164551, | |
| "rewards/rejected": -3.979783535003662, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.6304347826086958, | |
| "grad_norm": 18.295935717153018, | |
| "learning_rate": 4.495406199863217e-07, | |
| "logits/chosen": -2.3297061920166016, | |
| "logits/rejected": -2.3166041374206543, | |
| "logps/chosen": -180.5458221435547, | |
| "logps/rejected": -259.0296630859375, | |
| "loss": 0.1258, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 0.7601513266563416, | |
| "rewards/margins": 5.343691825866699, | |
| "rewards/rejected": -4.583540916442871, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 1.6666666666666665, | |
| "grad_norm": 20.56329538942416, | |
| "learning_rate": 4.302423305280385e-07, | |
| "logits/chosen": -2.342526912689209, | |
| "logits/rejected": -2.293287992477417, | |
| "logps/chosen": -181.458251953125, | |
| "logps/rejected": -268.34442138671875, | |
| "loss": 0.1301, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 0.30324190855026245, | |
| "rewards/margins": 5.840424060821533, | |
| "rewards/rejected": -5.537181854248047, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 1.7028985507246377, | |
| "grad_norm": 18.749044820633504, | |
| "learning_rate": 4.1104948322499386e-07, | |
| "logits/chosen": -2.36098575592041, | |
| "logits/rejected": -2.310408353805542, | |
| "logps/chosen": -187.14903259277344, | |
| "logps/rejected": -245.0007781982422, | |
| "loss": 0.1278, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.39717116951942444, | |
| "rewards/margins": 5.458386421203613, | |
| "rewards/rejected": -5.061215400695801, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 1.7391304347826086, | |
| "grad_norm": 19.17022127722479, | |
| "learning_rate": 3.919910890117584e-07, | |
| "logits/chosen": -2.3718903064727783, | |
| "logits/rejected": -2.348759889602661, | |
| "logps/chosen": -180.43243408203125, | |
| "logps/rejected": -232.767333984375, | |
| "loss": 0.1358, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 0.7176074981689453, | |
| "rewards/margins": 5.331349849700928, | |
| "rewards/rejected": -4.613741874694824, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.7391304347826086, | |
| "eval_logits/chosen": -2.3806025981903076, | |
| "eval_logits/rejected": -2.360522508621216, | |
| "eval_logps/chosen": -203.0333251953125, | |
| "eval_logps/rejected": -234.9227294921875, | |
| "eval_loss": 0.4247341454029083, | |
| "eval_rewards/accuracies": 0.8326612710952759, | |
| "eval_rewards/chosen": -0.27138766646385193, | |
| "eval_rewards/margins": 3.4000766277313232, | |
| "eval_rewards/rejected": -3.671464204788208, | |
| "eval_runtime": 247.4296, | |
| "eval_samples_per_second": 15.851, | |
| "eval_steps_per_second": 0.251, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 1.7753623188405796, | |
| "grad_norm": 16.252609084074862, | |
| "learning_rate": 3.7309595559042973e-07, | |
| "logits/chosen": -2.3761582374572754, | |
| "logits/rejected": -2.371837615966797, | |
| "logps/chosen": -185.31297302246094, | |
| "logps/rejected": -231.5406951904297, | |
| "loss": 0.1311, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 0.5640401840209961, | |
| "rewards/margins": 5.0947370529174805, | |
| "rewards/rejected": -4.530695915222168, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 1.8115942028985508, | |
| "grad_norm": 20.868443403720576, | |
| "learning_rate": 3.54392643886374e-07, | |
| "logits/chosen": -2.4362902641296387, | |
| "logits/rejected": -2.405050277709961, | |
| "logps/chosen": -186.8355255126953, | |
| "logps/rejected": -232.0409698486328, | |
| "loss": 0.1252, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 0.699629545211792, | |
| "rewards/margins": 5.528279781341553, | |
| "rewards/rejected": -4.82865047454834, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 1.8478260869565217, | |
| "grad_norm": 20.235567534724474, | |
| "learning_rate": 3.3590942487697765e-07, | |
| "logits/chosen": -2.4567503929138184, | |
| "logits/rejected": -2.444530725479126, | |
| "logps/chosen": -186.48788452148438, | |
| "logps/rejected": -223.74258422851562, | |
| "loss": 0.1613, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": 0.8208659291267395, | |
| "rewards/margins": 4.932476997375488, | |
| "rewards/rejected": -4.111611366271973, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 1.8840579710144927, | |
| "grad_norm": 17.41545455778026, | |
| "learning_rate": 3.176742368586725e-07, | |
| "logits/chosen": -2.452131986618042, | |
| "logits/rejected": -2.415022373199463, | |
| "logps/chosen": -196.01910400390625, | |
| "logps/rejected": -217.8632354736328, | |
| "loss": 0.1348, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 0.8574349284172058, | |
| "rewards/margins": 5.437888145446777, | |
| "rewards/rejected": -4.580452919006348, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 1.9202898550724639, | |
| "grad_norm": 16.095749631795933, | |
| "learning_rate": 2.997146432168236e-07, | |
| "logits/chosen": -2.44551420211792, | |
| "logits/rejected": -2.410526990890503, | |
| "logps/chosen": -192.61859130859375, | |
| "logps/rejected": -246.4084014892578, | |
| "loss": 0.1427, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 0.6556909084320068, | |
| "rewards/margins": 6.007452964782715, | |
| "rewards/rejected": -5.351761817932129, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 1.9565217391304348, | |
| "grad_norm": 20.740869553463348, | |
| "learning_rate": 2.8205779076231446e-07, | |
| "logits/chosen": -2.430431842803955, | |
| "logits/rejected": -2.4140846729278564, | |
| "logps/chosen": -191.99880981445312, | |
| "logps/rejected": -238.8526611328125, | |
| "loss": 0.1257, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.9606460332870483, | |
| "rewards/margins": 5.507210731506348, | |
| "rewards/rejected": -4.54656457901001, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 1.9927536231884058, | |
| "grad_norm": 19.416516281755985, | |
| "learning_rate": 2.647303686978035e-07, | |
| "logits/chosen": -2.3993430137634277, | |
| "logits/rejected": -2.376300811767578, | |
| "logps/chosen": -185.33132934570312, | |
| "logps/rejected": -219.64968872070312, | |
| "loss": 0.1492, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 0.5704915523529053, | |
| "rewards/margins": 4.9512529373168945, | |
| "rewards/rejected": -4.38076114654541, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.028985507246377, | |
| "grad_norm": 8.699243855503047, | |
| "learning_rate": 2.4775856827568014e-07, | |
| "logits/chosen": -2.397235155105591, | |
| "logits/rejected": -2.38557767868042, | |
| "logps/chosen": -187.7579803466797, | |
| "logps/rejected": -224.9508514404297, | |
| "loss": 0.0938, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.7930246591567993, | |
| "rewards/margins": 5.4677629470825195, | |
| "rewards/rejected": -4.674737930297852, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.028985507246377, | |
| "eval_logits/chosen": -2.3933310508728027, | |
| "eval_logits/rejected": -2.372490167617798, | |
| "eval_logps/chosen": -203.4556121826172, | |
| "eval_logps/rejected": -235.21469116210938, | |
| "eval_loss": 0.41276049613952637, | |
| "eval_rewards/accuracies": 0.8266128897666931, | |
| "eval_rewards/chosen": -0.313615620136261, | |
| "eval_rewards/margins": 3.3870484828948975, | |
| "eval_rewards/rejected": -3.7006635665893555, | |
| "eval_runtime": 247.2901, | |
| "eval_samples_per_second": 15.86, | |
| "eval_steps_per_second": 0.251, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 2.0652173913043477, | |
| "grad_norm": 8.518008709307914, | |
| "learning_rate": 2.3116804320869464e-07, | |
| "logits/chosen": -2.3717570304870605, | |
| "logits/rejected": -2.358806610107422, | |
| "logps/chosen": -189.88809204101562, | |
| "logps/rejected": -226.0210418701172, | |
| "loss": 0.072, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": 0.5044176578521729, | |
| "rewards/margins": 5.758059978485107, | |
| "rewards/rejected": -5.253642559051514, | |
| "step": 285 | |
| }, | |
| { | |
| "epoch": 2.101449275362319, | |
| "grad_norm": 10.070711388481515, | |
| "learning_rate": 2.1498387089310865e-07, | |
| "logits/chosen": -2.3519985675811768, | |
| "logits/rejected": -2.361879348754883, | |
| "logps/chosen": -195.57469177246094, | |
| "logps/rejected": -251.92440795898438, | |
| "loss": 0.0702, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 1.01186203956604, | |
| "rewards/margins": 6.105987548828125, | |
| "rewards/rejected": -5.094125747680664, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 2.13768115942029, | |
| "grad_norm": 9.472204652777094, | |
| "learning_rate": 1.9923051450297336e-07, | |
| "logits/chosen": -2.384852647781372, | |
| "logits/rejected": -2.346426486968994, | |
| "logps/chosen": -194.14443969726562, | |
| "logps/rejected": -229.8656463623047, | |
| "loss": 0.0725, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": 1.0537793636322021, | |
| "rewards/margins": 6.0652875900268555, | |
| "rewards/rejected": -5.011508464813232, | |
| "step": 295 | |
| }, | |
| { | |
| "epoch": 2.1739130434782608, | |
| "grad_norm": 11.964654423069804, | |
| "learning_rate": 1.839317860128368e-07, | |
| "logits/chosen": -2.368323802947998, | |
| "logits/rejected": -2.3484463691711426, | |
| "logps/chosen": -190.7874298095703, | |
| "logps/rejected": -249.8258514404297, | |
| "loss": 0.0649, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": 0.9586502909660339, | |
| "rewards/margins": 6.444998264312744, | |
| "rewards/rejected": -5.4863481521606445, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 2.210144927536232, | |
| "grad_norm": 14.363847590302074, | |
| "learning_rate": 1.6911081020477176e-07, | |
| "logits/chosen": -2.350074291229248, | |
| "logits/rejected": -2.3347928524017334, | |
| "logps/chosen": -188.36700439453125, | |
| "logps/rejected": -253.87844848632812, | |
| "loss": 0.0638, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 0.843727707862854, | |
| "rewards/margins": 6.143573760986328, | |
| "rewards/rejected": -5.299846649169922, | |
| "step": 305 | |
| }, | |
| { | |
| "epoch": 2.246376811594203, | |
| "grad_norm": 12.207090391342035, | |
| "learning_rate": 1.5478998971412666e-07, | |
| "logits/chosen": -2.3135275840759277, | |
| "logits/rejected": -2.302867889404297, | |
| "logps/chosen": -190.8241424560547, | |
| "logps/rejected": -256.7770080566406, | |
| "loss": 0.0644, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": 0.7945331931114197, | |
| "rewards/margins": 6.6735687255859375, | |
| "rewards/rejected": -5.879034996032715, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 2.282608695652174, | |
| "grad_norm": 17.814039974196998, | |
| "learning_rate": 1.4099097116683873e-07, | |
| "logits/chosen": -2.3371047973632812, | |
| "logits/rejected": -2.318176746368408, | |
| "logps/chosen": -211.5010986328125, | |
| "logps/rejected": -276.4696350097656, | |
| "loss": 0.0609, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.2731298208236694, | |
| "rewards/margins": 7.202207088470459, | |
| "rewards/rejected": -5.9290771484375, | |
| "step": 315 | |
| }, | |
| { | |
| "epoch": 2.318840579710145, | |
| "grad_norm": 12.87715616525982, | |
| "learning_rate": 1.2773461245949247e-07, | |
| "logits/chosen": -2.339965343475342, | |
| "logits/rejected": -2.3124070167541504, | |
| "logps/chosen": -203.6749725341797, | |
| "logps/rejected": -248.37149047851562, | |
| "loss": 0.0592, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": 0.8933491706848145, | |
| "rewards/margins": 6.6305975914001465, | |
| "rewards/rejected": -5.737248420715332, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.318840579710145, | |
| "eval_logits/chosen": -2.335758686065674, | |
| "eval_logits/rejected": -2.310899257659912, | |
| "eval_logps/chosen": -206.08692932128906, | |
| "eval_logps/rejected": -239.44287109375, | |
| "eval_loss": 0.4438334107398987, | |
| "eval_rewards/accuracies": 0.8165322542190552, | |
| "eval_rewards/chosen": -0.576747715473175, | |
| "eval_rewards/margins": 3.546731948852539, | |
| "eval_rewards/rejected": -4.123479843139648, | |
| "eval_runtime": 247.2469, | |
| "eval_samples_per_second": 15.863, | |
| "eval_steps_per_second": 0.251, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 2.355072463768116, | |
| "grad_norm": 12.127468615559982, | |
| "learning_rate": 1.1504095123158014e-07, | |
| "logits/chosen": -2.3464112281799316, | |
| "logits/rejected": -2.3510231971740723, | |
| "logps/chosen": -194.9401397705078, | |
| "logps/rejected": -256.2391052246094, | |
| "loss": 0.0642, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": 0.7289390563964844, | |
| "rewards/margins": 6.520253658294678, | |
| "rewards/rejected": -5.791314125061035, | |
| "step": 325 | |
| }, | |
| { | |
| "epoch": 2.391304347826087, | |
| "grad_norm": 9.771341581861016, | |
| "learning_rate": 1.0292917457762323e-07, | |
| "logits/chosen": -2.326569080352783, | |
| "logits/rejected": -2.2930023670196533, | |
| "logps/chosen": -188.14108276367188, | |
| "logps/rejected": -246.42495727539062, | |
| "loss": 0.0575, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": 0.9616584777832031, | |
| "rewards/margins": 6.517239570617676, | |
| "rewards/rejected": -5.555581092834473, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 2.427536231884058, | |
| "grad_norm": 14.867872053501438, | |
| "learning_rate": 9.141759004493282e-08, | |
| "logits/chosen": -2.304109573364258, | |
| "logits/rejected": -2.3050553798675537, | |
| "logps/chosen": -179.35818481445312, | |
| "logps/rejected": -243.79220581054688, | |
| "loss": 0.0539, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": 0.44949406385421753, | |
| "rewards/margins": 6.40158224105835, | |
| "rewards/rejected": -5.952088832855225, | |
| "step": 335 | |
| }, | |
| { | |
| "epoch": 2.463768115942029, | |
| "grad_norm": 15.245309497391487, | |
| "learning_rate": 8.052359796084951e-08, | |
| "logits/chosen": -2.3356070518493652, | |
| "logits/rejected": -2.2882630825042725, | |
| "logps/chosen": -191.533447265625, | |
| "logps/rejected": -248.5685577392578, | |
| "loss": 0.0649, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 1.079332947731018, | |
| "rewards/margins": 6.897538185119629, | |
| "rewards/rejected": -5.818203926086426, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 2.5, | |
| "grad_norm": 15.656947068033539, | |
| "learning_rate": 7.026366513129139e-08, | |
| "logits/chosen": -2.320748805999756, | |
| "logits/rejected": -2.2670602798461914, | |
| "logps/chosen": -184.66661071777344, | |
| "logps/rejected": -229.6528778076172, | |
| "loss": 0.0628, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.8999651670455933, | |
| "rewards/margins": 6.056799411773682, | |
| "rewards/rejected": -5.156834125518799, | |
| "step": 345 | |
| }, | |
| { | |
| "epoch": 2.536231884057971, | |
| "grad_norm": 10.866006807463382, | |
| "learning_rate": 6.065329995036572e-08, | |
| "logits/chosen": -2.352247953414917, | |
| "logits/rejected": -2.3084654808044434, | |
| "logps/chosen": -191.09317016601562, | |
| "logps/rejected": -235.8540496826172, | |
| "loss": 0.0577, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.519118070602417, | |
| "rewards/margins": 5.989134311676025, | |
| "rewards/rejected": -5.470016002655029, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 2.572463768115942, | |
| "grad_norm": 13.773720162250582, | |
| "learning_rate": 5.170702895866591e-08, | |
| "logits/chosen": -2.3459525108337402, | |
| "logits/rejected": -2.2808985710144043, | |
| "logps/chosen": -181.71859741210938, | |
| "logps/rejected": -234.2205352783203, | |
| "loss": 0.0527, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": 0.7595393061637878, | |
| "rewards/margins": 6.467637062072754, | |
| "rewards/rejected": -5.708098411560059, | |
| "step": 355 | |
| }, | |
| { | |
| "epoch": 2.608695652173913, | |
| "grad_norm": 17.050512305140668, | |
| "learning_rate": 4.343837488569057e-08, | |
| "logits/chosen": -2.3642101287841797, | |
| "logits/rejected": -2.334510087966919, | |
| "logps/chosen": -188.16355895996094, | |
| "logps/rejected": -243.60531616210938, | |
| "loss": 0.0673, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 0.8965447545051575, | |
| "rewards/margins": 6.54379940032959, | |
| "rewards/rejected": -5.647254943847656, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.608695652173913, | |
| "eval_logits/chosen": -2.3496623039245605, | |
| "eval_logits/rejected": -2.3254337310791016, | |
| "eval_logps/chosen": -206.58370971679688, | |
| "eval_logps/rejected": -241.21258544921875, | |
| "eval_loss": 0.455331414937973, | |
| "eval_rewards/accuracies": 0.8205645084381104, | |
| "eval_rewards/chosen": -0.6264265775680542, | |
| "eval_rewards/margins": 3.674025774002075, | |
| "eval_rewards/rejected": -4.300451755523682, | |
| "eval_runtime": 247.544, | |
| "eval_samples_per_second": 15.844, | |
| "eval_steps_per_second": 0.25, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 2.644927536231884, | |
| "grad_norm": 13.189480623721069, | |
| "learning_rate": 3.585983620957112e-08, | |
| "logits/chosen": -2.361184597015381, | |
| "logits/rejected": -2.3199973106384277, | |
| "logps/chosen": -183.58596801757812, | |
| "logps/rejected": -238.3034210205078, | |
| "loss": 0.0632, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": 0.6551353931427002, | |
| "rewards/margins": 6.403916358947754, | |
| "rewards/rejected": -5.748780250549316, | |
| "step": 365 | |
| }, | |
| { | |
| "epoch": 2.681159420289855, | |
| "grad_norm": 10.993397358198846, | |
| "learning_rate": 2.8982868265005454e-08, | |
| "logits/chosen": -2.3595023155212402, | |
| "logits/rejected": -2.3268685340881348, | |
| "logps/chosen": -189.55947875976562, | |
| "logps/rejected": -237.3144073486328, | |
| "loss": 0.0587, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.8602153658866882, | |
| "rewards/margins": 6.659050941467285, | |
| "rewards/rejected": -5.798834800720215, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 2.717391304347826, | |
| "grad_norm": 13.4301289511478, | |
| "learning_rate": 2.2817865927956092e-08, | |
| "logits/chosen": -2.3729586601257324, | |
| "logits/rejected": -2.333218574523926, | |
| "logps/chosen": -182.43051147460938, | |
| "logps/rejected": -235.4917449951172, | |
| "loss": 0.0602, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": 0.953597903251648, | |
| "rewards/margins": 6.269543647766113, | |
| "rewards/rejected": -5.315945148468018, | |
| "step": 375 | |
| }, | |
| { | |
| "epoch": 2.753623188405797, | |
| "grad_norm": 13.712934546586954, | |
| "learning_rate": 1.7374147903282176e-08, | |
| "logits/chosen": -2.352264881134033, | |
| "logits/rejected": -2.317417860031128, | |
| "logps/chosen": -196.2283477783203, | |
| "logps/rejected": -243.51541137695312, | |
| "loss": 0.0483, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.6484888792037964, | |
| "rewards/margins": 6.524939060211182, | |
| "rewards/rejected": -5.876450538635254, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 2.789855072463768, | |
| "grad_norm": 7.887220453233511, | |
| "learning_rate": 1.2659942639057952e-08, | |
| "logits/chosen": -2.3527908325195312, | |
| "logits/rejected": -2.324645757675171, | |
| "logps/chosen": -197.2644500732422, | |
| "logps/rejected": -250.51913452148438, | |
| "loss": 0.0557, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.7034837603569031, | |
| "rewards/margins": 6.565079689025879, | |
| "rewards/rejected": -5.861596584320068, | |
| "step": 385 | |
| }, | |
| { | |
| "epoch": 2.8260869565217392, | |
| "grad_norm": 10.716157650003963, | |
| "learning_rate": 8.682375888868166e-09, | |
| "logits/chosen": -2.352689027786255, | |
| "logits/rejected": -2.3404457569122314, | |
| "logps/chosen": -190.28585815429688, | |
| "logps/rejected": -256.22747802734375, | |
| "loss": 0.0505, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": 0.4607642590999603, | |
| "rewards/margins": 6.127005577087402, | |
| "rewards/rejected": -5.666241645812988, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 2.86231884057971, | |
| "grad_norm": 10.578934340288717, | |
| "learning_rate": 5.447459940880084e-09, | |
| "logits/chosen": -2.3586277961730957, | |
| "logits/rejected": -2.3280324935913086, | |
| "logps/chosen": -183.1387481689453, | |
| "logps/rejected": -248.5103759765625, | |
| "loss": 0.0541, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": 0.48908624053001404, | |
| "rewards/margins": 6.304218292236328, | |
| "rewards/rejected": -5.815131664276123, | |
| "step": 395 | |
| }, | |
| { | |
| "epoch": 2.898550724637681, | |
| "grad_norm": 13.682225413217974, | |
| "learning_rate": 2.9600845299737053e-09, | |
| "logits/chosen": -2.340261936187744, | |
| "logits/rejected": -2.314548969268799, | |
| "logps/chosen": -176.70303344726562, | |
| "logps/rejected": -237.1485595703125, | |
| "loss": 0.0728, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 0.3482286036014557, | |
| "rewards/margins": 6.158910274505615, | |
| "rewards/rejected": -5.810681343078613, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 2.898550724637681, | |
| "eval_logits/chosen": -2.3491718769073486, | |
| "eval_logits/rejected": -2.3246846199035645, | |
| "eval_logps/chosen": -207.1744384765625, | |
| "eval_logps/rejected": -242.15025329589844, | |
| "eval_loss": 0.45203983783721924, | |
| "eval_rewards/accuracies": 0.8185483813285828, | |
| "eval_rewards/chosen": -0.6855015158653259, | |
| "eval_rewards/margins": 3.708717107772827, | |
| "eval_rewards/rejected": -4.394218921661377, | |
| "eval_runtime": 247.4572, | |
| "eval_samples_per_second": 15.849, | |
| "eval_steps_per_second": 0.251, | |
| "step": 400 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 414, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 40, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4716694698196992.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |