| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984301412872841, | |
| "eval_steps": 100, | |
| "global_step": 477, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0020931449502878076, | |
| "grad_norm": 2.741797685623169, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": -3.1097278594970703, | |
| "logits/rejected": -3.080122470855713, | |
| "logps/chosen": -389.3681640625, | |
| "logps/rejected": -352.00482177734375, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.020931449502878074, | |
| "grad_norm": 2.746737480163574, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -2.9152116775512695, | |
| "logits/rejected": -2.905562400817871, | |
| "logps/chosen": -323.6693420410156, | |
| "logps/rejected": -297.4331359863281, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.4201388955116272, | |
| "rewards/chosen": -0.00030070680077187717, | |
| "rewards/margins": -0.00017917003424372524, | |
| "rewards/rejected": -0.00012153676652815193, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04186289900575615, | |
| "grad_norm": 2.543320655822754, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -2.9843814373016357, | |
| "logits/rejected": -2.9781064987182617, | |
| "logps/chosen": -335.30413818359375, | |
| "logps/rejected": -316.4299621582031, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.5218750238418579, | |
| "rewards/chosen": 0.0004041799402330071, | |
| "rewards/margins": 0.0005227966466918588, | |
| "rewards/rejected": -0.0001186166555271484, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06279434850863422, | |
| "grad_norm": 2.5893585681915283, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.997720241546631, | |
| "logits/rejected": -2.9677913188934326, | |
| "logps/chosen": -345.15240478515625, | |
| "logps/rejected": -295.96966552734375, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.528124988079071, | |
| "rewards/chosen": -0.0004641309496946633, | |
| "rewards/margins": 0.000602933403570205, | |
| "rewards/rejected": -0.0010670643532648683, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.0837257980115123, | |
| "grad_norm": 2.6134979724884033, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.9335551261901855, | |
| "logits/rejected": -2.9302399158477783, | |
| "logps/chosen": -321.5934143066406, | |
| "logps/rejected": -289.6072998046875, | |
| "loss": 0.6919, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.0015825815498828888, | |
| "rewards/margins": 0.0028794072568416595, | |
| "rewards/rejected": -0.004461988341063261, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "grad_norm": 2.6024513244628906, | |
| "learning_rate": 4.999731868769026e-07, | |
| "logits/chosen": -2.9198384284973145, | |
| "logits/rejected": -2.8993093967437744, | |
| "logps/chosen": -328.54644775390625, | |
| "logps/rejected": -309.58087158203125, | |
| "loss": 0.6901, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": -0.005653353873640299, | |
| "rewards/margins": 0.004262409172952175, | |
| "rewards/rejected": -0.009915763512253761, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12558869701726844, | |
| "grad_norm": 2.767859697341919, | |
| "learning_rate": 4.990353313429303e-07, | |
| "logits/chosen": -2.982860565185547, | |
| "logits/rejected": -2.969365119934082, | |
| "logps/chosen": -307.7032470703125, | |
| "logps/rejected": -304.3933410644531, | |
| "loss": 0.6874, | |
| "rewards/accuracies": 0.578125, | |
| "rewards/chosen": -0.015346085652709007, | |
| "rewards/margins": 0.010624411515891552, | |
| "rewards/rejected": -0.025970498099923134, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14652014652014653, | |
| "grad_norm": 2.7651147842407227, | |
| "learning_rate": 4.967625656594781e-07, | |
| "logits/chosen": -2.9431705474853516, | |
| "logits/rejected": -2.953648090362549, | |
| "logps/chosen": -346.7774658203125, | |
| "logps/rejected": -312.93865966796875, | |
| "loss": 0.6826, | |
| "rewards/accuracies": 0.653124988079071, | |
| "rewards/chosen": -0.04352644830942154, | |
| "rewards/margins": 0.024388128891587257, | |
| "rewards/rejected": -0.06791457533836365, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.1674515960230246, | |
| "grad_norm": 2.939892053604126, | |
| "learning_rate": 4.93167072587771e-07, | |
| "logits/chosen": -2.973877429962158, | |
| "logits/rejected": -2.9495885372161865, | |
| "logps/chosen": -362.0614929199219, | |
| "logps/rejected": -301.11322021484375, | |
| "loss": 0.6771, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.08665237575769424, | |
| "rewards/margins": 0.0347968153655529, | |
| "rewards/rejected": -0.12144919484853745, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18838304552590268, | |
| "grad_norm": 2.9581210613250732, | |
| "learning_rate": 4.882681251368548e-07, | |
| "logits/chosen": -2.9634270668029785, | |
| "logits/rejected": -2.9433810710906982, | |
| "logps/chosen": -323.24359130859375, | |
| "logps/rejected": -298.01763916015625, | |
| "loss": 0.6728, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.13328225910663605, | |
| "rewards/margins": 0.03973008692264557, | |
| "rewards/rejected": -0.17301234602928162, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "grad_norm": 3.1116480827331543, | |
| "learning_rate": 4.820919832540181e-07, | |
| "logits/chosen": -2.9381461143493652, | |
| "logits/rejected": -2.9422552585601807, | |
| "logps/chosen": -342.9717102050781, | |
| "logps/rejected": -310.58758544921875, | |
| "loss": 0.6685, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.14851507544517517, | |
| "rewards/margins": 0.057429663836956024, | |
| "rewards/rejected": -0.2059447318315506, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "eval_logits/chosen": -2.950716972351074, | |
| "eval_logits/rejected": -2.9254934787750244, | |
| "eval_logps/chosen": -355.60272216796875, | |
| "eval_logps/rejected": -317.9681701660156, | |
| "eval_loss": 0.6694281101226807, | |
| "eval_rewards/accuracies": 0.6527777910232544, | |
| "eval_rewards/chosen": -0.1298944652080536, | |
| "eval_rewards/margins": 0.06442055851221085, | |
| "eval_rewards/rejected": -0.19431501626968384, | |
| "eval_runtime": 86.0849, | |
| "eval_samples_per_second": 23.233, | |
| "eval_steps_per_second": 0.732, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2302459445316588, | |
| "grad_norm": 3.7738163471221924, | |
| "learning_rate": 4.7467175306295647e-07, | |
| "logits/chosen": -2.926741123199463, | |
| "logits/rejected": -2.913757562637329, | |
| "logps/chosen": -356.9674377441406, | |
| "logps/rejected": -310.5869140625, | |
| "loss": 0.6697, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.14377950131893158, | |
| "rewards/margins": 0.06109999865293503, | |
| "rewards/rejected": -0.204879492521286, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25117739403453687, | |
| "grad_norm": 3.45920467376709, | |
| "learning_rate": 4.6604720940421207e-07, | |
| "logits/chosen": -2.854685068130493, | |
| "logits/rejected": -2.851743459701538, | |
| "logps/chosen": -349.94189453125, | |
| "logps/rejected": -327.46820068359375, | |
| "loss": 0.6557, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.174159973859787, | |
| "rewards/margins": 0.08476099371910095, | |
| "rewards/rejected": -0.25892096757888794, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.272108843537415, | |
| "grad_norm": 4.213115692138672, | |
| "learning_rate": 4.5626458262912735e-07, | |
| "logits/chosen": -2.8723671436309814, | |
| "logits/rejected": -2.8682103157043457, | |
| "logps/chosen": -364.5023498535156, | |
| "logps/rejected": -344.17413330078125, | |
| "loss": 0.65, | |
| "rewards/accuracies": 0.640625, | |
| "rewards/chosen": -0.26356348395347595, | |
| "rewards/margins": 0.08755739033222198, | |
| "rewards/rejected": -0.35112088918685913, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29304029304029305, | |
| "grad_norm": 4.351049900054932, | |
| "learning_rate": 4.453763107901675e-07, | |
| "logits/chosen": -2.8395862579345703, | |
| "logits/rejected": -2.8383350372314453, | |
| "logps/chosen": -386.54754638671875, | |
| "logps/rejected": -353.8211975097656, | |
| "loss": 0.6576, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.30667954683303833, | |
| "rewards/margins": 0.11848233640193939, | |
| "rewards/rejected": -0.4251618981361389, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "grad_norm": 5.293145656585693, | |
| "learning_rate": 4.3344075855595097e-07, | |
| "logits/chosen": -2.837963581085205, | |
| "logits/rejected": -2.8277411460876465, | |
| "logps/chosen": -363.89532470703125, | |
| "logps/rejected": -334.4164733886719, | |
| "loss": 0.6562, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.3559018671512604, | |
| "rewards/margins": 0.0798763707280159, | |
| "rewards/rejected": -0.43577829003334045, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.3349031920460492, | |
| "grad_norm": 5.583098411560059, | |
| "learning_rate": 4.2052190435769554e-07, | |
| "logits/chosen": -2.8631508350372314, | |
| "logits/rejected": -2.8266890048980713, | |
| "logps/chosen": -359.99200439453125, | |
| "logps/rejected": -320.49151611328125, | |
| "loss": 0.6453, | |
| "rewards/accuracies": 0.628125011920929, | |
| "rewards/chosen": -0.34305575489997864, | |
| "rewards/margins": 0.13357409834861755, | |
| "rewards/rejected": -0.4766298234462738, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.35583464154892724, | |
| "grad_norm": 5.565986633300781, | |
| "learning_rate": 4.0668899744407567e-07, | |
| "logits/chosen": -2.800814628601074, | |
| "logits/rejected": -2.7905514240264893, | |
| "logps/chosen": -349.87725830078125, | |
| "logps/rejected": -323.04150390625, | |
| "loss": 0.6516, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.4009243845939636, | |
| "rewards/margins": 0.12320031225681305, | |
| "rewards/rejected": -0.5241247415542603, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.37676609105180536, | |
| "grad_norm": 6.890281677246094, | |
| "learning_rate": 3.920161866827889e-07, | |
| "logits/chosen": -2.8282012939453125, | |
| "logits/rejected": -2.8248977661132812, | |
| "logps/chosen": -382.04364013671875, | |
| "logps/rejected": -350.888427734375, | |
| "loss": 0.64, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.4164988100528717, | |
| "rewards/margins": 0.15513385832309723, | |
| "rewards/rejected": -0.5716326832771301, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3976975405546834, | |
| "grad_norm": 5.489590167999268, | |
| "learning_rate": 3.765821230985757e-07, | |
| "logits/chosen": -2.8130276203155518, | |
| "logits/rejected": -2.7892467975616455, | |
| "logps/chosen": -350.63873291015625, | |
| "logps/rejected": -333.55718994140625, | |
| "loss": 0.6408, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.3725319802761078, | |
| "rewards/margins": 0.13300345838069916, | |
| "rewards/rejected": -0.5055354833602905, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "grad_norm": 6.967275142669678, | |
| "learning_rate": 3.604695382782159e-07, | |
| "logits/chosen": -2.787351608276367, | |
| "logits/rejected": -2.789077043533325, | |
| "logps/chosen": -370.17864990234375, | |
| "logps/rejected": -361.43988037109375, | |
| "loss": 0.642, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.38552746176719666, | |
| "rewards/margins": 0.16969361901283264, | |
| "rewards/rejected": -0.5552210807800293, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "eval_logits/chosen": -2.8207905292510986, | |
| "eval_logits/rejected": -2.7925562858581543, | |
| "eval_logps/chosen": -385.3469543457031, | |
| "eval_logps/rejected": -360.2872619628906, | |
| "eval_loss": 0.6406751871109009, | |
| "eval_rewards/accuracies": 0.6726190447807312, | |
| "eval_rewards/chosen": -0.4273369312286377, | |
| "eval_rewards/margins": 0.19016937911510468, | |
| "eval_rewards/rejected": -0.6175063252449036, | |
| "eval_runtime": 86.2297, | |
| "eval_samples_per_second": 23.194, | |
| "eval_steps_per_second": 0.731, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.43956043956043955, | |
| "grad_norm": 8.497602462768555, | |
| "learning_rate": 3.4376480090239047e-07, | |
| "logits/chosen": -2.7831473350524902, | |
| "logits/rejected": -2.7467095851898193, | |
| "logps/chosen": -387.8429870605469, | |
| "logps/rejected": -351.0397644042969, | |
| "loss": 0.6471, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.5111196041107178, | |
| "rewards/margins": 0.1388254463672638, | |
| "rewards/rejected": -0.649945080280304, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.4604918890633176, | |
| "grad_norm": 6.599562168121338, | |
| "learning_rate": 3.265574537815398e-07, | |
| "logits/chosen": -2.7829642295837402, | |
| "logits/rejected": -2.7764458656311035, | |
| "logps/chosen": -379.44464111328125, | |
| "logps/rejected": -354.76092529296875, | |
| "loss": 0.6423, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.49129724502563477, | |
| "rewards/margins": 0.16036386787891388, | |
| "rewards/rejected": -0.6516611576080322, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48142333856619574, | |
| "grad_norm": 6.966579914093018, | |
| "learning_rate": 3.0893973387735683e-07, | |
| "logits/chosen": -2.7712459564208984, | |
| "logits/rejected": -2.760551929473877, | |
| "logps/chosen": -370.9303283691406, | |
| "logps/rejected": -359.6304626464844, | |
| "loss": 0.6429, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.436328649520874, | |
| "rewards/margins": 0.17412248253822327, | |
| "rewards/rejected": -0.6104511022567749, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5023547880690737, | |
| "grad_norm": 6.429652690887451, | |
| "learning_rate": 2.910060778827554e-07, | |
| "logits/chosen": -2.7925572395324707, | |
| "logits/rejected": -2.7672150135040283, | |
| "logps/chosen": -370.3599548339844, | |
| "logps/rejected": -349.8984680175781, | |
| "loss": 0.6292, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -0.4170507490634918, | |
| "rewards/margins": 0.2068098783493042, | |
| "rewards/rejected": -0.6238606572151184, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "grad_norm": 6.210602283477783, | |
| "learning_rate": 2.7285261601056697e-07, | |
| "logits/chosen": -2.791151285171509, | |
| "logits/rejected": -2.761967897415161, | |
| "logps/chosen": -386.1891784667969, | |
| "logps/rejected": -342.8139343261719, | |
| "loss": 0.6327, | |
| "rewards/accuracies": 0.6781250238418579, | |
| "rewards/chosen": -0.42713984847068787, | |
| "rewards/margins": 0.22475413978099823, | |
| "rewards/rejected": -0.6518939137458801, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.54421768707483, | |
| "grad_norm": 6.485888481140137, | |
| "learning_rate": 2.5457665670441937e-07, | |
| "logits/chosen": -2.8333914279937744, | |
| "logits/rejected": -2.8468918800354004, | |
| "logps/chosen": -381.37078857421875, | |
| "logps/rejected": -365.30963134765625, | |
| "loss": 0.6332, | |
| "rewards/accuracies": 0.684374988079071, | |
| "rewards/chosen": -0.45365315675735474, | |
| "rewards/margins": 0.20745711028575897, | |
| "rewards/rejected": -0.6611102223396301, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.565149136577708, | |
| "grad_norm": 7.725765705108643, | |
| "learning_rate": 2.3627616503391812e-07, | |
| "logits/chosen": -2.8207736015319824, | |
| "logits/rejected": -2.8081822395324707, | |
| "logps/chosen": -386.81976318359375, | |
| "logps/rejected": -361.7131652832031, | |
| "loss": 0.6314, | |
| "rewards/accuracies": 0.659375011920929, | |
| "rewards/chosen": -0.44246116280555725, | |
| "rewards/margins": 0.18473473191261292, | |
| "rewards/rejected": -0.6271958351135254, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5860805860805861, | |
| "grad_norm": 7.315251350402832, | |
| "learning_rate": 2.1804923757009882e-07, | |
| "logits/chosen": -2.7474558353424072, | |
| "logits/rejected": -2.7452392578125, | |
| "logps/chosen": -373.4273681640625, | |
| "logps/rejected": -346.6417541503906, | |
| "loss": 0.6341, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -0.4729071259498596, | |
| "rewards/margins": 0.2166450470685959, | |
| "rewards/rejected": -0.6895521879196167, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6070120355834642, | |
| "grad_norm": 7.481057643890381, | |
| "learning_rate": 1.9999357655598891e-07, | |
| "logits/chosen": -2.8017992973327637, | |
| "logits/rejected": -2.7807507514953613, | |
| "logps/chosen": -372.48126220703125, | |
| "logps/rejected": -341.9537048339844, | |
| "loss": 0.6264, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.5083416700363159, | |
| "rewards/margins": 0.176277756690979, | |
| "rewards/rejected": -0.6846194863319397, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "grad_norm": 8.147472381591797, | |
| "learning_rate": 1.8220596619089573e-07, | |
| "logits/chosen": -2.7440390586853027, | |
| "logits/rejected": -2.717979907989502, | |
| "logps/chosen": -375.93035888671875, | |
| "logps/rejected": -361.93121337890625, | |
| "loss": 0.6285, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.4719668924808502, | |
| "rewards/margins": 0.17040875554084778, | |
| "rewards/rejected": -0.642375648021698, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "eval_logits/chosen": -2.801197052001953, | |
| "eval_logits/rejected": -2.773117780685425, | |
| "eval_logps/chosen": -389.8437805175781, | |
| "eval_logps/rejected": -368.0482177734375, | |
| "eval_loss": 0.6331161856651306, | |
| "eval_rewards/accuracies": 0.6646825671195984, | |
| "eval_rewards/chosen": -0.4723050594329834, | |
| "eval_rewards/margins": 0.22281017899513245, | |
| "eval_rewards/rejected": -0.6951152086257935, | |
| "eval_runtime": 85.9334, | |
| "eval_samples_per_second": 23.274, | |
| "eval_steps_per_second": 0.733, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6488749345892203, | |
| "grad_norm": 11.20168399810791, | |
| "learning_rate": 1.647817538357072e-07, | |
| "logits/chosen": -2.736398935317993, | |
| "logits/rejected": -2.738858461380005, | |
| "logps/chosen": -391.0700378417969, | |
| "logps/rejected": -347.84356689453125, | |
| "loss": 0.622, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.4913768768310547, | |
| "rewards/margins": 0.23215806484222412, | |
| "rewards/rejected": -0.723534882068634, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6698063840920984, | |
| "grad_norm": 7.335338592529297, | |
| "learning_rate": 1.478143389201113e-07, | |
| "logits/chosen": -2.7694268226623535, | |
| "logits/rejected": -2.74495267868042, | |
| "logps/chosen": -382.05419921875, | |
| "logps/rejected": -351.9586486816406, | |
| "loss": 0.6179, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.5782755017280579, | |
| "rewards/margins": 0.24550530314445496, | |
| "rewards/rejected": -0.8237808346748352, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6907378335949764, | |
| "grad_norm": 9.174397468566895, | |
| "learning_rate": 1.3139467229135998e-07, | |
| "logits/chosen": -2.7542152404785156, | |
| "logits/rejected": -2.74013352394104, | |
| "logps/chosen": -391.40618896484375, | |
| "logps/rejected": -386.9635314941406, | |
| "loss": 0.63, | |
| "rewards/accuracies": 0.609375, | |
| "rewards/chosen": -0.6302947998046875, | |
| "rewards/margins": 0.17986619472503662, | |
| "rewards/rejected": -0.8101609945297241, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7116692830978545, | |
| "grad_norm": 8.58338451385498, | |
| "learning_rate": 1.1561076868822755e-07, | |
| "logits/chosen": -2.75496768951416, | |
| "logits/rejected": -2.7243494987487793, | |
| "logps/chosen": -427.29315185546875, | |
| "logps/rejected": -422.3783264160156, | |
| "loss": 0.6206, | |
| "rewards/accuracies": 0.659375011920929, | |
| "rewards/chosen": -0.6146808862686157, | |
| "rewards/margins": 0.28181296586990356, | |
| "rewards/rejected": -0.8964937925338745, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "grad_norm": 7.745279788970947, | |
| "learning_rate": 1.0054723495346482e-07, | |
| "logits/chosen": -2.7472786903381348, | |
| "logits/rejected": -2.727072238922119, | |
| "logps/chosen": -380.8338623046875, | |
| "logps/rejected": -356.40875244140625, | |
| "loss": 0.6116, | |
| "rewards/accuracies": 0.6656249761581421, | |
| "rewards/chosen": -0.6483052968978882, | |
| "rewards/margins": 0.22691123187541962, | |
| "rewards/rejected": -0.8752166628837585, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7535321821036107, | |
| "grad_norm": 8.188222885131836, | |
| "learning_rate": 8.628481651367875e-08, | |
| "logits/chosen": -2.7281575202941895, | |
| "logits/rejected": -2.714717388153076, | |
| "logps/chosen": -399.79217529296875, | |
| "logps/rejected": -374.2989807128906, | |
| "loss": 0.6313, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.7062140107154846, | |
| "rewards/margins": 0.22601504623889923, | |
| "rewards/rejected": -0.9322290420532227, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7744636316064888, | |
| "grad_norm": 7.2593889236450195, | |
| "learning_rate": 7.289996455765748e-08, | |
| "logits/chosen": -2.753688097000122, | |
| "logits/rejected": -2.7259750366210938, | |
| "logps/chosen": -383.2185974121094, | |
| "logps/rejected": -367.9438781738281, | |
| "loss": 0.6152, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.7225135564804077, | |
| "rewards/margins": 0.21898405253887177, | |
| "rewards/rejected": -0.9414976239204407, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7953950811093669, | |
| "grad_norm": 7.146636009216309, | |
| "learning_rate": 6.046442623320145e-08, | |
| "logits/chosen": -2.6859023571014404, | |
| "logits/rejected": -2.695509433746338, | |
| "logps/chosen": -363.49493408203125, | |
| "logps/rejected": -377.5242004394531, | |
| "loss": 0.6159, | |
| "rewards/accuracies": 0.6656249761581421, | |
| "rewards/chosen": -0.7377739548683167, | |
| "rewards/margins": 0.25747013092041016, | |
| "rewards/rejected": -0.995244026184082, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8163265306122449, | |
| "grad_norm": 7.638937473297119, | |
| "learning_rate": 4.904486005914027e-08, | |
| "logits/chosen": -2.6865615844726562, | |
| "logits/rejected": -2.6732540130615234, | |
| "logps/chosen": -433.0126037597656, | |
| "logps/rejected": -432.28448486328125, | |
| "loss": 0.6203, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.6989966034889221, | |
| "rewards/margins": 0.23467639088630676, | |
| "rewards/rejected": -0.9336729049682617, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "grad_norm": 8.645991325378418, | |
| "learning_rate": 3.8702478614051345e-08, | |
| "logits/chosen": -2.6967921257019043, | |
| "logits/rejected": -2.6760878562927246, | |
| "logps/chosen": -396.04595947265625, | |
| "logps/rejected": -385.61810302734375, | |
| "loss": 0.6222, | |
| "rewards/accuracies": 0.6781250238418579, | |
| "rewards/chosen": -0.674105167388916, | |
| "rewards/margins": 0.25409311056137085, | |
| "rewards/rejected": -0.9281982183456421, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "eval_logits/chosen": -2.7399401664733887, | |
| "eval_logits/rejected": -2.7107229232788086, | |
| "eval_logps/chosen": -412.686767578125, | |
| "eval_logps/rejected": -395.38079833984375, | |
| "eval_loss": 0.6240472197532654, | |
| "eval_rewards/accuracies": 0.682539701461792, | |
| "eval_rewards/chosen": -0.7007347345352173, | |
| "eval_rewards/margins": 0.2677067816257477, | |
| "eval_rewards/rejected": -0.9684414863586426, | |
| "eval_runtime": 86.0349, | |
| "eval_samples_per_second": 23.246, | |
| "eval_steps_per_second": 0.732, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.858189429618001, | |
| "grad_norm": 8.63017749786377, | |
| "learning_rate": 2.9492720416985e-08, | |
| "logits/chosen": -2.72432279586792, | |
| "logits/rejected": -2.709479808807373, | |
| "logps/chosen": -423.2232971191406, | |
| "logps/rejected": -398.17779541015625, | |
| "loss": 0.6347, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.6851434111595154, | |
| "rewards/margins": 0.2306232750415802, | |
| "rewards/rejected": -0.915766716003418, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8791208791208791, | |
| "grad_norm": 8.239731788635254, | |
| "learning_rate": 2.1464952759020856e-08, | |
| "logits/chosen": -2.6693649291992188, | |
| "logits/rejected": -2.672886371612549, | |
| "logps/chosen": -392.8204040527344, | |
| "logps/rejected": -393.4344177246094, | |
| "loss": 0.6118, | |
| "rewards/accuracies": 0.640625, | |
| "rewards/chosen": -0.7332924008369446, | |
| "rewards/margins": 0.24412448704242706, | |
| "rewards/rejected": -0.9774168133735657, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9000523286237572, | |
| "grad_norm": 8.310417175292969, | |
| "learning_rate": 1.4662207078575684e-08, | |
| "logits/chosen": -2.6678881645202637, | |
| "logits/rejected": -2.6358609199523926, | |
| "logps/chosen": -409.0691833496094, | |
| "logps/rejected": -372.57183837890625, | |
| "loss": 0.6136, | |
| "rewards/accuracies": 0.653124988079071, | |
| "rewards/chosen": -0.7174139618873596, | |
| "rewards/margins": 0.21988801658153534, | |
| "rewards/rejected": -0.9373019933700562, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9209837781266352, | |
| "grad_norm": 13.682318687438965, | |
| "learning_rate": 9.12094829893642e-09, | |
| "logits/chosen": -2.718200445175171, | |
| "logits/rejected": -2.689492702484131, | |
| "logps/chosen": -413.35528564453125, | |
| "logps/rejected": -406.2598876953125, | |
| "loss": 0.6068, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.7059027552604675, | |
| "rewards/margins": 0.23966805636882782, | |
| "rewards/rejected": -0.9455708265304565, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "grad_norm": 8.156304359436035, | |
| "learning_rate": 4.8708793644441086e-09, | |
| "logits/chosen": -2.63850736618042, | |
| "logits/rejected": -2.6505093574523926, | |
| "logps/chosen": -413.0054626464844, | |
| "logps/rejected": -420.00018310546875, | |
| "loss": 0.6209, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -0.7106130719184875, | |
| "rewards/margins": 0.24646687507629395, | |
| "rewards/rejected": -0.957080066204071, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9628466771323915, | |
| "grad_norm": 10.356741905212402, | |
| "learning_rate": 1.9347820230782295e-09, | |
| "logits/chosen": -2.698603630065918, | |
| "logits/rejected": -2.671870231628418, | |
| "logps/chosen": -374.42828369140625, | |
| "logps/rejected": -359.28692626953125, | |
| "loss": 0.6226, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.7496173977851868, | |
| "rewards/margins": 0.20760098099708557, | |
| "rewards/rejected": -0.95721834897995, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.9837781266352695, | |
| "grad_norm": 9.971595764160156, | |
| "learning_rate": 3.2839470889836627e-10, | |
| "logits/chosen": -2.7027974128723145, | |
| "logits/rejected": -2.684480905532837, | |
| "logps/chosen": -412.5672302246094, | |
| "logps/rejected": -383.63372802734375, | |
| "loss": 0.6113, | |
| "rewards/accuracies": 0.640625, | |
| "rewards/chosen": -0.7107462882995605, | |
| "rewards/margins": 0.1883649080991745, | |
| "rewards/rejected": -0.8991111516952515, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.9984301412872841, | |
| "step": 477, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6428392008415558, | |
| "train_runtime": 6631.0546, | |
| "train_samples_per_second": 9.219, | |
| "train_steps_per_second": 0.072 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 477, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |