| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9990817263544536, | |
| "eval_steps": 100, | |
| "global_step": 204, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 2.3809523809523811e-07, | |
| "logits/chosen": -2.476858377456665, | |
| "logits/rejected": -2.312589406967163, | |
| "logps/chosen": -256.081787109375, | |
| "logps/rejected": -237.47195434570312, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.380952380952381e-06, | |
| "logits/chosen": -2.2722971439361572, | |
| "logits/rejected": -2.247359275817871, | |
| "logps/chosen": -218.2034454345703, | |
| "logps/rejected": -239.2631072998047, | |
| "loss": 0.6892, | |
| "rewards/accuracies": 0.5277777910232544, | |
| "rewards/chosen": 0.0009012501686811447, | |
| "rewards/margins": 0.007593905553221703, | |
| "rewards/rejected": -0.0066926549188792706, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.761904761904762e-06, | |
| "logits/chosen": -1.9851748943328857, | |
| "logits/rejected": -1.7117283344268799, | |
| "logps/chosen": -327.2151794433594, | |
| "logps/rejected": -315.92840576171875, | |
| "loss": 0.6161, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.29775306582450867, | |
| "rewards/margins": 0.21288084983825684, | |
| "rewards/rejected": -0.5106339454650879, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.970219740227693e-06, | |
| "logits/chosen": -1.6041641235351562, | |
| "logits/rejected": -1.3207042217254639, | |
| "logps/chosen": -298.0397644042969, | |
| "logps/rejected": -315.13543701171875, | |
| "loss": 0.5858, | |
| "rewards/accuracies": 0.703125, | |
| "rewards/chosen": -0.27262887358665466, | |
| "rewards/margins": 0.3364975154399872, | |
| "rewards/rejected": -0.6091263294219971, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.868186180746792e-06, | |
| "logits/chosen": -1.2411524057388306, | |
| "logits/rejected": -0.8121637105941772, | |
| "logps/chosen": -354.04315185546875, | |
| "logps/rejected": -402.13787841796875, | |
| "loss": 0.5475, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.5969960689544678, | |
| "rewards/margins": 0.565431535243988, | |
| "rewards/rejected": -1.1624276638031006, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.696530612642871e-06, | |
| "logits/chosen": -1.148498296737671, | |
| "logits/rejected": -0.7622831463813782, | |
| "logps/chosen": -342.1610107421875, | |
| "logps/rejected": -401.2055969238281, | |
| "loss": 0.5384, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.5624964833259583, | |
| "rewards/margins": 0.587742030620575, | |
| "rewards/rejected": -1.1502383947372437, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.460299516441777e-06, | |
| "logits/chosen": -1.0859754085540771, | |
| "logits/rejected": -0.6681124567985535, | |
| "logps/chosen": -328.865234375, | |
| "logps/rejected": -390.8894348144531, | |
| "loss": 0.5481, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": -0.5670114159584045, | |
| "rewards/margins": 0.5836409330368042, | |
| "rewards/rejected": -1.150652289390564, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.1664378205239085e-06, | |
| "logits/chosen": -0.855512797832489, | |
| "logits/rejected": -0.3595692217350006, | |
| "logps/chosen": -353.48675537109375, | |
| "logps/rejected": -402.8636779785156, | |
| "loss": 0.5327, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.8263834714889526, | |
| "rewards/margins": 0.5548129081726074, | |
| "rewards/rejected": -1.38119637966156, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 3.8235847280454626e-06, | |
| "logits/chosen": -1.166483998298645, | |
| "logits/rejected": -0.6420685052871704, | |
| "logps/chosen": -332.43975830078125, | |
| "logps/rejected": -391.0145263671875, | |
| "loss": 0.4927, | |
| "rewards/accuracies": 0.753125011920929, | |
| "rewards/chosen": -0.5521113276481628, | |
| "rewards/margins": 0.6991725564002991, | |
| "rewards/rejected": -1.251283884048462, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.441819734087963e-06, | |
| "logits/chosen": -1.4997479915618896, | |
| "logits/rejected": -0.9499446153640747, | |
| "logps/chosen": -348.1226806640625, | |
| "logps/rejected": -402.1637268066406, | |
| "loss": 0.5287, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.5159438848495483, | |
| "rewards/margins": 0.7751549482345581, | |
| "rewards/rejected": -1.2910988330841064, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 3.0323662998460396e-06, | |
| "logits/chosen": -1.3590983152389526, | |
| "logits/rejected": -0.7285458445549011, | |
| "logps/chosen": -362.32666015625, | |
| "logps/rejected": -408.63885498046875, | |
| "loss": 0.515, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.7177928686141968, | |
| "rewards/margins": 0.6953709125518799, | |
| "rewards/rejected": -1.4131637811660767, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "eval_logits/chosen": -1.2092429399490356, | |
| "eval_logits/rejected": -0.5798258185386658, | |
| "eval_logps/chosen": -365.3294677734375, | |
| "eval_logps/rejected": -413.1092834472656, | |
| "eval_loss": 0.544935405254364, | |
| "eval_rewards/accuracies": 0.7160000205039978, | |
| "eval_rewards/chosen": -0.7798795700073242, | |
| "eval_rewards/margins": 0.681567370891571, | |
| "eval_rewards/rejected": -1.4614468812942505, | |
| "eval_runtime": 384.0193, | |
| "eval_samples_per_second": 5.208, | |
| "eval_steps_per_second": 0.651, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.6072618954988867e-06, | |
| "logits/chosen": -1.1533472537994385, | |
| "logits/rejected": -0.3894789218902588, | |
| "logps/chosen": -369.21466064453125, | |
| "logps/rejected": -427.78204345703125, | |
| "loss": 0.5476, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.81017005443573, | |
| "rewards/margins": 0.6706447601318359, | |
| "rewards/rejected": -1.4808146953582764, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.1790041121336223e-06, | |
| "logits/chosen": -1.2282952070236206, | |
| "logits/rejected": -0.6969107985496521, | |
| "logps/chosen": -362.3310852050781, | |
| "logps/rejected": -415.934814453125, | |
| "loss": 0.5101, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.7746846079826355, | |
| "rewards/margins": 0.6679913401603699, | |
| "rewards/rejected": -1.4426759481430054, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.760183246631777e-06, | |
| "logits/chosen": -1.172527551651001, | |
| "logits/rejected": -0.597676694393158, | |
| "logps/chosen": -354.2320556640625, | |
| "logps/rejected": -422.12841796875, | |
| "loss": 0.5083, | |
| "rewards/accuracies": 0.7406250238418579, | |
| "rewards/chosen": -0.6290267109870911, | |
| "rewards/margins": 0.8170250654220581, | |
| "rewards/rejected": -1.446051836013794, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3631121611097364e-06, | |
| "logits/chosen": -1.188578486442566, | |
| "logits/rejected": -0.816562294960022, | |
| "logps/chosen": -359.2029113769531, | |
| "logps/rejected": -407.4598693847656, | |
| "loss": 0.5267, | |
| "rewards/accuracies": 0.715624988079071, | |
| "rewards/chosen": -0.699051558971405, | |
| "rewards/margins": 0.6616160273551941, | |
| "rewards/rejected": -1.3606675863265991, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 9.994642986290797e-07, | |
| "logits/chosen": -1.323190450668335, | |
| "logits/rejected": -0.7437389492988586, | |
| "logps/chosen": -343.61151123046875, | |
| "logps/rejected": -402.93597412109375, | |
| "loss": 0.5295, | |
| "rewards/accuracies": 0.746874988079071, | |
| "rewards/chosen": -0.6734719276428223, | |
| "rewards/margins": 0.7382540702819824, | |
| "rewards/rejected": -1.4117259979248047, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 6.799304971075383e-07, | |
| "logits/chosen": -1.250248670578003, | |
| "logits/rejected": -0.6498929262161255, | |
| "logps/chosen": -375.78753662109375, | |
| "logps/rejected": -424.0221252441406, | |
| "loss": 0.5198, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.7862330675125122, | |
| "rewards/margins": 0.6770610213279724, | |
| "rewards/rejected": -1.463294267654419, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.1390469071538183e-07, | |
| "logits/chosen": -1.101661205291748, | |
| "logits/rejected": -0.5704807043075562, | |
| "logps/chosen": -360.23687744140625, | |
| "logps/rejected": -435.20220947265625, | |
| "loss": 0.5199, | |
| "rewards/accuracies": 0.765625, | |
| "rewards/chosen": -0.7304506897926331, | |
| "rewards/margins": 0.9151493310928345, | |
| "rewards/rejected": -1.6455999612808228, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.092077387824884e-07, | |
| "logits/chosen": -1.2315480709075928, | |
| "logits/rejected": -0.714310348033905, | |
| "logps/chosen": -371.4505310058594, | |
| "logps/rejected": -436.5375061035156, | |
| "loss": 0.5269, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.7898632287979126, | |
| "rewards/margins": 0.7547814249992371, | |
| "rewards/rejected": -1.5446445941925049, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 7.185750133542168e-08, | |
| "logits/chosen": -1.1423877477645874, | |
| "logits/rejected": -0.7209599018096924, | |
| "logps/chosen": -351.42633056640625, | |
| "logps/rejected": -417.2574768066406, | |
| "loss": 0.5275, | |
| "rewards/accuracies": 0.703125, | |
| "rewards/chosen": -0.8558732867240906, | |
| "rewards/margins": 0.733085572719574, | |
| "rewards/rejected": -1.588958978652954, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 5.891920784984184e-09, | |
| "logits/chosen": -1.2168794870376587, | |
| "logits/rejected": -0.6400166749954224, | |
| "logps/chosen": -360.48065185546875, | |
| "logps/rejected": -428.3067932128906, | |
| "loss": 0.4925, | |
| "rewards/accuracies": 0.778124988079071, | |
| "rewards/chosen": -0.7727445960044861, | |
| "rewards/margins": 0.8154066205024719, | |
| "rewards/rejected": -1.5881513357162476, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "eval_logits/chosen": -1.180118203163147, | |
| "eval_logits/rejected": -0.6068284511566162, | |
| "eval_logps/chosen": -367.6444396972656, | |
| "eval_logps/rejected": -415.7151794433594, | |
| "eval_loss": 0.5366359949111938, | |
| "eval_rewards/accuracies": 0.7319999933242798, | |
| "eval_rewards/chosen": -0.8030293583869934, | |
| "eval_rewards/margins": 0.6844774484634399, | |
| "eval_rewards/rejected": -1.4875068664550781, | |
| "eval_runtime": 384.016, | |
| "eval_samples_per_second": 5.208, | |
| "eval_steps_per_second": 0.651, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 204, | |
| "total_flos": 0.0, | |
| "train_loss": 0.54004861034599, | |
| "train_runtime": 9618.5725, | |
| "train_samples_per_second": 2.717, | |
| "train_steps_per_second": 0.021 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 204, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |