| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.0010206338135982445, | |
| "eval_steps": 500, | |
| "global_step": 30, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 3.4021127119941484e-05, | |
| "grad_norm": 33.091712951660156, | |
| "learning_rate": 1.0000000000000002e-06, | |
| "logits/chosen": -3.133824586868286, | |
| "logits/rejected": -3.1356313228607178, | |
| "logps/chosen": -317.8847961425781, | |
| "logps/rejected": -306.8866271972656, | |
| "loss": 0.8544, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.10576057434082031, | |
| "rewards/margins": 0.2207910716533661, | |
| "rewards/rejected": -0.3265516459941864, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 6.804225423988297e-05, | |
| "grad_norm": 38.49541473388672, | |
| "learning_rate": 2.0000000000000003e-06, | |
| "logits/chosen": -3.1433358192443848, | |
| "logits/rejected": -3.1124918460845947, | |
| "logps/chosen": -212.93820190429688, | |
| "logps/rejected": -198.22511291503906, | |
| "loss": 1.02, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.28638574481010437, | |
| "rewards/margins": -0.37813109159469604, | |
| "rewards/rejected": 0.0917452871799469, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.00010206338135982445, | |
| "grad_norm": 27.203516006469727, | |
| "learning_rate": 3e-06, | |
| "logits/chosen": -3.2104451656341553, | |
| "logits/rejected": -3.2103281021118164, | |
| "logps/chosen": -235.89254760742188, | |
| "logps/rejected": -206.0758514404297, | |
| "loss": 0.605, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.40017586946487427, | |
| "rewards/margins": 0.24887371063232422, | |
| "rewards/rejected": 0.15130215883255005, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.00013608450847976594, | |
| "grad_norm": 26.42736053466797, | |
| "learning_rate": 4.000000000000001e-06, | |
| "logits/chosen": -3.0741634368896484, | |
| "logits/rejected": -2.953117847442627, | |
| "logps/chosen": -428.7364196777344, | |
| "logps/rejected": -192.357666015625, | |
| "loss": 0.4914, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.4332427978515625, | |
| "rewards/margins": 0.5804362893104553, | |
| "rewards/rejected": -0.1471935510635376, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.00017010563559970743, | |
| "grad_norm": 35.623634338378906, | |
| "learning_rate": 5e-06, | |
| "logits/chosen": -2.985285758972168, | |
| "logits/rejected": -2.8854246139526367, | |
| "logps/chosen": -342.38604736328125, | |
| "logps/rejected": -211.78237915039062, | |
| "loss": 0.8273, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.3475755751132965, | |
| "rewards/margins": -0.19371166825294495, | |
| "rewards/rejected": -0.15386392176151276, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0002041267627196489, | |
| "grad_norm": 31.70409393310547, | |
| "learning_rate": 6e-06, | |
| "logits/chosen": -3.074589967727661, | |
| "logits/rejected": -2.85581636428833, | |
| "logps/chosen": -593.6435546875, | |
| "logps/rejected": -259.5284729003906, | |
| "loss": 0.5473, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.2495414763689041, | |
| "rewards/margins": 0.35888367891311646, | |
| "rewards/rejected": -0.10934218764305115, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.00023814788983959038, | |
| "grad_norm": 32.41214370727539, | |
| "learning_rate": 7.000000000000001e-06, | |
| "logits/chosen": -2.976675033569336, | |
| "logits/rejected": -2.978133201599121, | |
| "logps/chosen": -218.11688232421875, | |
| "logps/rejected": -231.7698974609375, | |
| "loss": 0.7124, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.08671779930591583, | |
| "rewards/margins": 0.21950224041938782, | |
| "rewards/rejected": -0.30622005462646484, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.00027216901695953187, | |
| "grad_norm": 27.637351989746094, | |
| "learning_rate": 8.000000000000001e-06, | |
| "logits/chosen": -2.997600793838501, | |
| "logits/rejected": -3.1078853607177734, | |
| "logps/chosen": -165.71444702148438, | |
| "logps/rejected": -335.5938720703125, | |
| "loss": 0.6085, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.057393088936805725, | |
| "rewards/margins": 0.34109365940093994, | |
| "rewards/rejected": -0.3984867036342621, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.00030619014407947336, | |
| "grad_norm": 32.246559143066406, | |
| "learning_rate": 9e-06, | |
| "logits/chosen": -3.0760390758514404, | |
| "logits/rejected": -2.938197135925293, | |
| "logps/chosen": -463.1333312988281, | |
| "logps/rejected": -173.99951171875, | |
| "loss": 0.6398, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.5581258535385132, | |
| "rewards/margins": 0.7653559446334839, | |
| "rewards/rejected": -0.20723000168800354, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.00034021127119941485, | |
| "grad_norm": 39.84626388549805, | |
| "learning_rate": 1e-05, | |
| "logits/chosen": -3.192596673965454, | |
| "logits/rejected": -2.9751501083374023, | |
| "logps/chosen": -473.14532470703125, | |
| "logps/rejected": -110.90155029296875, | |
| "loss": 0.9794, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.3494076132774353, | |
| "rewards/margins": -0.31546899676322937, | |
| "rewards/rejected": -0.03393859416246414, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.00037423239831935634, | |
| "grad_norm": 39.444889068603516, | |
| "learning_rate": 1.1000000000000001e-05, | |
| "logits/chosen": -3.096883535385132, | |
| "logits/rejected": -3.043616771697998, | |
| "logps/chosen": -391.7703552246094, | |
| "logps/rejected": -271.8331604003906, | |
| "loss": 0.8029, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.6240158081054688, | |
| "rewards/margins": 0.4651091992855072, | |
| "rewards/rejected": 0.15890654921531677, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.0004082535254392978, | |
| "grad_norm": 40.78356170654297, | |
| "learning_rate": 1.2e-05, | |
| "logits/chosen": -3.000021457672119, | |
| "logits/rejected": -3.0893115997314453, | |
| "logps/chosen": -255.94607543945312, | |
| "logps/rejected": -405.3093566894531, | |
| "loss": 0.6297, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.33697396516799927, | |
| "rewards/margins": 0.3315456509590149, | |
| "rewards/rejected": 0.005428321659564972, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.00044227465255923927, | |
| "grad_norm": 35.69674301147461, | |
| "learning_rate": 1.3000000000000001e-05, | |
| "logits/chosen": -3.011014699935913, | |
| "logits/rejected": -3.0214414596557617, | |
| "logps/chosen": -223.4480438232422, | |
| "logps/rejected": -261.62469482421875, | |
| "loss": 0.815, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.7416718006134033, | |
| "rewards/margins": 0.19626998901367188, | |
| "rewards/rejected": 0.5454018115997314, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.00047629577967918076, | |
| "grad_norm": 32.88405990600586, | |
| "learning_rate": 1.4000000000000001e-05, | |
| "logits/chosen": -3.065549373626709, | |
| "logits/rejected": -3.1094508171081543, | |
| "logps/chosen": -169.08047485351562, | |
| "logps/rejected": -227.85162353515625, | |
| "loss": 0.8115, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": 0.37659722566604614, | |
| "rewards/margins": -0.11390496790409088, | |
| "rewards/rejected": 0.4905022084712982, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.0005103169067991223, | |
| "grad_norm": 20.918012619018555, | |
| "learning_rate": 1.5e-05, | |
| "logits/chosen": -3.045846462249756, | |
| "logits/rejected": -2.9114990234375, | |
| "logps/chosen": -573.1253662109375, | |
| "logps/rejected": -273.9326477050781, | |
| "loss": 0.3105, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.7755428552627563, | |
| "rewards/margins": 1.082038164138794, | |
| "rewards/rejected": -0.3064952790737152, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.0005443380339190637, | |
| "grad_norm": 38.434749603271484, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "logits/chosen": -3.1250810623168945, | |
| "logits/rejected": -2.96527099609375, | |
| "logps/chosen": -396.9795837402344, | |
| "logps/rejected": -174.92918395996094, | |
| "loss": 0.9971, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.2244594693183899, | |
| "rewards/margins": -0.37520748376846313, | |
| "rewards/rejected": 0.5996670126914978, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.0005783591610390052, | |
| "grad_norm": 37.71885681152344, | |
| "learning_rate": 1.7000000000000003e-05, | |
| "logits/chosen": -2.970050096511841, | |
| "logits/rejected": -3.0346829891204834, | |
| "logps/chosen": -236.67071533203125, | |
| "logps/rejected": -403.0212097167969, | |
| "loss": 1.0307, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.9259212613105774, | |
| "rewards/margins": -0.23367927968502045, | |
| "rewards/rejected": 1.1596004962921143, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.0006123802881589467, | |
| "grad_norm": 25.964290618896484, | |
| "learning_rate": 1.8e-05, | |
| "logits/chosen": -3.0690722465515137, | |
| "logits/rejected": -3.0348947048187256, | |
| "logps/chosen": -312.81317138671875, | |
| "logps/rejected": -259.230712890625, | |
| "loss": 0.6989, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.6348915100097656, | |
| "rewards/margins": 0.6418477892875671, | |
| "rewards/rejected": -0.006956290453672409, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.0006464014152788882, | |
| "grad_norm": 27.076108932495117, | |
| "learning_rate": 1.9e-05, | |
| "logits/chosen": -3.135136842727661, | |
| "logits/rejected": -2.911625862121582, | |
| "logps/chosen": -473.7462158203125, | |
| "logps/rejected": -141.9491729736328, | |
| "loss": 0.6027, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.9424793720245361, | |
| "rewards/margins": 0.9380611181259155, | |
| "rewards/rejected": 0.00441819429397583, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.0006804225423988297, | |
| "grad_norm": 51.820159912109375, | |
| "learning_rate": 2e-05, | |
| "logits/chosen": -2.8403027057647705, | |
| "logits/rejected": -3.0664803981781006, | |
| "logps/chosen": -219.79153442382812, | |
| "logps/rejected": -231.70004272460938, | |
| "loss": 1.1022, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.052263081073760986, | |
| "rewards/margins": -0.6957410573959351, | |
| "rewards/rejected": 0.7480041980743408, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.0007144436695187712, | |
| "grad_norm": 45.9146614074707, | |
| "learning_rate": 2.1e-05, | |
| "logits/chosen": -2.9764764308929443, | |
| "logits/rejected": -2.9391565322875977, | |
| "logps/chosen": -458.0147705078125, | |
| "logps/rejected": -218.37648010253906, | |
| "loss": 1.0772, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.8049196004867554, | |
| "rewards/margins": -0.448437362909317, | |
| "rewards/rejected": 1.2533570528030396, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.0007484647966387127, | |
| "grad_norm": 22.854082107543945, | |
| "learning_rate": 2.2000000000000003e-05, | |
| "logits/chosen": -3.1029181480407715, | |
| "logits/rejected": -3.044419527053833, | |
| "logps/chosen": -154.8139190673828, | |
| "logps/rejected": -155.980712890625, | |
| "loss": 0.4065, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.45841503143310547, | |
| "rewards/margins": 0.8753025531768799, | |
| "rewards/rejected": -0.41688746213912964, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.0007824859237586541, | |
| "grad_norm": 43.94650650024414, | |
| "learning_rate": 2.3000000000000003e-05, | |
| "logits/chosen": -2.9984076023101807, | |
| "logits/rejected": -3.114562511444092, | |
| "logps/chosen": -263.8316650390625, | |
| "logps/rejected": -340.1427307128906, | |
| "loss": 1.0427, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": 0.6933631896972656, | |
| "rewards/margins": -0.14637237787246704, | |
| "rewards/rejected": 0.8397356271743774, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.0008165070508785956, | |
| "grad_norm": 49.11759948730469, | |
| "learning_rate": 2.4e-05, | |
| "logits/chosen": -3.025609016418457, | |
| "logits/rejected": -3.0236239433288574, | |
| "logps/chosen": -412.2383117675781, | |
| "logps/rejected": -643.138427734375, | |
| "loss": 1.0374, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 1.1175220012664795, | |
| "rewards/margins": -0.31503716111183167, | |
| "rewards/rejected": 1.4325592517852783, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.000850528177998537, | |
| "grad_norm": 37.57383728027344, | |
| "learning_rate": 2.5e-05, | |
| "logits/chosen": -3.024618625640869, | |
| "logits/rejected": -3.123701572418213, | |
| "logps/chosen": -201.70530700683594, | |
| "logps/rejected": -224.99038696289062, | |
| "loss": 0.7477, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": 0.9206753373146057, | |
| "rewards/margins": 0.5947309732437134, | |
| "rewards/rejected": 0.32594436407089233, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.0008845493051184785, | |
| "grad_norm": 40.297794342041016, | |
| "learning_rate": 2.6000000000000002e-05, | |
| "logits/chosen": -3.1100215911865234, | |
| "logits/rejected": -3.0195024013519287, | |
| "logps/chosen": -226.2320098876953, | |
| "logps/rejected": -193.0676727294922, | |
| "loss": 0.8151, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.9122087359428406, | |
| "rewards/margins": 0.0203973650932312, | |
| "rewards/rejected": 0.8918113708496094, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.00091857043223842, | |
| "grad_norm": 43.394229888916016, | |
| "learning_rate": 2.7000000000000002e-05, | |
| "logits/chosen": -3.0187630653381348, | |
| "logits/rejected": -3.053678035736084, | |
| "logps/chosen": -153.02883911132812, | |
| "logps/rejected": -260.0540771484375, | |
| "loss": 1.3457, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.7145153284072876, | |
| "rewards/margins": -0.806611180305481, | |
| "rewards/rejected": 1.5211265087127686, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.0009525915593583615, | |
| "grad_norm": 28.59680938720703, | |
| "learning_rate": 2.8000000000000003e-05, | |
| "logits/chosen": -3.032409191131592, | |
| "logits/rejected": -3.0502161979675293, | |
| "logps/chosen": -240.1644287109375, | |
| "logps/rejected": -245.81436157226562, | |
| "loss": 0.5501, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 2.0826761722564697, | |
| "rewards/margins": 0.3433869779109955, | |
| "rewards/rejected": 1.7392891645431519, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.000986612686478303, | |
| "grad_norm": 37.069583892822266, | |
| "learning_rate": 2.9e-05, | |
| "logits/chosen": -3.0516104698181152, | |
| "logits/rejected": -3.144774913787842, | |
| "logps/chosen": -296.03125, | |
| "logps/rejected": -292.9052429199219, | |
| "loss": 1.1087, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.8324771523475647, | |
| "rewards/margins": -0.34529268741607666, | |
| "rewards/rejected": 1.1777698993682861, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.0010206338135982445, | |
| "grad_norm": 30.238658905029297, | |
| "learning_rate": 3e-05, | |
| "logits/chosen": -3.064671039581299, | |
| "logits/rejected": -2.8963613510131836, | |
| "logps/chosen": -415.6085510253906, | |
| "logps/rejected": -149.9669952392578, | |
| "loss": 0.4892, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 1.0249770879745483, | |
| "rewards/margins": 0.7549355030059814, | |
| "rewards/rejected": 0.27004164457321167, | |
| "step": 30 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 500, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 10, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |