| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.2461059190031152, | |
| "eval_steps": 50, | |
| "global_step": 100, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.06230529595015576, | |
| "grad_norm": 61.7364860897157, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": -2.726062059402466, | |
| "logits/rejected": -2.7418746948242188, | |
| "logps/chosen": -264.51171875, | |
| "logps/rejected": -202.7697296142578, | |
| "loss": 0.6899, | |
| "rewards/accuracies": 0.38749998807907104, | |
| "rewards/chosen": 0.020583948120474815, | |
| "rewards/margins": 0.009055671282112598, | |
| "rewards/rejected": 0.011528277769684792, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.12461059190031153, | |
| "grad_norm": 49.11239347044956, | |
| "learning_rate": 1e-06, | |
| "logits/chosen": -2.6860156059265137, | |
| "logits/rejected": -2.666875123977661, | |
| "logps/chosen": -245.5350799560547, | |
| "logps/rejected": -233.3320770263672, | |
| "loss": 0.6338, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": 0.5292676687240601, | |
| "rewards/margins": 0.1557035893201828, | |
| "rewards/rejected": 0.37356406450271606, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.18691588785046728, | |
| "grad_norm": 42.978157224616126, | |
| "learning_rate": 9.988343845952696e-07, | |
| "logits/chosen": -2.480203151702881, | |
| "logits/rejected": -2.471378803253174, | |
| "logps/chosen": -235.6896514892578, | |
| "logps/rejected": -183.0069122314453, | |
| "loss": 0.6089, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 1.5770940780639648, | |
| "rewards/margins": 0.9468550682067871, | |
| "rewards/rejected": 0.6302390098571777, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.24922118380062305, | |
| "grad_norm": 44.40442856187882, | |
| "learning_rate": 9.953429730181652e-07, | |
| "logits/chosen": -2.3026254177093506, | |
| "logits/rejected": -2.2847046852111816, | |
| "logps/chosen": -223.083251953125, | |
| "logps/rejected": -198.93948364257812, | |
| "loss": 0.667, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 1.5826599597930908, | |
| "rewards/margins": 1.0483930110931396, | |
| "rewards/rejected": 0.5342669486999512, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.3115264797507788, | |
| "grad_norm": 41.20960146376919, | |
| "learning_rate": 9.895420438411615e-07, | |
| "logits/chosen": -2.3145434856414795, | |
| "logits/rejected": -2.3113181591033936, | |
| "logps/chosen": -252.1824951171875, | |
| "logps/rejected": -197.47164916992188, | |
| "loss": 0.6233, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 1.545325756072998, | |
| "rewards/margins": 0.8761142492294312, | |
| "rewards/rejected": 0.6692114472389221, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.37383177570093457, | |
| "grad_norm": 38.93893290936411, | |
| "learning_rate": 9.814586436738997e-07, | |
| "logits/chosen": -2.3786375522613525, | |
| "logits/rejected": -2.3869917392730713, | |
| "logps/chosen": -237.798828125, | |
| "logps/rejected": -212.68075561523438, | |
| "loss": 0.6269, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 1.5185164213180542, | |
| "rewards/margins": 1.2036529779434204, | |
| "rewards/rejected": 0.3148634433746338, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.43613707165109034, | |
| "grad_norm": 39.19127764379894, | |
| "learning_rate": 9.711304610594102e-07, | |
| "logits/chosen": -2.440690517425537, | |
| "logits/rejected": -2.4353909492492676, | |
| "logps/chosen": -243.35546875, | |
| "logps/rejected": -212.8701629638672, | |
| "loss": 0.5609, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.8680373430252075, | |
| "rewards/margins": 1.0491241216659546, | |
| "rewards/rejected": -0.1810867041349411, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.4984423676012461, | |
| "grad_norm": 37.39468496881122, | |
| "learning_rate": 9.586056507527264e-07, | |
| "logits/chosen": -2.4579291343688965, | |
| "logits/rejected": -2.4394490718841553, | |
| "logps/chosen": -245.80770874023438, | |
| "logps/rejected": -226.79977416992188, | |
| "loss": 0.5379, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.9454948306083679, | |
| "rewards/margins": 1.2604901790618896, | |
| "rewards/rejected": -0.31499534845352173, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.5607476635514018, | |
| "grad_norm": 36.57554162682888, | |
| "learning_rate": 9.439426092011875e-07, | |
| "logits/chosen": -2.4525513648986816, | |
| "logits/rejected": -2.426307201385498, | |
| "logps/chosen": -268.70843505859375, | |
| "logps/rejected": -200.7107391357422, | |
| "loss": 0.5435, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.9689223170280457, | |
| "rewards/margins": 1.6185741424560547, | |
| "rewards/rejected": -0.6496520042419434, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.6230529595015576, | |
| "grad_norm": 35.85551989361158, | |
| "learning_rate": 9.272097022732443e-07, | |
| "logits/chosen": -2.3485758304595947, | |
| "logits/rejected": -2.3458869457244873, | |
| "logps/chosen": -254.47097778320312, | |
| "logps/rejected": -219.59909057617188, | |
| "loss": 0.5367, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 1.0948731899261475, | |
| "rewards/margins": 1.4575660228729248, | |
| "rewards/rejected": -0.3626927435398102, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.6230529595015576, | |
| "eval_logits/chosen": -2.3203799724578857, | |
| "eval_logits/rejected": -2.3100695610046387, | |
| "eval_logps/chosen": -242.62884521484375, | |
| "eval_logps/rejected": -204.77874755859375, | |
| "eval_loss": 0.5754929780960083, | |
| "eval_rewards/accuracies": 0.7604166865348816, | |
| "eval_rewards/chosen": 1.1959108114242554, | |
| "eval_rewards/margins": 1.3574522733688354, | |
| "eval_rewards/rejected": -0.16154144704341888, | |
| "eval_runtime": 153.3612, | |
| "eval_samples_per_second": 14.867, | |
| "eval_steps_per_second": 0.235, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.6853582554517134, | |
| "grad_norm": 38.90911338348836, | |
| "learning_rate": 9.084849465052209e-07, | |
| "logits/chosen": -2.343371629714966, | |
| "logits/rejected": -2.3140501976013184, | |
| "logps/chosen": -232.2624969482422, | |
| "logps/rejected": -216.8009033203125, | |
| "loss": 0.5809, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 1.4138634204864502, | |
| "rewards/margins": 1.6005926132202148, | |
| "rewards/rejected": -0.18672919273376465, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.7476635514018691, | |
| "grad_norm": 47.023889326268446, | |
| "learning_rate": 8.878556453522099e-07, | |
| "logits/chosen": -2.3421199321746826, | |
| "logits/rejected": -2.3611013889312744, | |
| "logps/chosen": -240.03848266601562, | |
| "logps/rejected": -183.4457244873047, | |
| "loss": 0.5443, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": 1.0544302463531494, | |
| "rewards/margins": 1.1835196018218994, | |
| "rewards/rejected": -0.12908944487571716, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.8099688473520249, | |
| "grad_norm": 34.095725802748696, | |
| "learning_rate": 8.654179821390621e-07, | |
| "logits/chosen": -2.3532650470733643, | |
| "logits/rejected": -2.328545570373535, | |
| "logps/chosen": -241.86349487304688, | |
| "logps/rejected": -211.8879852294922, | |
| "loss": 0.5031, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.9804493188858032, | |
| "rewards/margins": 1.4752452373504639, | |
| "rewards/rejected": -0.4947957992553711, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.8722741433021807, | |
| "grad_norm": 38.629908145531026, | |
| "learning_rate": 8.41276571609327e-07, | |
| "logits/chosen": -2.37026309967041, | |
| "logits/rejected": -2.341160535812378, | |
| "logps/chosen": -264.19927978515625, | |
| "logps/rejected": -228.86752319335938, | |
| "loss": 0.5605, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.7270997762680054, | |
| "rewards/margins": 1.4355695247650146, | |
| "rewards/rejected": -0.708469808101654, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.9345794392523364, | |
| "grad_norm": 30.048777436008372, | |
| "learning_rate": 8.155439721630264e-07, | |
| "logits/chosen": -2.3301854133605957, | |
| "logits/rejected": -2.333800792694092, | |
| "logps/chosen": -232.375732421875, | |
| "logps/rejected": -189.99459838867188, | |
| "loss": 0.4699, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": 0.896081268787384, | |
| "rewards/margins": 1.9089847803115845, | |
| "rewards/rejected": -1.0129035711288452, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.9968847352024922, | |
| "grad_norm": 33.228661608807776, | |
| "learning_rate": 7.883401610574336e-07, | |
| "logits/chosen": -2.417353391647339, | |
| "logits/rejected": -2.390876054763794, | |
| "logps/chosen": -240.1530303955078, | |
| "logps/rejected": -214.0542449951172, | |
| "loss": 0.4852, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.922646701335907, | |
| "rewards/margins": 1.8798067569732666, | |
| "rewards/rejected": -0.9571601748466492, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.0591900311526479, | |
| "grad_norm": 18.48400971470129, | |
| "learning_rate": 7.597919750177168e-07, | |
| "logits/chosen": -2.5199341773986816, | |
| "logits/rejected": -2.4877076148986816, | |
| "logps/chosen": -234.12948608398438, | |
| "logps/rejected": -236.8078155517578, | |
| "loss": 0.2646, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 1.3459094762802124, | |
| "rewards/margins": 3.2457852363586426, | |
| "rewards/rejected": -1.8998762369155884, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 1.1214953271028036, | |
| "grad_norm": 17.398245773283783, | |
| "learning_rate": 7.30032518865576e-07, | |
| "logits/chosen": -2.5517895221710205, | |
| "logits/rejected": -2.541144847869873, | |
| "logps/chosen": -249.17434692382812, | |
| "logps/rejected": -215.4199676513672, | |
| "loss": 0.2414, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.5442664623260498, | |
| "rewards/margins": 3.122053623199463, | |
| "rewards/rejected": -1.577787160873413, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.1838006230529594, | |
| "grad_norm": 22.634622959038673, | |
| "learning_rate": 6.992005449231207e-07, | |
| "logits/chosen": -2.577732801437378, | |
| "logits/rejected": -2.5622453689575195, | |
| "logps/chosen": -236.0395050048828, | |
| "logps/rejected": -224.38516235351562, | |
| "loss": 0.2744, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 2.1463866233825684, | |
| "rewards/margins": 3.273930072784424, | |
| "rewards/rejected": -1.1275434494018555, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 1.2461059190031152, | |
| "grad_norm": 22.411472968651037, | |
| "learning_rate": 6.67439806085493e-07, | |
| "logits/chosen": -2.611746311187744, | |
| "logits/rejected": -2.6113743782043457, | |
| "logps/chosen": -241.65982055664062, | |
| "logps/rejected": -233.66464233398438, | |
| "loss": 0.2278, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 2.5751986503601074, | |
| "rewards/margins": 3.778069019317627, | |
| "rewards/rejected": -1.2028696537017822, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.2461059190031152, | |
| "eval_logits/chosen": -2.6071712970733643, | |
| "eval_logits/rejected": -2.605973958969116, | |
| "eval_logps/chosen": -238.49212646484375, | |
| "eval_logps/rejected": -209.84432983398438, | |
| "eval_loss": 0.5324757099151611, | |
| "eval_rewards/accuracies": 0.7986111044883728, | |
| "eval_rewards/chosen": 1.6095833778381348, | |
| "eval_rewards/margins": 2.2776832580566406, | |
| "eval_rewards/rejected": -0.6680999398231506, | |
| "eval_runtime": 153.2854, | |
| "eval_samples_per_second": 14.874, | |
| "eval_steps_per_second": 0.235, | |
| "step": 100 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 240, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1178822762299392.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |