| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.971563981042654, | |
| "eval_steps": 100, | |
| "global_step": 104, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.018957345971563982, | |
| "grad_norm": 132.15360444004384, | |
| "learning_rate": 4.545454545454545e-08, | |
| "logits/chosen": 117.53560638427734, | |
| "logits/rejected": 126.8960952758789, | |
| "logps/chosen": -335.40118408203125, | |
| "logps/rejected": -439.16552734375, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.1895734597156398, | |
| "grad_norm": 132.3674027987073, | |
| "learning_rate": 4.545454545454545e-07, | |
| "logits/chosen": 135.01699829101562, | |
| "logits/rejected": 138.37664794921875, | |
| "logps/chosen": -396.05718994140625, | |
| "logps/rejected": -439.1203918457031, | |
| "loss": 0.7127, | |
| "rewards/accuracies": 0.4583333432674408, | |
| "rewards/chosen": -0.0030322629027068615, | |
| "rewards/margins": -0.013390823267400265, | |
| "rewards/rejected": 0.010358559899032116, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.3791469194312796, | |
| "grad_norm": 131.21733523095625, | |
| "learning_rate": 4.885348141000122e-07, | |
| "logits/chosen": 121.60444641113281, | |
| "logits/rejected": 125.29842376708984, | |
| "logps/chosen": -370.2664489746094, | |
| "logps/rejected": -422.78851318359375, | |
| "loss": 0.6459, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.10727670043706894, | |
| "rewards/margins": 0.247134730219841, | |
| "rewards/rejected": -0.13985800743103027, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.5687203791469194, | |
| "grad_norm": 117.90232463642135, | |
| "learning_rate": 4.5025027361734613e-07, | |
| "logits/chosen": 142.974853515625, | |
| "logits/rejected": 136.52386474609375, | |
| "logps/chosen": -424.7781677246094, | |
| "logps/rejected": -469.64813232421875, | |
| "loss": 0.5746, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.6156466007232666, | |
| "rewards/margins": 0.8666501045227051, | |
| "rewards/rejected": -2.4822967052459717, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.7582938388625592, | |
| "grad_norm": 104.91283452119073, | |
| "learning_rate": 3.893311157806091e-07, | |
| "logits/chosen": 126.9936752319336, | |
| "logits/rejected": 115.53365325927734, | |
| "logps/chosen": -399.81353759765625, | |
| "logps/rejected": -426.99853515625, | |
| "loss": 0.5456, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.2809689044952393, | |
| "rewards/margins": 1.1751956939697266, | |
| "rewards/rejected": -3.456164598464966, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.9478672985781991, | |
| "grad_norm": 123.57780236639618, | |
| "learning_rate": 3.126631330646801e-07, | |
| "logits/chosen": 142.1190643310547, | |
| "logits/rejected": 146.2515411376953, | |
| "logps/chosen": -456.97979736328125, | |
| "logps/rejected": -540.1392822265625, | |
| "loss": 0.489, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -1.891798734664917, | |
| "rewards/margins": 1.2988468408584595, | |
| "rewards/rejected": -3.190645456314087, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.1374407582938388, | |
| "grad_norm": 67.1680971334559, | |
| "learning_rate": 2.2891223348923882e-07, | |
| "logits/chosen": 133.56114196777344, | |
| "logits/rejected": 137.20738220214844, | |
| "logps/chosen": -449.55303955078125, | |
| "logps/rejected": -534.8367919921875, | |
| "loss": 0.3117, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -2.5773684978485107, | |
| "rewards/margins": 2.346193790435791, | |
| "rewards/rejected": -4.923562049865723, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.3270142180094786, | |
| "grad_norm": 48.54475300946312, | |
| "learning_rate": 1.4754491880085317e-07, | |
| "logits/chosen": 125.71492004394531, | |
| "logits/rejected": 127.68719482421875, | |
| "logps/chosen": -426.90228271484375, | |
| "logps/rejected": -528.0679321289062, | |
| "loss": 0.195, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -2.9879212379455566, | |
| "rewards/margins": 2.7397806644439697, | |
| "rewards/rejected": -5.727701663970947, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.5165876777251186, | |
| "grad_norm": 55.371866892062, | |
| "learning_rate": 7.775827023107834e-08, | |
| "logits/chosen": 111.7248306274414, | |
| "logits/rejected": 128.3420867919922, | |
| "logps/chosen": -427.53106689453125, | |
| "logps/rejected": -546.7640991210938, | |
| "loss": 0.1651, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.693999767303467, | |
| "rewards/margins": 2.9823195934295654, | |
| "rewards/rejected": -6.676319122314453, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 1.7061611374407581, | |
| "grad_norm": 40.99464664899818, | |
| "learning_rate": 2.7440387297912122e-08, | |
| "logits/chosen": 110.8941879272461, | |
| "logits/rejected": 123.70848083496094, | |
| "logps/chosen": -457.2183532714844, | |
| "logps/rejected": -575.8634033203125, | |
| "loss": 0.1557, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.5393664836883545, | |
| "rewards/margins": 3.369715929031372, | |
| "rewards/rejected": -6.909082889556885, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.8957345971563981, | |
| "grad_norm": 45.385328063823785, | |
| "learning_rate": 2.27878296044029e-09, | |
| "logits/chosen": 117.1551284790039, | |
| "logits/rejected": 117.0487060546875, | |
| "logps/chosen": -446.9934997558594, | |
| "logps/rejected": -541.2728881835938, | |
| "loss": 0.1585, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -3.250919818878174, | |
| "rewards/margins": 2.9745240211486816, | |
| "rewards/rejected": -6.2254438400268555, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.8957345971563981, | |
| "eval_logits/chosen": 96.73149871826172, | |
| "eval_logits/rejected": 91.1323013305664, | |
| "eval_logps/chosen": -446.253662109375, | |
| "eval_logps/rejected": -476.3663635253906, | |
| "eval_loss": 0.46732592582702637, | |
| "eval_rewards/accuracies": 0.6979166865348816, | |
| "eval_rewards/chosen": -4.1292724609375, | |
| "eval_rewards/margins": 1.5854991674423218, | |
| "eval_rewards/rejected": -5.714771270751953, | |
| "eval_runtime": 120.4793, | |
| "eval_samples_per_second": 6.225, | |
| "eval_steps_per_second": 0.199, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.971563981042654, | |
| "step": 104, | |
| "total_flos": 0.0, | |
| "train_loss": 0.39153398688022906, | |
| "train_runtime": 2311.0387, | |
| "train_samples_per_second": 5.842, | |
| "train_steps_per_second": 0.045 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 104, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 500, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |