| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9983361064891847, | |
| "eval_steps": 500, | |
| "global_step": 150, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.01, | |
| "grad_norm": 1142.1729750161728, | |
| "learning_rate": 6.666666666666666e-10, | |
| "logits/chosen": -4.106247425079346, | |
| "logits/rejected": -4.200438499450684, | |
| "logps/chosen": -382.81439208984375, | |
| "logps/rejected": -357.65960693359375, | |
| "loss": 0.685, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 1162.0154294843069, | |
| "learning_rate": 6.666666666666667e-09, | |
| "logits/chosen": -4.2175726890563965, | |
| "logits/rejected": -4.321321487426758, | |
| "logps/chosen": -334.61383056640625, | |
| "logps/rejected": -313.4597473144531, | |
| "loss": 0.7288, | |
| "rewards/accuracies": 0.4756944477558136, | |
| "rewards/chosen": -0.016012493520975113, | |
| "rewards/margins": 0.006440857890993357, | |
| "rewards/rejected": -0.022453350946307182, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 1254.20259356522, | |
| "learning_rate": 9.966191788709716e-09, | |
| "logits/chosen": -4.266871452331543, | |
| "logits/rejected": -4.419375896453857, | |
| "logps/chosen": -313.91156005859375, | |
| "logps/rejected": -288.8208923339844, | |
| "loss": 0.7239, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.0004551798047032207, | |
| "rewards/margins": 0.007867029868066311, | |
| "rewards/rejected": -0.0074118501506745815, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 1168.2123312853032, | |
| "learning_rate": 9.698463103929542e-09, | |
| "logits/chosen": -4.302128791809082, | |
| "logits/rejected": -4.365870475769043, | |
| "logps/chosen": -308.2377014160156, | |
| "logps/rejected": -285.6295471191406, | |
| "loss": 0.7344, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.019400831311941147, | |
| "rewards/margins": 0.019786948338150978, | |
| "rewards/rejected": -0.0003861159202642739, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 1099.047513823573, | |
| "learning_rate": 9.177439057064682e-09, | |
| "logits/chosen": -4.187483310699463, | |
| "logits/rejected": -4.270766735076904, | |
| "logps/chosen": -332.38433837890625, | |
| "logps/rejected": -307.4942321777344, | |
| "loss": 0.7115, | |
| "rewards/accuracies": 0.5406249761581421, | |
| "rewards/chosen": 0.0541040301322937, | |
| "rewards/margins": 0.07320869714021683, | |
| "rewards/rejected": -0.01910465955734253, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 1141.2520247434832, | |
| "learning_rate": 8.431208189343668e-09, | |
| "logits/chosen": -4.198305606842041, | |
| "logits/rejected": -4.367269992828369, | |
| "logps/chosen": -333.6199645996094, | |
| "logps/rejected": -308.95989990234375, | |
| "loss": 0.7163, | |
| "rewards/accuracies": 0.4906249940395355, | |
| "rewards/chosen": 0.06264184415340424, | |
| "rewards/margins": -0.024443484842777252, | |
| "rewards/rejected": 0.08708532154560089, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 1099.3679350302498, | |
| "learning_rate": 7.500000000000001e-09, | |
| "logits/chosen": -4.187924385070801, | |
| "logits/rejected": -4.2703022956848145, | |
| "logps/chosen": -323.7719421386719, | |
| "logps/rejected": -308.23748779296875, | |
| "loss": 0.7118, | |
| "rewards/accuracies": 0.515625, | |
| "rewards/chosen": 0.11530591547489166, | |
| "rewards/margins": 0.03647974878549576, | |
| "rewards/rejected": 0.07882615178823471, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 1096.6613245075011, | |
| "learning_rate": 6.434016163555451e-09, | |
| "logits/chosen": -4.1484785079956055, | |
| "logits/rejected": -4.303661346435547, | |
| "logps/chosen": -344.68658447265625, | |
| "logps/rejected": -314.47064208984375, | |
| "loss": 0.6973, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.1820925623178482, | |
| "rewards/margins": 0.10691970586776733, | |
| "rewards/rejected": 0.07517284899950027, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "grad_norm": 1050.1170453783911, | |
| "learning_rate": 5.290724144552379e-09, | |
| "logits/chosen": -4.223504543304443, | |
| "logits/rejected": -4.3897480964660645, | |
| "logps/chosen": -334.43511962890625, | |
| "logps/rejected": -305.484375, | |
| "loss": 0.6912, | |
| "rewards/accuracies": 0.559374988079071, | |
| "rewards/chosen": 0.20204909145832062, | |
| "rewards/margins": 0.08096315711736679, | |
| "rewards/rejected": 0.12108592689037323, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 1080.8957995779926, | |
| "learning_rate": 4.131759111665349e-09, | |
| "logits/chosen": -4.228762626647949, | |
| "logits/rejected": -4.349400997161865, | |
| "logps/chosen": -327.1580810546875, | |
| "logps/rejected": -307.87689208984375, | |
| "loss": 0.674, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.24468111991882324, | |
| "rewards/margins": 0.05739554762840271, | |
| "rewards/rejected": 0.18728554248809814, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 1086.0982504773885, | |
| "learning_rate": 3.0196011698042157e-09, | |
| "logits/chosen": -4.210589408874512, | |
| "logits/rejected": -4.420603275299072, | |
| "logps/chosen": -320.5531311035156, | |
| "logps/rejected": -289.7651062011719, | |
| "loss": 0.6799, | |
| "rewards/accuracies": 0.578125, | |
| "rewards/chosen": 0.2572989761829376, | |
| "rewards/margins": 0.10579316318035126, | |
| "rewards/rejected": 0.15150579810142517, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 1114.0041176823654, | |
| "learning_rate": 2.0142070414860704e-09, | |
| "logits/chosen": -4.225996971130371, | |
| "logits/rejected": -4.2789506912231445, | |
| "logps/chosen": -314.6085205078125, | |
| "logps/rejected": -303.3541564941406, | |
| "loss": 0.6851, | |
| "rewards/accuracies": 0.546875, | |
| "rewards/chosen": 0.27688390016555786, | |
| "rewards/margins": 0.041596584022045135, | |
| "rewards/rejected": 0.23528733849525452, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 1097.8240578626694, | |
| "learning_rate": 1.1697777844051105e-09, | |
| "logits/chosen": -4.1503801345825195, | |
| "logits/rejected": -4.306635856628418, | |
| "logps/chosen": -338.7808837890625, | |
| "logps/rejected": -313.2768249511719, | |
| "loss": 0.6758, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.3373282849788666, | |
| "rewards/margins": 0.1483292281627655, | |
| "rewards/rejected": 0.18899908661842346, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "grad_norm": 1113.3294478605874, | |
| "learning_rate": 5.318367983829391e-10, | |
| "logits/chosen": -4.178295135498047, | |
| "logits/rejected": -4.3724284172058105, | |
| "logps/chosen": -323.9390563964844, | |
| "logps/rejected": -304.91119384765625, | |
| "loss": 0.6643, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.340619832277298, | |
| "rewards/margins": 0.12920674681663513, | |
| "rewards/rejected": 0.21141307055950165, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 1112.5811910392042, | |
| "learning_rate": 1.3477564710088098e-10, | |
| "logits/chosen": -4.293918609619141, | |
| "logits/rejected": -4.359633445739746, | |
| "logps/chosen": -307.56317138671875, | |
| "logps/rejected": -297.0579833984375, | |
| "loss": 0.6697, | |
| "rewards/accuracies": 0.596875011920929, | |
| "rewards/chosen": 0.34358957409858704, | |
| "rewards/margins": 0.11470258235931396, | |
| "rewards/rejected": 0.22888696193695068, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 1082.9191655386894, | |
| "learning_rate": 0.0, | |
| "logits/chosen": -4.258730411529541, | |
| "logits/rejected": -4.332475185394287, | |
| "logps/chosen": -312.3280029296875, | |
| "logps/rejected": -300.03082275390625, | |
| "loss": 0.6661, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.3249002993106842, | |
| "rewards/margins": 0.10186745971441269, | |
| "rewards/rejected": 0.2230328619480133, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 150, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6950656716028849, | |
| "train_runtime": 4446.5407, | |
| "train_samples_per_second": 8.646, | |
| "train_steps_per_second": 0.034 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 150, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |