| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 223, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 9.89770231320471, | |
| "learning_rate": 2.1739130434782606e-08, | |
| "logits/chosen": -1.3066810369491577, | |
| "logits/rejected": -1.1899611949920654, | |
| "logps/chosen": -190.14166259765625, | |
| "logps/rejected": -191.2518310546875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "grad_norm": 9.33253553043054, | |
| "learning_rate": 2.1739130434782607e-07, | |
| "logits/chosen": -0.7182029485702515, | |
| "logits/rejected": -1.1984152793884277, | |
| "logps/chosen": -120.43441772460938, | |
| "logps/rejected": -182.3992919921875, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.4583333432674408, | |
| "rewards/chosen": 3.514537092996761e-05, | |
| "rewards/margins": 0.00027891554054804146, | |
| "rewards/rejected": -0.00024377032241318375, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 9.34106704706082, | |
| "learning_rate": 4.3478260869565214e-07, | |
| "logits/chosen": -0.7554597854614258, | |
| "logits/rejected": -1.2149486541748047, | |
| "logps/chosen": -119.57357025146484, | |
| "logps/rejected": -192.81492614746094, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": -0.00020422041416168213, | |
| "rewards/margins": -0.00024611633853055537, | |
| "rewards/rejected": 4.189582614344545e-05, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "grad_norm": 9.104090024380893, | |
| "learning_rate": 4.984902388637949e-07, | |
| "logits/chosen": -0.8330415487289429, | |
| "logits/rejected": -1.2031960487365723, | |
| "logps/chosen": -122.64117431640625, | |
| "logps/rejected": -195.4859161376953, | |
| "loss": 0.6922, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -0.0009992063278332353, | |
| "rewards/margins": 0.0014079783577471972, | |
| "rewards/rejected": -0.002407185034826398, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 9.627025658314095, | |
| "learning_rate": 4.911393546144495e-07, | |
| "logits/chosen": -0.8064976930618286, | |
| "logits/rejected": -1.1510051488876343, | |
| "logps/chosen": -132.39263916015625, | |
| "logps/rejected": -191.39283752441406, | |
| "loss": 0.6902, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.0007292800582945347, | |
| "rewards/margins": 0.0054230657406151295, | |
| "rewards/rejected": -0.004693786613643169, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "grad_norm": 9.23487552340627, | |
| "learning_rate": 4.778508191588613e-07, | |
| "logits/chosen": -0.8550766706466675, | |
| "logits/rejected": -1.1777122020721436, | |
| "logps/chosen": -133.31373596191406, | |
| "logps/rejected": -179.41244506835938, | |
| "loss": 0.6882, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.0029656877741217613, | |
| "rewards/margins": 0.012276900932192802, | |
| "rewards/rejected": -0.009311212226748466, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 9.05100661494371, | |
| "learning_rate": 4.5895184034206755e-07, | |
| "logits/chosen": -0.8925403356552124, | |
| "logits/rejected": -1.131866216659546, | |
| "logps/chosen": -136.6233673095703, | |
| "logps/rejected": -188.6793975830078, | |
| "loss": 0.6856, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.001518428442068398, | |
| "rewards/margins": 0.012228612788021564, | |
| "rewards/rejected": -0.010710184462368488, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "grad_norm": 9.997440253666376, | |
| "learning_rate": 4.3490777374465244e-07, | |
| "logits/chosen": -0.9242467880249023, | |
| "logits/rejected": -1.2503801584243774, | |
| "logps/chosen": -145.55833435058594, | |
| "logps/rejected": -184.1634063720703, | |
| "loss": 0.6833, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.002613705350086093, | |
| "rewards/margins": 0.020717555657029152, | |
| "rewards/rejected": -0.01810384914278984, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 9.187427456431445, | |
| "learning_rate": 4.063106640839263e-07, | |
| "logits/chosen": -0.8415228724479675, | |
| "logits/rejected": -1.3067677021026611, | |
| "logps/chosen": -135.7071990966797, | |
| "logps/rejected": -211.23416137695312, | |
| "loss": 0.6784, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.003357144072651863, | |
| "rewards/margins": 0.03158879652619362, | |
| "rewards/rejected": -0.028231654316186905, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "grad_norm": 9.183369355368722, | |
| "learning_rate": 3.7386466710810187e-07, | |
| "logits/chosen": -0.8030783534049988, | |
| "logits/rejected": -1.106172800064087, | |
| "logps/chosen": -124.84532165527344, | |
| "logps/rejected": -189.9628143310547, | |
| "loss": 0.6759, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.0018527141073718667, | |
| "rewards/margins": 0.03642088547348976, | |
| "rewards/rejected": -0.034568168222904205, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "grad_norm": 9.558099135591291, | |
| "learning_rate": 3.383687109448143e-07, | |
| "logits/chosen": -0.8545591235160828, | |
| "logits/rejected": -1.2025887966156006, | |
| "logps/chosen": -160.20260620117188, | |
| "logps/rejected": -195.79226684570312, | |
| "loss": 0.6732, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -0.006974806543439627, | |
| "rewards/margins": 0.03974348306655884, | |
| "rewards/rejected": -0.0467182882130146, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 9.470887743050254, | |
| "learning_rate": 3.0069682383912807e-07, | |
| "logits/chosen": -0.8946328163146973, | |
| "logits/rejected": -1.2082103490829468, | |
| "logps/chosen": -136.27890014648438, | |
| "logps/rejected": -195.939453125, | |
| "loss": 0.6686, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.005538312252610922, | |
| "rewards/margins": 0.05113464593887329, | |
| "rewards/rejected": -0.05667296051979065, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "grad_norm": 9.426731682830118, | |
| "learning_rate": 2.6177661267741063e-07, | |
| "logits/chosen": -0.9082379341125488, | |
| "logits/rejected": -1.268019437789917, | |
| "logps/chosen": -129.68191528320312, | |
| "logps/rejected": -195.88900756835938, | |
| "loss": 0.6638, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.004824039526283741, | |
| "rewards/margins": 0.06547943502664566, | |
| "rewards/rejected": -0.07030347734689713, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "grad_norm": 10.012422462946702, | |
| "learning_rate": 2.2256642222723867e-07, | |
| "logits/chosen": -0.9008873701095581, | |
| "logits/rejected": -1.2844858169555664, | |
| "logps/chosen": -122.87223052978516, | |
| "logps/rejected": -198.33155822753906, | |
| "loss": 0.6586, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.01183926872909069, | |
| "rewards/margins": 0.07219536602497101, | |
| "rewards/rejected": -0.08403463661670685, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "grad_norm": 10.278151003618936, | |
| "learning_rate": 1.8403173750865683e-07, | |
| "logits/chosen": -0.8554606437683105, | |
| "logits/rejected": -1.2466466426849365, | |
| "logps/chosen": -143.69813537597656, | |
| "logps/rejected": -194.7081298828125, | |
| "loss": 0.66, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.028567075729370117, | |
| "rewards/margins": 0.06817030906677246, | |
| "rewards/rejected": -0.09673738479614258, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "grad_norm": 10.058138299560008, | |
| "learning_rate": 1.471214103487228e-07, | |
| "logits/chosen": -0.948029637336731, | |
| "logits/rejected": -1.2443821430206299, | |
| "logps/chosen": -126.2611312866211, | |
| "logps/rejected": -193.74984741210938, | |
| "loss": 0.6567, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -0.02120281383395195, | |
| "rewards/margins": 0.08252517133951187, | |
| "rewards/rejected": -0.10372798144817352, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "grad_norm": 10.590989952026893, | |
| "learning_rate": 1.1274429550046701e-07, | |
| "logits/chosen": -0.9630361795425415, | |
| "logits/rejected": -1.3080346584320068, | |
| "logps/chosen": -137.56187438964844, | |
| "logps/rejected": -211.8583526611328, | |
| "loss": 0.6551, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -0.041766177862882614, | |
| "rewards/margins": 0.0846499651670456, | |
| "rewards/rejected": -0.1264161467552185, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "grad_norm": 10.046412745054047, | |
| "learning_rate": 8.174687162255673e-08, | |
| "logits/chosen": -0.8078781366348267, | |
| "logits/rejected": -1.2546970844268799, | |
| "logps/chosen": -131.09913635253906, | |
| "logps/rejected": -185.79489135742188, | |
| "loss": 0.6495, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.03623218461871147, | |
| "rewards/margins": 0.08339719474315643, | |
| "rewards/rejected": -0.11962936818599701, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "grad_norm": 10.575942205861123, | |
| "learning_rate": 5.489239816541755e-08, | |
| "logits/chosen": -0.8742658495903015, | |
| "logits/rejected": -1.3499236106872559, | |
| "logps/chosen": -126.0824203491211, | |
| "logps/rejected": -215.39071655273438, | |
| "loss": 0.646, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -0.035958122462034225, | |
| "rewards/margins": 0.10524071753025055, | |
| "rewards/rejected": -0.14119884371757507, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "grad_norm": 11.01986954869563, | |
| "learning_rate": 3.2842121390452224e-08, | |
| "logits/chosen": -0.9610059857368469, | |
| "logits/rejected": -1.3607670068740845, | |
| "logps/chosen": -133.7125244140625, | |
| "logps/rejected": -217.3642120361328, | |
| "loss": 0.6496, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -0.049020860344171524, | |
| "rewards/margins": 0.09572702646255493, | |
| "rewards/rejected": -0.14474789798259735, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "grad_norm": 10.150344499203792, | |
| "learning_rate": 1.6138992292533182e-08, | |
| "logits/chosen": -0.8730287551879883, | |
| "logits/rejected": -1.3922489881515503, | |
| "logps/chosen": -132.81820678710938, | |
| "logps/rejected": -212.6441192626953, | |
| "loss": 0.6445, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.046734295785427094, | |
| "rewards/margins": 0.10609643161296844, | |
| "rewards/rejected": -0.15283071994781494, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 10.624851033611936, | |
| "learning_rate": 5.194297344558534e-09, | |
| "logits/chosen": -0.9464960098266602, | |
| "logits/rejected": -1.3372732400894165, | |
| "logps/chosen": -142.20260620117188, | |
| "logps/rejected": -218.54898071289062, | |
| "loss": 0.648, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.04380584880709648, | |
| "rewards/margins": 0.10490747541189194, | |
| "rewards/rejected": -0.14871332049369812, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "grad_norm": 10.584763250355945, | |
| "learning_rate": 2.775312595075241e-10, | |
| "logits/chosen": -0.9528037309646606, | |
| "logits/rejected": -1.2510448694229126, | |
| "logps/chosen": -146.1761932373047, | |
| "logps/rejected": -200.77236938476562, | |
| "loss": 0.6465, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -0.05798276513814926, | |
| "rewards/margins": 0.09351377934217453, | |
| "rewards/rejected": -0.1514965295791626, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 223, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6677660370086875, | |
| "train_runtime": 2913.0719, | |
| "train_samples_per_second": 4.894, | |
| "train_steps_per_second": 0.077 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 223, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |