{ "best_metric": null, "best_model_checkpoint": null, "epoch": 1.998381162619573, "eval_steps": 1, "global_step": 848, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.0235467255334805, "grad_norm": 4.103935241699219, "learning_rate": 3.060971262053483e-06, "logits/chosen": -4.40625, "logits/rejected": -4.4375, "logps/chosen": -652.0, "logps/rejected": -592.0, "loss": 0.6951, "rewards/accuracies": 0.08124999701976776, "rewards/chosen": -0.0006256103515625, "rewards/margins": -0.0089111328125, "rewards/rejected": 0.00830078125, "step": 10 }, { "epoch": 0.047093451066961, "grad_norm": 3.962085247039795, "learning_rate": 3.9824154277970135e-06, "logits/chosen": -4.5, "logits/rejected": -4.5, "logps/chosen": -592.0, "logps/rejected": -520.0, "loss": 0.6868, "rewards/accuracies": 0.20000000298023224, "rewards/chosen": 0.0006256103515625, "rewards/margins": 0.00830078125, "rewards/rejected": -0.007659912109375, "step": 20 }, { "epoch": 0.0706401766004415, "grad_norm": 4.004490852355957, "learning_rate": 4.521425711265269e-06, "logits/chosen": -4.46875, "logits/rejected": -4.5, "logps/chosen": -608.0, "logps/rejected": -524.0, "loss": 0.6591, "rewards/accuracies": 0.3812499940395355, "rewards/chosen": 0.053955078125, "rewards/margins": 0.076171875, "rewards/rejected": -0.0220947265625, "step": 30 }, { "epoch": 0.094186902133922, "grad_norm": 3.424196481704712, "learning_rate": 4.903859593540544e-06, "logits/chosen": -4.4375, "logits/rejected": -4.40625, "logps/chosen": -584.0, "logps/rejected": -504.0, "loss": 0.6295, "rewards/accuracies": 0.5625, "rewards/chosen": 0.09521484375, "rewards/margins": 0.1455078125, "rewards/rejected": -0.05078125, "step": 40 }, { "epoch": 0.1177336276674025, "grad_norm": 3.229916572570801, "learning_rate": 4.962732919254658e-06, "logits/chosen": -4.5625, "logits/rejected": -4.53125, "logps/chosen": -548.0, "logps/rejected": -502.0, "loss": 0.58, "rewards/accuracies": 0.6499999761581421, "rewards/chosen": 0.15234375, "rewards/margins": 0.28125, "rewards/rejected": -0.12890625, "step": 50 }, { "epoch": 0.141280353200883, "grad_norm": 3.3798811435699463, "learning_rate": 4.900621118012423e-06, "logits/chosen": -4.40625, "logits/rejected": -4.40625, "logps/chosen": -580.0, "logps/rejected": -454.0, "loss": 0.5537, "rewards/accuracies": 0.7250000238418579, "rewards/chosen": 0.302734375, "rewards/margins": 0.423828125, "rewards/rejected": -0.12158203125, "step": 60 }, { "epoch": 0.1648270787343635, "grad_norm": 3.2893781661987305, "learning_rate": 4.8385093167701866e-06, "logits/chosen": -4.40625, "logits/rejected": -4.40625, "logps/chosen": -608.0, "logps/rejected": -508.0, "loss": 0.5351, "rewards/accuracies": 0.7250000238418579, "rewards/chosen": 0.251953125, "rewards/margins": 0.4609375, "rewards/rejected": -0.208984375, "step": 70 }, { "epoch": 0.188373804267844, "grad_norm": 3.3645501136779785, "learning_rate": 4.776397515527951e-06, "logits/chosen": -4.5, "logits/rejected": -4.53125, "logps/chosen": -588.0, "logps/rejected": -468.0, "loss": 0.5096, "rewards/accuracies": 0.7749999761581421, "rewards/chosen": 0.34375, "rewards/margins": 0.6796875, "rewards/rejected": -0.3359375, "step": 80 }, { "epoch": 0.2119205298013245, "grad_norm": 3.14998197555542, "learning_rate": 4.714285714285715e-06, "logits/chosen": -4.40625, "logits/rejected": -4.4375, "logps/chosen": -632.0, "logps/rejected": -490.0, "loss": 0.4901, "rewards/accuracies": 0.7749999761581421, "rewards/chosen": 0.375, "rewards/margins": 0.69140625, "rewards/rejected": -0.318359375, "step": 90 }, { "epoch": 0.235467255334805, "grad_norm": 3.2294790744781494, "learning_rate": 4.652173913043478e-06, "logits/chosen": -4.5, "logits/rejected": -4.5, "logps/chosen": -580.0, "logps/rejected": -506.0, "loss": 0.4736, "rewards/accuracies": 0.762499988079071, "rewards/chosen": 0.369140625, "rewards/margins": 0.6328125, "rewards/rejected": -0.263671875, "step": 100 }, { "epoch": 0.25901398086828553, "grad_norm": 2.738917827606201, "learning_rate": 4.590062111801243e-06, "logits/chosen": -4.5, "logits/rejected": -4.59375, "logps/chosen": -604.0, "logps/rejected": -516.0, "loss": 0.4703, "rewards/accuracies": 0.762499988079071, "rewards/chosen": 0.5, "rewards/margins": 0.9296875, "rewards/rejected": -0.431640625, "step": 110 }, { "epoch": 0.282560706401766, "grad_norm": 3.060237169265747, "learning_rate": 4.527950310559007e-06, "logits/chosen": -4.40625, "logits/rejected": -4.40625, "logps/chosen": -632.0, "logps/rejected": -552.0, "loss": 0.4649, "rewards/accuracies": 0.7437499761581421, "rewards/chosen": 0.53515625, "rewards/margins": 0.99609375, "rewards/rejected": -0.462890625, "step": 120 }, { "epoch": 0.30610743193524653, "grad_norm": 3.196960926055908, "learning_rate": 4.46583850931677e-06, "logits/chosen": -4.34375, "logits/rejected": -4.34375, "logps/chosen": -644.0, "logps/rejected": -524.0, "loss": 0.4824, "rewards/accuracies": 0.762499988079071, "rewards/chosen": 0.54296875, "rewards/margins": 1.0, "rewards/rejected": -0.45703125, "step": 130 }, { "epoch": 0.329654157468727, "grad_norm": 3.1361074447631836, "learning_rate": 4.4037267080745344e-06, "logits/chosen": -4.375, "logits/rejected": -4.40625, "logps/chosen": -628.0, "logps/rejected": -548.0, "loss": 0.4539, "rewards/accuracies": 0.7749999761581421, "rewards/chosen": 0.447265625, "rewards/margins": 0.9921875, "rewards/rejected": -0.546875, "step": 140 }, { "epoch": 0.35320088300220753, "grad_norm": 2.6170785427093506, "learning_rate": 4.3416149068322985e-06, "logits/chosen": -4.46875, "logits/rejected": -4.4375, "logps/chosen": -580.0, "logps/rejected": -512.0, "loss": 0.4478, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.65625, "rewards/margins": 1.0390625, "rewards/rejected": -0.3828125, "step": 150 }, { "epoch": 0.376747608535688, "grad_norm": 2.7828519344329834, "learning_rate": 4.279503105590063e-06, "logits/chosen": -4.375, "logits/rejected": -4.34375, "logps/chosen": -608.0, "logps/rejected": -506.0, "loss": 0.4407, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": 0.5859375, "rewards/margins": 1.09375, "rewards/rejected": -0.5078125, "step": 160 }, { "epoch": 0.40029433406916853, "grad_norm": 2.906342029571533, "learning_rate": 4.217391304347827e-06, "logits/chosen": -4.5, "logits/rejected": -4.4375, "logps/chosen": -584.0, "logps/rejected": -520.0, "loss": 0.4309, "rewards/accuracies": 0.824999988079071, "rewards/chosen": 0.6796875, "rewards/margins": 1.2578125, "rewards/rejected": -0.58203125, "step": 170 }, { "epoch": 0.423841059602649, "grad_norm": 2.827556848526001, "learning_rate": 4.15527950310559e-06, "logits/chosen": -4.5, "logits/rejected": -4.53125, "logps/chosen": -616.0, "logps/rejected": -506.0, "loss": 0.4508, "rewards/accuracies": 0.78125, "rewards/chosen": 0.6875, "rewards/margins": 1.140625, "rewards/rejected": -0.453125, "step": 180 }, { "epoch": 0.44738778513612953, "grad_norm": 2.7354469299316406, "learning_rate": 4.093167701863354e-06, "logits/chosen": -4.4375, "logits/rejected": -4.5, "logps/chosen": -616.0, "logps/rejected": -510.0, "loss": 0.4341, "rewards/accuracies": 0.78125, "rewards/chosen": 0.78515625, "rewards/margins": 1.3046875, "rewards/rejected": -0.51953125, "step": 190 }, { "epoch": 0.47093451066961, "grad_norm": 2.674173593521118, "learning_rate": 4.031055900621118e-06, "logits/chosen": -4.4375, "logits/rejected": -4.4375, "logps/chosen": -600.0, "logps/rejected": -474.0, "loss": 0.4224, "rewards/accuracies": 0.8125, "rewards/chosen": 0.72265625, "rewards/margins": 1.140625, "rewards/rejected": -0.4140625, "step": 200 }, { "epoch": 0.49448123620309054, "grad_norm": 3.0752134323120117, "learning_rate": 3.968944099378882e-06, "logits/chosen": -4.4375, "logits/rejected": -4.46875, "logps/chosen": -628.0, "logps/rejected": -584.0, "loss": 0.4401, "rewards/accuracies": 0.75, "rewards/chosen": 0.71875, "rewards/margins": 1.2109375, "rewards/rejected": -0.494140625, "step": 210 }, { "epoch": 0.5180279617365711, "grad_norm": 3.02925181388855, "learning_rate": 3.906832298136646e-06, "logits/chosen": -4.34375, "logits/rejected": -4.40625, "logps/chosen": -644.0, "logps/rejected": -532.0, "loss": 0.4338, "rewards/accuracies": 0.7437499761581421, "rewards/chosen": 0.71875, "rewards/margins": 1.0078125, "rewards/rejected": -0.29296875, "step": 220 }, { "epoch": 0.5415746872700515, "grad_norm": 2.621617078781128, "learning_rate": 3.8447204968944105e-06, "logits/chosen": -4.40625, "logits/rejected": -4.4375, "logps/chosen": -616.0, "logps/rejected": -552.0, "loss": 0.4285, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": 0.81640625, "rewards/margins": 1.1640625, "rewards/rejected": -0.345703125, "step": 230 }, { "epoch": 0.565121412803532, "grad_norm": 2.96905779838562, "learning_rate": 3.782608695652174e-06, "logits/chosen": -4.46875, "logits/rejected": -4.53125, "logps/chosen": -560.0, "logps/rejected": -486.0, "loss": 0.4185, "rewards/accuracies": 0.8062499761581421, "rewards/chosen": 0.84375, "rewards/margins": 1.40625, "rewards/rejected": -0.55859375, "step": 240 }, { "epoch": 0.5886681383370125, "grad_norm": 3.1430623531341553, "learning_rate": 3.7204968944099383e-06, "logits/chosen": -4.46875, "logits/rejected": -4.53125, "logps/chosen": -624.0, "logps/rejected": -556.0, "loss": 0.4288, "rewards/accuracies": 0.768750011920929, "rewards/chosen": 0.85546875, "rewards/margins": 1.2890625, "rewards/rejected": -0.435546875, "step": 250 }, { "epoch": 0.6122148638704931, "grad_norm": 2.8485913276672363, "learning_rate": 3.6583850931677024e-06, "logits/chosen": -4.40625, "logits/rejected": -4.4375, "logps/chosen": -584.0, "logps/rejected": -552.0, "loss": 0.4362, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 0.90625, "rewards/margins": 1.3125, "rewards/rejected": -0.41015625, "step": 260 }, { "epoch": 0.6357615894039735, "grad_norm": 2.8013317584991455, "learning_rate": 3.596273291925466e-06, "logits/chosen": -4.4375, "logits/rejected": -4.4375, "logps/chosen": -584.0, "logps/rejected": -492.0, "loss": 0.4193, "rewards/accuracies": 0.7562500238418579, "rewards/chosen": 0.75, "rewards/margins": 1.25, "rewards/rejected": -0.498046875, "step": 270 }, { "epoch": 0.659308314937454, "grad_norm": 2.7474491596221924, "learning_rate": 3.53416149068323e-06, "logits/chosen": -4.4375, "logits/rejected": -4.5, "logps/chosen": -604.0, "logps/rejected": -472.0, "loss": 0.4029, "rewards/accuracies": 0.793749988079071, "rewards/chosen": 0.8046875, "rewards/margins": 1.3125, "rewards/rejected": -0.5078125, "step": 280 }, { "epoch": 0.6828550404709345, "grad_norm": 2.721045732498169, "learning_rate": 3.472049689440994e-06, "logits/chosen": -4.40625, "logits/rejected": -4.46875, "logps/chosen": -592.0, "logps/rejected": -510.0, "loss": 0.4111, "rewards/accuracies": 0.8062499761581421, "rewards/chosen": 1.09375, "rewards/margins": 1.6015625, "rewards/rejected": -0.51171875, "step": 290 }, { "epoch": 0.7064017660044151, "grad_norm": 2.9466845989227295, "learning_rate": 3.4099378881987584e-06, "logits/chosen": -4.40625, "logits/rejected": -4.40625, "logps/chosen": -568.0, "logps/rejected": -468.0, "loss": 0.4108, "rewards/accuracies": 0.7749999761581421, "rewards/chosen": 1.1015625, "rewards/margins": 1.578125, "rewards/rejected": -0.474609375, "step": 300 }, { "epoch": 0.7299484915378955, "grad_norm": 2.3606228828430176, "learning_rate": 3.347826086956522e-06, "logits/chosen": -4.40625, "logits/rejected": -4.40625, "logps/chosen": -596.0, "logps/rejected": -488.0, "loss": 0.4041, "rewards/accuracies": 0.78125, "rewards/chosen": 1.078125, "rewards/margins": 1.4453125, "rewards/rejected": -0.36328125, "step": 310 }, { "epoch": 0.753495217071376, "grad_norm": 3.1279854774475098, "learning_rate": 3.285714285714286e-06, "logits/chosen": -4.5, "logits/rejected": -4.46875, "logps/chosen": -564.0, "logps/rejected": -510.0, "loss": 0.4247, "rewards/accuracies": 0.762499988079071, "rewards/chosen": 0.9609375, "rewards/margins": 1.2265625, "rewards/rejected": -0.26171875, "step": 320 }, { "epoch": 0.7770419426048565, "grad_norm": 2.6707963943481445, "learning_rate": 3.22360248447205e-06, "logits/chosen": -4.375, "logits/rejected": -4.4375, "logps/chosen": -580.0, "logps/rejected": -492.0, "loss": 0.3787, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 1.0390625, "rewards/margins": 1.4453125, "rewards/rejected": -0.400390625, "step": 330 }, { "epoch": 0.8005886681383371, "grad_norm": 2.8069331645965576, "learning_rate": 3.1614906832298135e-06, "logits/chosen": -4.46875, "logits/rejected": -4.46875, "logps/chosen": -632.0, "logps/rejected": -600.0, "loss": 0.4048, "rewards/accuracies": 0.78125, "rewards/chosen": 1.078125, "rewards/margins": 1.3125, "rewards/rejected": -0.228515625, "step": 340 }, { "epoch": 0.8241353936718175, "grad_norm": 2.5299603939056396, "learning_rate": 3.099378881987578e-06, "logits/chosen": -4.53125, "logits/rejected": -4.5, "logps/chosen": -600.0, "logps/rejected": -532.0, "loss": 0.3885, "rewards/accuracies": 0.78125, "rewards/chosen": 1.1875, "rewards/margins": 1.4375, "rewards/rejected": -0.251953125, "step": 350 }, { "epoch": 0.847682119205298, "grad_norm": 2.7665863037109375, "learning_rate": 3.037267080745342e-06, "logits/chosen": -4.46875, "logits/rejected": -4.4375, "logps/chosen": -592.0, "logps/rejected": -472.0, "loss": 0.3892, "rewards/accuracies": 0.862500011920929, "rewards/chosen": 1.3515625, "rewards/margins": 1.7578125, "rewards/rejected": -0.404296875, "step": 360 }, { "epoch": 0.8712288447387785, "grad_norm": 2.8109049797058105, "learning_rate": 2.975155279503106e-06, "logits/chosen": -4.46875, "logits/rejected": -4.4375, "logps/chosen": -580.0, "logps/rejected": -524.0, "loss": 0.4115, "rewards/accuracies": 0.8062499761581421, "rewards/chosen": 1.171875, "rewards/margins": 1.5234375, "rewards/rejected": -0.349609375, "step": 370 }, { "epoch": 0.8947755702722591, "grad_norm": 2.657787799835205, "learning_rate": 2.9130434782608695e-06, "logits/chosen": -4.40625, "logits/rejected": -4.46875, "logps/chosen": -568.0, "logps/rejected": -532.0, "loss": 0.3831, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": 1.1953125, "rewards/margins": 1.578125, "rewards/rejected": -0.38671875, "step": 380 }, { "epoch": 0.9183222958057395, "grad_norm": 2.7588131427764893, "learning_rate": 2.850931677018634e-06, "logits/chosen": -4.46875, "logits/rejected": -4.46875, "logps/chosen": -588.0, "logps/rejected": -544.0, "loss": 0.3865, "rewards/accuracies": 0.8125, "rewards/chosen": 1.234375, "rewards/margins": 1.4453125, "rewards/rejected": -0.216796875, "step": 390 }, { "epoch": 0.94186902133922, "grad_norm": 3.0615074634552, "learning_rate": 2.788819875776398e-06, "logits/chosen": -4.4375, "logits/rejected": -4.5, "logps/chosen": -576.0, "logps/rejected": -502.0, "loss": 0.3967, "rewards/accuracies": 0.800000011920929, "rewards/chosen": 1.1875, "rewards/margins": 1.546875, "rewards/rejected": -0.359375, "step": 400 }, { "epoch": 0.9654157468727005, "grad_norm": 3.119372844696045, "learning_rate": 2.7267080745341618e-06, "logits/chosen": -4.53125, "logits/rejected": -4.5, "logps/chosen": -604.0, "logps/rejected": -576.0, "loss": 0.4029, "rewards/accuracies": 0.8125, "rewards/chosen": 1.28125, "rewards/margins": 1.7265625, "rewards/rejected": -0.451171875, "step": 410 }, { "epoch": 0.9889624724061811, "grad_norm": 3.135815382003784, "learning_rate": 2.6645962732919255e-06, "logits/chosen": -4.375, "logits/rejected": -4.4375, "logps/chosen": -628.0, "logps/rejected": -544.0, "loss": 0.3687, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": 1.46875, "rewards/margins": 1.671875, "rewards/rejected": -0.1982421875, "step": 420 }, { "epoch": 1.0141280353200883, "grad_norm": 1.4070848226547241, "learning_rate": 2.6024844720496896e-06, "logits/chosen": -4.40625, "logits/rejected": -4.4375, "logps/chosen": -596.0, "logps/rejected": -544.0, "loss": 0.3031, "rewards/accuracies": 0.8695651888847351, "rewards/chosen": 1.4140625, "rewards/margins": 1.8203125, "rewards/rejected": -0.40234375, "step": 430 }, { "epoch": 1.0376747608535688, "grad_norm": 1.7027989625930786, "learning_rate": 2.5403726708074537e-06, "logits/chosen": -4.375, "logits/rejected": -4.40625, "logps/chosen": -584.0, "logps/rejected": -502.0, "loss": 0.2702, "rewards/accuracies": 0.9312499761581421, "rewards/chosen": 1.578125, "rewards/margins": 2.21875, "rewards/rejected": -0.64453125, "step": 440 }, { "epoch": 1.0612214863870493, "grad_norm": 1.9917528629302979, "learning_rate": 2.4782608695652178e-06, "logits/chosen": -4.4375, "logits/rejected": -4.40625, "logps/chosen": -584.0, "logps/rejected": -508.0, "loss": 0.2617, "rewards/accuracies": 0.8812500238418579, "rewards/chosen": 1.53125, "rewards/margins": 2.046875, "rewards/rejected": -0.51953125, "step": 450 }, { "epoch": 1.0847682119205297, "grad_norm": 1.8857314586639404, "learning_rate": 2.4161490683229814e-06, "logits/chosen": -4.40625, "logits/rejected": -4.46875, "logps/chosen": -580.0, "logps/rejected": -524.0, "loss": 0.2539, "rewards/accuracies": 0.925000011920929, "rewards/chosen": 1.703125, "rewards/margins": 2.171875, "rewards/rejected": -0.462890625, "step": 460 }, { "epoch": 1.1083149374540102, "grad_norm": 1.778446912765503, "learning_rate": 2.3540372670807455e-06, "logits/chosen": -4.4375, "logits/rejected": -4.53125, "logps/chosen": -588.0, "logps/rejected": -472.0, "loss": 0.2617, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": 1.5703125, "rewards/margins": 2.03125, "rewards/rejected": -0.46484375, "step": 470 }, { "epoch": 1.131861662987491, "grad_norm": 1.5340267419815063, "learning_rate": 2.2919254658385096e-06, "logits/chosen": -4.34375, "logits/rejected": -4.40625, "logps/chosen": -644.0, "logps/rejected": -608.0, "loss": 0.2414, "rewards/accuracies": 0.9375, "rewards/chosen": 1.71875, "rewards/margins": 2.40625, "rewards/rejected": -0.6875, "step": 480 }, { "epoch": 1.1554083885209714, "grad_norm": 1.7361700534820557, "learning_rate": 2.2298136645962733e-06, "logits/chosen": -4.375, "logits/rejected": -4.34375, "logps/chosen": -612.0, "logps/rejected": -564.0, "loss": 0.2463, "rewards/accuracies": 0.918749988079071, "rewards/chosen": 1.5859375, "rewards/margins": 2.234375, "rewards/rejected": -0.65625, "step": 490 }, { "epoch": 1.1789551140544519, "grad_norm": 1.7070685625076294, "learning_rate": 2.1677018633540374e-06, "logits/chosen": -4.34375, "logits/rejected": -4.3125, "logps/chosen": -568.0, "logps/rejected": -528.0, "loss": 0.2385, "rewards/accuracies": 0.918749988079071, "rewards/chosen": 1.6015625, "rewards/margins": 2.09375, "rewards/rejected": -0.5, "step": 500 }, { "epoch": 1.2025018395879323, "grad_norm": 2.132335901260376, "learning_rate": 2.1055900621118015e-06, "logits/chosen": -4.4375, "logits/rejected": -4.40625, "logps/chosen": -636.0, "logps/rejected": -548.0, "loss": 0.2412, "rewards/accuracies": 0.90625, "rewards/chosen": 1.734375, "rewards/margins": 2.203125, "rewards/rejected": -0.46484375, "step": 510 }, { "epoch": 1.2260485651214128, "grad_norm": 2.0063838958740234, "learning_rate": 2.0434782608695656e-06, "logits/chosen": -4.46875, "logits/rejected": -4.40625, "logps/chosen": -576.0, "logps/rejected": -552.0, "loss": 0.2437, "rewards/accuracies": 0.918749988079071, "rewards/chosen": 1.484375, "rewards/margins": 2.171875, "rewards/rejected": -0.69140625, "step": 520 }, { "epoch": 1.2495952906548933, "grad_norm": 1.715623378753662, "learning_rate": 1.9813664596273293e-06, "logits/chosen": -4.375, "logits/rejected": -4.4375, "logps/chosen": -620.0, "logps/rejected": -556.0, "loss": 0.2371, "rewards/accuracies": 0.9125000238418579, "rewards/chosen": 1.6328125, "rewards/margins": 2.09375, "rewards/rejected": -0.466796875, "step": 530 }, { "epoch": 1.2731420161883737, "grad_norm": 1.8319090604782104, "learning_rate": 1.9192546583850934e-06, "logits/chosen": -4.375, "logits/rejected": -4.34375, "logps/chosen": -564.0, "logps/rejected": -470.0, "loss": 0.2456, "rewards/accuracies": 0.949999988079071, "rewards/chosen": 1.6015625, "rewards/margins": 2.3125, "rewards/rejected": -0.70703125, "step": 540 }, { "epoch": 1.2966887417218542, "grad_norm": 1.757511854171753, "learning_rate": 1.8571428571428573e-06, "logits/chosen": -4.4375, "logits/rejected": -4.46875, "logps/chosen": -604.0, "logps/rejected": -528.0, "loss": 0.2382, "rewards/accuracies": 0.925000011920929, "rewards/chosen": 1.7265625, "rewards/margins": 2.21875, "rewards/rejected": -0.494140625, "step": 550 }, { "epoch": 1.3202354672553347, "grad_norm": 1.6837247610092163, "learning_rate": 1.7950310559006212e-06, "logits/chosen": -4.40625, "logits/rejected": -4.40625, "logps/chosen": -616.0, "logps/rejected": -528.0, "loss": 0.2339, "rewards/accuracies": 0.887499988079071, "rewards/chosen": 1.8125, "rewards/margins": 2.203125, "rewards/rejected": -0.390625, "step": 560 }, { "epoch": 1.3437821927888154, "grad_norm": 1.5681090354919434, "learning_rate": 1.7329192546583853e-06, "logits/chosen": -4.4375, "logits/rejected": -4.4375, "logps/chosen": -572.0, "logps/rejected": -496.0, "loss": 0.2341, "rewards/accuracies": 0.9375, "rewards/chosen": 1.7890625, "rewards/margins": 2.1875, "rewards/rejected": -0.40625, "step": 570 }, { "epoch": 1.3673289183222959, "grad_norm": 1.8752034902572632, "learning_rate": 1.6708074534161492e-06, "logits/chosen": -4.5, "logits/rejected": -4.5, "logps/chosen": -532.0, "logps/rejected": -512.0, "loss": 0.2257, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": 1.671875, "rewards/margins": 2.390625, "rewards/rejected": -0.71875, "step": 580 }, { "epoch": 1.3908756438557763, "grad_norm": 1.8631969690322876, "learning_rate": 1.608695652173913e-06, "logits/chosen": -4.40625, "logits/rejected": -4.40625, "logps/chosen": -596.0, "logps/rejected": -488.0, "loss": 0.2405, "rewards/accuracies": 0.9125000238418579, "rewards/chosen": 1.671875, "rewards/margins": 2.28125, "rewards/rejected": -0.61328125, "step": 590 }, { "epoch": 1.4144223693892568, "grad_norm": 1.8408665657043457, "learning_rate": 1.5465838509316772e-06, "logits/chosen": -4.4375, "logits/rejected": -4.40625, "logps/chosen": -568.0, "logps/rejected": -508.0, "loss": 0.2346, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": 1.6796875, "rewards/margins": 2.34375, "rewards/rejected": -0.6640625, "step": 600 }, { "epoch": 1.4379690949227373, "grad_norm": 1.8455615043640137, "learning_rate": 1.484472049689441e-06, "logits/chosen": -4.5, "logits/rejected": -4.5, "logps/chosen": -632.0, "logps/rejected": -548.0, "loss": 0.2415, "rewards/accuracies": 0.925000011920929, "rewards/chosen": 1.8359375, "rewards/margins": 2.25, "rewards/rejected": -0.41015625, "step": 610 }, { "epoch": 1.4615158204562178, "grad_norm": 1.7498400211334229, "learning_rate": 1.4223602484472052e-06, "logits/chosen": -4.40625, "logits/rejected": -4.4375, "logps/chosen": -588.0, "logps/rejected": -496.0, "loss": 0.2428, "rewards/accuracies": 0.887499988079071, "rewards/chosen": 1.8125, "rewards/margins": 2.328125, "rewards/rejected": -0.515625, "step": 620 }, { "epoch": 1.4850625459896984, "grad_norm": 1.8195767402648926, "learning_rate": 1.360248447204969e-06, "logits/chosen": -4.40625, "logits/rejected": -4.4375, "logps/chosen": -616.0, "logps/rejected": -520.0, "loss": 0.2408, "rewards/accuracies": 0.9125000238418579, "rewards/chosen": 1.8125, "rewards/margins": 2.40625, "rewards/rejected": -0.59375, "step": 630 }, { "epoch": 1.508609271523179, "grad_norm": 1.84929621219635, "learning_rate": 1.2981366459627332e-06, "logits/chosen": -4.375, "logits/rejected": -4.40625, "logps/chosen": -596.0, "logps/rejected": -492.0, "loss": 0.2462, "rewards/accuracies": 0.893750011920929, "rewards/chosen": 1.5859375, "rewards/margins": 2.328125, "rewards/rejected": -0.73046875, "step": 640 }, { "epoch": 1.5321559970566594, "grad_norm": 1.943724513053894, "learning_rate": 1.236024844720497e-06, "logits/chosen": -4.5, "logits/rejected": -4.5, "logps/chosen": -624.0, "logps/rejected": -556.0, "loss": 0.2502, "rewards/accuracies": 0.918749988079071, "rewards/chosen": 1.734375, "rewards/margins": 2.296875, "rewards/rejected": -0.56640625, "step": 650 }, { "epoch": 1.5557027225901399, "grad_norm": 1.7159101963043213, "learning_rate": 1.173913043478261e-06, "logits/chosen": -4.40625, "logits/rejected": -4.4375, "logps/chosen": -536.0, "logps/rejected": -494.0, "loss": 0.2423, "rewards/accuracies": 0.90625, "rewards/chosen": 1.71875, "rewards/margins": 2.34375, "rewards/rejected": -0.62109375, "step": 660 }, { "epoch": 1.5792494481236203, "grad_norm": 1.5926669836044312, "learning_rate": 1.111801242236025e-06, "logits/chosen": -4.4375, "logits/rejected": -4.46875, "logps/chosen": -568.0, "logps/rejected": -502.0, "loss": 0.234, "rewards/accuracies": 0.887499988079071, "rewards/chosen": 1.515625, "rewards/margins": 2.03125, "rewards/rejected": -0.515625, "step": 670 }, { "epoch": 1.6027961736571008, "grad_norm": 1.9127466678619385, "learning_rate": 1.049689440993789e-06, "logits/chosen": -4.46875, "logits/rejected": -4.5, "logps/chosen": -612.0, "logps/rejected": -508.0, "loss": 0.2337, "rewards/accuracies": 0.918749988079071, "rewards/chosen": 1.9375, "rewards/margins": 2.375, "rewards/rejected": -0.4296875, "step": 680 }, { "epoch": 1.6263428991905813, "grad_norm": 1.569267749786377, "learning_rate": 9.875776397515528e-07, "logits/chosen": -4.4375, "logits/rejected": -4.4375, "logps/chosen": -596.0, "logps/rejected": -548.0, "loss": 0.2294, "rewards/accuracies": 0.90625, "rewards/chosen": 1.8984375, "rewards/margins": 2.28125, "rewards/rejected": -0.375, "step": 690 }, { "epoch": 1.6498896247240618, "grad_norm": 1.7182821035385132, "learning_rate": 9.254658385093168e-07, "logits/chosen": -4.4375, "logits/rejected": -4.4375, "logps/chosen": -632.0, "logps/rejected": -520.0, "loss": 0.2386, "rewards/accuracies": 0.9437500238418579, "rewards/chosen": 1.8046875, "rewards/margins": 2.546875, "rewards/rejected": -0.734375, "step": 700 }, { "epoch": 1.6734363502575422, "grad_norm": 1.6479544639587402, "learning_rate": 8.633540372670808e-07, "logits/chosen": -4.40625, "logits/rejected": -4.5, "logps/chosen": -576.0, "logps/rejected": -510.0, "loss": 0.2277, "rewards/accuracies": 0.9375, "rewards/chosen": 1.703125, "rewards/margins": 2.421875, "rewards/rejected": -0.71484375, "step": 710 }, { "epoch": 1.6969830757910227, "grad_norm": 1.6700725555419922, "learning_rate": 8.012422360248448e-07, "logits/chosen": -4.46875, "logits/rejected": -4.46875, "logps/chosen": -572.0, "logps/rejected": -464.0, "loss": 0.2362, "rewards/accuracies": 0.918749988079071, "rewards/chosen": 1.765625, "rewards/margins": 2.375, "rewards/rejected": -0.609375, "step": 720 }, { "epoch": 1.7205298013245032, "grad_norm": 1.8397201299667358, "learning_rate": 7.391304347826088e-07, "logits/chosen": -4.4375, "logits/rejected": -4.5, "logps/chosen": -656.0, "logps/rejected": -556.0, "loss": 0.2306, "rewards/accuracies": 0.956250011920929, "rewards/chosen": 1.875, "rewards/margins": 2.65625, "rewards/rejected": -0.77734375, "step": 730 }, { "epoch": 1.7440765268579839, "grad_norm": 1.8021209239959717, "learning_rate": 6.770186335403728e-07, "logits/chosen": -4.46875, "logits/rejected": -4.46875, "logps/chosen": -600.0, "logps/rejected": -556.0, "loss": 0.2239, "rewards/accuracies": 0.9125000238418579, "rewards/chosen": 1.9296875, "rewards/margins": 2.234375, "rewards/rejected": -0.306640625, "step": 740 }, { "epoch": 1.7676232523914643, "grad_norm": 1.8263683319091797, "learning_rate": 6.149068322981367e-07, "logits/chosen": -4.4375, "logits/rejected": -4.34375, "logps/chosen": -548.0, "logps/rejected": -516.0, "loss": 0.2322, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": 1.6328125, "rewards/margins": 2.171875, "rewards/rejected": -0.54296875, "step": 750 }, { "epoch": 1.7911699779249448, "grad_norm": 1.5601938962936401, "learning_rate": 5.527950310559007e-07, "logits/chosen": -4.46875, "logits/rejected": -4.46875, "logps/chosen": -572.0, "logps/rejected": -494.0, "loss": 0.2401, "rewards/accuracies": 0.90625, "rewards/chosen": 1.84375, "rewards/margins": 2.171875, "rewards/rejected": -0.3359375, "step": 760 }, { "epoch": 1.8147167034584253, "grad_norm": 1.8713370561599731, "learning_rate": 4.906832298136646e-07, "logits/chosen": -4.5, "logits/rejected": -4.4375, "logps/chosen": -544.0, "logps/rejected": -494.0, "loss": 0.2345, "rewards/accuracies": 0.9624999761581421, "rewards/chosen": 1.71875, "rewards/margins": 2.296875, "rewards/rejected": -0.58203125, "step": 770 }, { "epoch": 1.838263428991906, "grad_norm": 1.3841131925582886, "learning_rate": 4.285714285714286e-07, "logits/chosen": -4.40625, "logits/rejected": -4.375, "logps/chosen": -576.0, "logps/rejected": -500.0, "loss": 0.2201, "rewards/accuracies": 0.887499988079071, "rewards/chosen": 1.6640625, "rewards/margins": 2.3125, "rewards/rejected": -0.640625, "step": 780 }, { "epoch": 1.8618101545253865, "grad_norm": 1.6164636611938477, "learning_rate": 3.664596273291926e-07, "logits/chosen": -4.46875, "logits/rejected": -4.53125, "logps/chosen": -576.0, "logps/rejected": -520.0, "loss": 0.2429, "rewards/accuracies": 0.90625, "rewards/chosen": 1.6953125, "rewards/margins": 2.1875, "rewards/rejected": -0.4921875, "step": 790 }, { "epoch": 1.885356880058867, "grad_norm": 1.6874057054519653, "learning_rate": 3.0434782608695656e-07, "logits/chosen": -4.5, "logits/rejected": -4.4375, "logps/chosen": -572.0, "logps/rejected": -524.0, "loss": 0.2396, "rewards/accuracies": 0.9375, "rewards/chosen": 1.7265625, "rewards/margins": 2.234375, "rewards/rejected": -0.51171875, "step": 800 }, { "epoch": 1.9089036055923474, "grad_norm": 1.7373439073562622, "learning_rate": 2.422360248447205e-07, "logits/chosen": -4.4375, "logits/rejected": -4.46875, "logps/chosen": -588.0, "logps/rejected": -508.0, "loss": 0.2511, "rewards/accuracies": 0.925000011920929, "rewards/chosen": 1.9765625, "rewards/margins": 2.34375, "rewards/rejected": -0.37109375, "step": 810 }, { "epoch": 1.9324503311258279, "grad_norm": 1.989230990409851, "learning_rate": 1.8012422360248447e-07, "logits/chosen": -4.375, "logits/rejected": -4.34375, "logps/chosen": -592.0, "logps/rejected": -504.0, "loss": 0.2427, "rewards/accuracies": 0.8812500238418579, "rewards/chosen": 1.6640625, "rewards/margins": 2.0625, "rewards/rejected": -0.392578125, "step": 820 }, { "epoch": 1.9559970566593083, "grad_norm": 1.692975640296936, "learning_rate": 1.1801242236024847e-07, "logits/chosen": -4.40625, "logits/rejected": -4.4375, "logps/chosen": -620.0, "logps/rejected": -512.0, "loss": 0.2267, "rewards/accuracies": 0.9437500238418579, "rewards/chosen": 1.7890625, "rewards/margins": 2.53125, "rewards/rejected": -0.7421875, "step": 830 }, { "epoch": 1.9795437821927888, "grad_norm": 1.7542818784713745, "learning_rate": 5.590062111801243e-08, "logits/chosen": -4.375, "logits/rejected": -4.46875, "logps/chosen": -604.0, "logps/rejected": -536.0, "loss": 0.2425, "rewards/accuracies": 0.893750011920929, "rewards/chosen": 1.8359375, "rewards/margins": 2.078125, "rewards/rejected": -0.244140625, "step": 840 } ], "logging_steps": 10, "max_steps": 848, "num_input_tokens_seen": 0, "num_train_epochs": 2, "save_steps": 250, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 1, "trial_name": null, "trial_params": null }