| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9994767137624281, | |
| "eval_steps": 100, | |
| "global_step": 955, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0010465724751439038, | |
| "grad_norm": 9.51283368668585, | |
| "learning_rate": 5.208333333333333e-09, | |
| "logits/chosen": -3.21875, | |
| "logits/rejected": -3.21875, | |
| "logps/chosen": -250.0, | |
| "logps/rejected": -364.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.010465724751439037, | |
| "grad_norm": 9.570547962527824, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -3.25, | |
| "logits/rejected": -3.28125, | |
| "logps/chosen": -298.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.6918, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": -0.0004787445068359375, | |
| "rewards/margins": -0.000843048095703125, | |
| "rewards/rejected": 0.0003643035888671875, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.020931449502878074, | |
| "grad_norm": 9.308254903442457, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -3.078125, | |
| "logits/rejected": -3.171875, | |
| "logps/chosen": -286.0, | |
| "logps/rejected": -294.0, | |
| "loss": 0.6916, | |
| "rewards/accuracies": 0.25, | |
| "rewards/chosen": -0.0016326904296875, | |
| "rewards/margins": 0.00031280517578125, | |
| "rewards/rejected": -0.00194549560546875, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03139717425431711, | |
| "grad_norm": 8.758887635948838, | |
| "learning_rate": 1.5624999999999999e-07, | |
| "logits/chosen": -3.21875, | |
| "logits/rejected": -3.25, | |
| "logps/chosen": -294.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.6899, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": -0.00150299072265625, | |
| "rewards/margins": 0.002899169921875, | |
| "rewards/rejected": -0.00439453125, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04186289900575615, | |
| "grad_norm": 9.50313648843035, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -3.1875, | |
| "logits/rejected": -3.125, | |
| "logps/chosen": -288.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.684, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.00579833984375, | |
| "rewards/margins": 0.0184326171875, | |
| "rewards/rejected": -0.0242919921875, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.052328623757195186, | |
| "grad_norm": 8.682508739901571, | |
| "learning_rate": 2.604166666666667e-07, | |
| "logits/chosen": -3.25, | |
| "logits/rejected": -3.21875, | |
| "logps/chosen": -320.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.6706, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.025146484375, | |
| "rewards/margins": 0.0498046875, | |
| "rewards/rejected": -0.07470703125, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06279434850863422, | |
| "grad_norm": 11.261511786558849, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -3.15625, | |
| "logits/rejected": -3.171875, | |
| "logps/chosen": -296.0, | |
| "logps/rejected": -316.0, | |
| "loss": 0.6462, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.047119140625, | |
| "rewards/margins": 0.09326171875, | |
| "rewards/rejected": -0.140625, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07326007326007326, | |
| "grad_norm": 12.224158124661995, | |
| "learning_rate": 3.645833333333333e-07, | |
| "logits/chosen": -3.109375, | |
| "logits/rejected": -3.109375, | |
| "logps/chosen": -310.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.5955, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -0.0294189453125, | |
| "rewards/margins": 0.26171875, | |
| "rewards/rejected": -0.29296875, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.0837257980115123, | |
| "grad_norm": 24.94058449952297, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -3.046875, | |
| "logits/rejected": -3.078125, | |
| "logps/chosen": -326.0, | |
| "logps/rejected": -348.0, | |
| "loss": 0.5178, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.10498046875, | |
| "rewards/margins": 0.57421875, | |
| "rewards/rejected": -0.6796875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.09419152276295134, | |
| "grad_norm": 25.450706815664475, | |
| "learning_rate": 4.6874999999999996e-07, | |
| "logits/chosen": -3.0625, | |
| "logits/rejected": -3.046875, | |
| "logps/chosen": -328.0, | |
| "logps/rejected": -356.0, | |
| "loss": 0.4379, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -0.283203125, | |
| "rewards/margins": 0.7109375, | |
| "rewards/rejected": -0.9921875, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "grad_norm": 27.228385560373255, | |
| "learning_rate": 4.999732492681437e-07, | |
| "logits/chosen": -2.953125, | |
| "logits/rejected": -3.0, | |
| "logps/chosen": -350.0, | |
| "logps/rejected": -498.0, | |
| "loss": 0.3176, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -0.498046875, | |
| "rewards/margins": 1.5078125, | |
| "rewards/rejected": -2.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1151229722658294, | |
| "grad_norm": 45.59466990825789, | |
| "learning_rate": 4.996723692767926e-07, | |
| "logits/chosen": -2.96875, | |
| "logits/rejected": -2.96875, | |
| "logps/chosen": -344.0, | |
| "logps/rejected": -540.0, | |
| "loss": 0.2646, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.64453125, | |
| "rewards/margins": 1.9609375, | |
| "rewards/rejected": -2.609375, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12558869701726844, | |
| "grad_norm": 27.15030514765242, | |
| "learning_rate": 4.990375746213598e-07, | |
| "logits/chosen": -2.875, | |
| "logits/rejected": -2.84375, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -604.0, | |
| "loss": 0.2701, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.87890625, | |
| "rewards/margins": 2.1875, | |
| "rewards/rejected": -3.0625, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.1360544217687075, | |
| "grad_norm": 27.883434269977027, | |
| "learning_rate": 4.980697142834314e-07, | |
| "logits/chosen": -2.796875, | |
| "logits/rejected": -2.765625, | |
| "logps/chosen": -390.0, | |
| "logps/rejected": -588.0, | |
| "loss": 0.2224, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -0.9296875, | |
| "rewards/margins": 2.296875, | |
| "rewards/rejected": -3.21875, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.14652014652014653, | |
| "grad_norm": 29.52150895374023, | |
| "learning_rate": 4.967700826904229e-07, | |
| "logits/chosen": -2.875, | |
| "logits/rejected": -2.796875, | |
| "logps/chosen": -368.0, | |
| "logps/rejected": -672.0, | |
| "loss": 0.2261, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.95703125, | |
| "rewards/margins": 2.921875, | |
| "rewards/rejected": -3.875, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15698587127158556, | |
| "grad_norm": 45.4969620914659, | |
| "learning_rate": 4.951404179843962e-07, | |
| "logits/chosen": -2.75, | |
| "logits/rejected": -2.625, | |
| "logps/chosen": -438.0, | |
| "logps/rejected": -688.0, | |
| "loss": 0.2018, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": -1.2734375, | |
| "rewards/margins": 3.015625, | |
| "rewards/rejected": -4.28125, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.1674515960230246, | |
| "grad_norm": 32.79270513780942, | |
| "learning_rate": 4.931828996974498e-07, | |
| "logits/chosen": -2.75, | |
| "logits/rejected": -2.65625, | |
| "logps/chosen": -440.0, | |
| "logps/rejected": -700.0, | |
| "loss": 0.2027, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.1640625, | |
| "rewards/margins": 3.015625, | |
| "rewards/rejected": -4.1875, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.17791732077446362, | |
| "grad_norm": 30.949462462321232, | |
| "learning_rate": 4.909001458367866e-07, | |
| "logits/chosen": -2.671875, | |
| "logits/rejected": -2.609375, | |
| "logps/chosen": -372.0, | |
| "logps/rejected": -724.0, | |
| "loss": 0.1997, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.0703125, | |
| "rewards/margins": 3.171875, | |
| "rewards/rejected": -4.25, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.18838304552590268, | |
| "grad_norm": 26.340401250142335, | |
| "learning_rate": 4.882952093833627e-07, | |
| "logits/chosen": -2.703125, | |
| "logits/rejected": -2.578125, | |
| "logps/chosen": -382.0, | |
| "logps/rejected": -716.0, | |
| "loss": 0.1926, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -1.078125, | |
| "rewards/margins": 3.359375, | |
| "rewards/rejected": -4.4375, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.1988487702773417, | |
| "grad_norm": 18.968097604368335, | |
| "learning_rate": 4.853715742087946e-07, | |
| "logits/chosen": -2.53125, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -428.0, | |
| "logps/rejected": -752.0, | |
| "loss": 0.1612, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -1.5, | |
| "rewards/margins": 3.53125, | |
| "rewards/rejected": -5.03125, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "grad_norm": 20.782646308534286, | |
| "learning_rate": 4.821331504159906e-07, | |
| "logits/chosen": -2.6875, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -462.0, | |
| "logps/rejected": -784.0, | |
| "loss": 0.1607, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -1.5703125, | |
| "rewards/margins": 3.5, | |
| "rewards/rejected": -5.09375, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21978021978021978, | |
| "grad_norm": 23.863300215477395, | |
| "learning_rate": 4.785842691097342e-07, | |
| "logits/chosen": -2.609375, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -410.0, | |
| "logps/rejected": -820.0, | |
| "loss": 0.1596, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -1.21875, | |
| "rewards/margins": 4.25, | |
| "rewards/rejected": -5.46875, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.2302459445316588, | |
| "grad_norm": 40.35606043050439, | |
| "learning_rate": 4.7472967660421603e-07, | |
| "logits/chosen": -2.59375, | |
| "logits/rejected": -2.46875, | |
| "logps/chosen": -494.0, | |
| "logps/rejected": -932.0, | |
| "loss": 0.1764, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.9609375, | |
| "rewards/margins": 4.46875, | |
| "rewards/rejected": -6.4375, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.24071166928309787, | |
| "grad_norm": 25.181446808734336, | |
| "learning_rate": 4.705745280752585e-07, | |
| "logits/chosen": -2.6875, | |
| "logits/rejected": -2.515625, | |
| "logps/chosen": -442.0, | |
| "logps/rejected": -832.0, | |
| "loss": 0.1588, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.46875, | |
| "rewards/margins": 4.0625, | |
| "rewards/rejected": -5.53125, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.25117739403453687, | |
| "grad_norm": 32.49967985504463, | |
| "learning_rate": 4.6612438066572555e-07, | |
| "logits/chosen": -2.4375, | |
| "logits/rejected": -2.1875, | |
| "logps/chosen": -480.0, | |
| "logps/rejected": -940.0, | |
| "loss": 0.16, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.6875, | |
| "rewards/margins": 4.78125, | |
| "rewards/rejected": -6.4375, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.2616431187859759, | |
| "grad_norm": 31.230972977495156, | |
| "learning_rate": 4.6138518605333664e-07, | |
| "logits/chosen": -2.5625, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -426.0, | |
| "logps/rejected": -800.0, | |
| "loss": 0.1537, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -1.421875, | |
| "rewards/margins": 4.0625, | |
| "rewards/rejected": -5.46875, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.272108843537415, | |
| "grad_norm": 29.71689511008149, | |
| "learning_rate": 4.5636328249082514e-07, | |
| "logits/chosen": -2.453125, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -474.0, | |
| "logps/rejected": -948.0, | |
| "loss": 0.1394, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.828125, | |
| "rewards/margins": 4.625, | |
| "rewards/rejected": -6.4375, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.282574568288854, | |
| "grad_norm": 28.295583463414996, | |
| "learning_rate": 4.510653863290871e-07, | |
| "logits/chosen": -2.40625, | |
| "logits/rejected": -2.078125, | |
| "logps/chosen": -438.0, | |
| "logps/rejected": -896.0, | |
| "loss": 0.1481, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.390625, | |
| "rewards/margins": 4.84375, | |
| "rewards/rejected": -6.25, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.29304029304029305, | |
| "grad_norm": 28.398306620399453, | |
| "learning_rate": 4.4549858303465737e-07, | |
| "logits/chosen": -2.5, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -484.0, | |
| "logps/rejected": -944.0, | |
| "loss": 0.1392, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.9453125, | |
| "rewards/margins": 4.5625, | |
| "rewards/rejected": -6.5, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.3035060177917321, | |
| "grad_norm": 40.00756374405055, | |
| "learning_rate": 4.396703177135261e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.109375, | |
| "logps/chosen": -544.0, | |
| "logps/rejected": -1072.0, | |
| "loss": 0.1322, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -2.734375, | |
| "rewards/margins": 5.53125, | |
| "rewards/rejected": -8.25, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "grad_norm": 37.30671040470767, | |
| "learning_rate": 4.335883851539693e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.015625, | |
| "logps/chosen": -680.0, | |
| "logps/rejected": -1200.0, | |
| "loss": 0.1326, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -3.65625, | |
| "rewards/margins": 5.46875, | |
| "rewards/rejected": -9.125, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32443746729461015, | |
| "grad_norm": 34.70455152314579, | |
| "learning_rate": 4.272609194017105e-07, | |
| "logits/chosen": -2.21875, | |
| "logits/rejected": -1.7265625, | |
| "logps/chosen": -832.0, | |
| "logps/rejected": -1368.0, | |
| "loss": 0.1371, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -5.25, | |
| "rewards/margins": 5.40625, | |
| "rewards/rejected": -10.625, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.3349031920460492, | |
| "grad_norm": 24.846518493181428, | |
| "learning_rate": 4.2069638288135547e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.109375, | |
| "logps/chosen": -696.0, | |
| "logps/rejected": -1264.0, | |
| "loss": 0.1284, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -3.9375, | |
| "rewards/margins": 5.71875, | |
| "rewards/rejected": -9.625, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.3453689167974882, | |
| "grad_norm": 45.84342601245407, | |
| "learning_rate": 4.139035550786494e-07, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -684.0, | |
| "logps/rejected": -1200.0, | |
| "loss": 0.1228, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -4.03125, | |
| "rewards/margins": 5.0625, | |
| "rewards/rejected": -9.125, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.35583464154892724, | |
| "grad_norm": 49.94021964049473, | |
| "learning_rate": 4.0689152079869306e-07, | |
| "logits/chosen": -2.21875, | |
| "logits/rejected": -1.8515625, | |
| "logps/chosen": -740.0, | |
| "logps/rejected": -1456.0, | |
| "loss": 0.1026, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 7.34375, | |
| "rewards/rejected": -12.1875, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.3663003663003663, | |
| "grad_norm": 39.72539996550781, | |
| "learning_rate": 3.99669658015821e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -1.9609375, | |
| "logps/chosen": -696.0, | |
| "logps/rejected": -1328.0, | |
| "loss": 0.1202, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -4.375, | |
| "rewards/margins": 6.59375, | |
| "rewards/rejected": -10.9375, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.37676609105180536, | |
| "grad_norm": 45.84961303802077, | |
| "learning_rate": 3.92247625331392e-07, | |
| "logits/chosen": -2.21875, | |
| "logits/rejected": -2.0, | |
| "logps/chosen": -764.0, | |
| "logps/rejected": -1384.0, | |
| "loss": 0.1073, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -4.71875, | |
| "rewards/margins": 6.40625, | |
| "rewards/rejected": -11.125, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.3872318158032444, | |
| "grad_norm": 20.54553148683269, | |
| "learning_rate": 3.846353490562664e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -1.796875, | |
| "logps/chosen": -680.0, | |
| "logps/rejected": -1360.0, | |
| "loss": 0.1308, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -4.1875, | |
| "rewards/margins": 6.90625, | |
| "rewards/rejected": -11.125, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.3976975405546834, | |
| "grad_norm": 30.114773714834385, | |
| "learning_rate": 3.768430099352445e-07, | |
| "logits/chosen": -2.4375, | |
| "logits/rejected": -2.234375, | |
| "logps/chosen": -668.0, | |
| "logps/rejected": -1200.0, | |
| "loss": 0.1285, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -3.609375, | |
| "rewards/margins": 5.53125, | |
| "rewards/rejected": -9.125, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.40816326530612246, | |
| "grad_norm": 22.860559479058303, | |
| "learning_rate": 3.6888102953122304e-07, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.15625, | |
| "logps/chosen": -652.0, | |
| "logps/rejected": -1272.0, | |
| "loss": 0.1018, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": -3.609375, | |
| "rewards/margins": 6.5, | |
| "rewards/rejected": -10.125, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "grad_norm": 19.494795072649943, | |
| "learning_rate": 3.607600562872785e-07, | |
| "logits/chosen": -2.4375, | |
| "logits/rejected": -2.03125, | |
| "logps/chosen": -892.0, | |
| "logps/rejected": -1640.0, | |
| "loss": 0.091, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -5.46875, | |
| "rewards/margins": 7.90625, | |
| "rewards/rejected": -13.375, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4290947148090005, | |
| "grad_norm": 34.802523467101764, | |
| "learning_rate": 3.5249095128531856e-07, | |
| "logits/chosen": -2.453125, | |
| "logits/rejected": -2.203125, | |
| "logps/chosen": -812.0, | |
| "logps/rejected": -1392.0, | |
| "loss": 0.1044, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 6.4375, | |
| "rewards/rejected": -11.1875, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.43956043956043955, | |
| "grad_norm": 16.064797665253533, | |
| "learning_rate": 3.4408477372034736e-07, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.03125, | |
| "logps/chosen": -856.0, | |
| "logps/rejected": -1536.0, | |
| "loss": 0.1125, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -5.40625, | |
| "rewards/margins": 7.375, | |
| "rewards/rejected": -12.75, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.4500261643118786, | |
| "grad_norm": 26.48155910496492, | |
| "learning_rate": 3.3555276610977276e-07, | |
| "logits/chosen": -2.40625, | |
| "logits/rejected": -2.171875, | |
| "logps/chosen": -776.0, | |
| "logps/rejected": -1456.0, | |
| "loss": 0.0939, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -4.90625, | |
| "rewards/margins": 6.78125, | |
| "rewards/rejected": -11.6875, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.4604918890633176, | |
| "grad_norm": 20.80477653317148, | |
| "learning_rate": 3.269063392575352e-07, | |
| "logits/chosen": -2.4375, | |
| "logits/rejected": -2.1875, | |
| "logps/chosen": -748.0, | |
| "logps/rejected": -1472.0, | |
| "loss": 0.1046, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 7.21875, | |
| "rewards/rejected": -12.0, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.47095761381475665, | |
| "grad_norm": 29.44200512203946, | |
| "learning_rate": 3.1815705699316964e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.0625, | |
| "logps/chosen": -792.0, | |
| "logps/rejected": -1496.0, | |
| "loss": 0.0963, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": -4.90625, | |
| "rewards/margins": 7.125, | |
| "rewards/rejected": -12.0625, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.48142333856619574, | |
| "grad_norm": 26.271372385880735, | |
| "learning_rate": 3.0931662070620794e-07, | |
| "logits/chosen": -2.203125, | |
| "logits/rejected": -2.0, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -1608.0, | |
| "loss": 0.0973, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -6.34375, | |
| "rewards/margins": 7.1875, | |
| "rewards/rejected": -13.5625, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.49188906331763477, | |
| "grad_norm": 31.154091413440536, | |
| "learning_rate": 3.003968536966078e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.109375, | |
| "logps/chosen": -868.0, | |
| "logps/rejected": -1600.0, | |
| "loss": 0.0846, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -5.53125, | |
| "rewards/margins": 7.40625, | |
| "rewards/rejected": -12.9375, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5023547880690737, | |
| "grad_norm": 25.37606130345961, | |
| "learning_rate": 2.9140968536213693e-07, | |
| "logits/chosen": -2.5, | |
| "logits/rejected": -2.125, | |
| "logps/chosen": -812.0, | |
| "logps/rejected": -1504.0, | |
| "loss": 0.0895, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -5.0625, | |
| "rewards/margins": 7.25, | |
| "rewards/rejected": -12.3125, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.5128205128205128, | |
| "grad_norm": 27.895788314669506, | |
| "learning_rate": 2.823671352438608e-07, | |
| "logits/chosen": -2.515625, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -820.0, | |
| "logps/rejected": -1488.0, | |
| "loss": 0.0997, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -5.34375, | |
| "rewards/margins": 6.8125, | |
| "rewards/rejected": -12.1875, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "grad_norm": 31.69949419586544, | |
| "learning_rate": 2.73281296951072e-07, | |
| "logits/chosen": -2.453125, | |
| "logits/rejected": -2.03125, | |
| "logps/chosen": -952.0, | |
| "logps/rejected": -1672.0, | |
| "loss": 0.0871, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -6.5625, | |
| "rewards/margins": 7.59375, | |
| "rewards/rejected": -14.125, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.533751962323391, | |
| "grad_norm": 18.91909774862979, | |
| "learning_rate": 2.641643219871597e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.046875, | |
| "logps/chosen": -824.0, | |
| "logps/rejected": -1472.0, | |
| "loss": 0.0982, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -5.1875, | |
| "rewards/margins": 6.84375, | |
| "rewards/rejected": -12.0, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.54421768707483, | |
| "grad_norm": 31.301196837566955, | |
| "learning_rate": 2.550284034980507e-07, | |
| "logits/chosen": -2.46875, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -736.0, | |
| "logps/rejected": -1368.0, | |
| "loss": 0.1158, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -4.34375, | |
| "rewards/margins": 6.5, | |
| "rewards/rejected": -10.8125, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.554683411826269, | |
| "grad_norm": 35.72913678676417, | |
| "learning_rate": 2.4588575996495794e-07, | |
| "logits/chosen": -2.640625, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -684.0, | |
| "logps/rejected": -1336.0, | |
| "loss": 0.1081, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 6.78125, | |
| "rewards/rejected": -10.9375, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.565149136577708, | |
| "grad_norm": 22.499329779979547, | |
| "learning_rate": 2.367486188632446e-07, | |
| "logits/chosen": -2.609375, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -668.0, | |
| "logps/rejected": -1336.0, | |
| "loss": 0.0892, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.84375, | |
| "rewards/margins": 6.65625, | |
| "rewards/rejected": -10.5, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.5756148613291471, | |
| "grad_norm": 18.90148104971076, | |
| "learning_rate": 2.276292003092593e-07, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.203125, | |
| "logps/chosen": -816.0, | |
| "logps/rejected": -1520.0, | |
| "loss": 0.0834, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -5.28125, | |
| "rewards/margins": 7.28125, | |
| "rewards/rejected": -12.5625, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5860805860805861, | |
| "grad_norm": 34.885386057496255, | |
| "learning_rate": 2.185397007170141e-07, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.0625, | |
| "logps/chosen": -952.0, | |
| "logps/rejected": -1680.0, | |
| "loss": 0.0802, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": -6.71875, | |
| "rewards/margins": 7.75, | |
| "rewards/rejected": -14.4375, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.5965463108320251, | |
| "grad_norm": 31.26740690349814, | |
| "learning_rate": 2.094922764865619e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.21875, | |
| "logps/chosen": -892.0, | |
| "logps/rejected": -1640.0, | |
| "loss": 0.073, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -5.9375, | |
| "rewards/margins": 7.71875, | |
| "rewards/rejected": -13.625, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.6070120355834642, | |
| "grad_norm": 26.503557168812993, | |
| "learning_rate": 2.0049902774588797e-07, | |
| "logits/chosen": -2.4375, | |
| "logits/rejected": -2.125, | |
| "logps/chosen": -892.0, | |
| "logps/rejected": -1680.0, | |
| "loss": 0.0786, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -6.1875, | |
| "rewards/margins": 8.0, | |
| "rewards/rejected": -14.25, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.6174777603349032, | |
| "grad_norm": 26.760637852696387, | |
| "learning_rate": 1.9157198216806238e-07, | |
| "logits/chosen": -2.453125, | |
| "logits/rejected": -2.09375, | |
| "logps/chosen": -888.0, | |
| "logps/rejected": -1736.0, | |
| "loss": 0.0796, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -6.125, | |
| "rewards/margins": 8.4375, | |
| "rewards/rejected": -14.5625, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "grad_norm": 33.575414387335336, | |
| "learning_rate": 1.8272307888529274e-07, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -1776.0, | |
| "loss": 0.0901, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -5.9375, | |
| "rewards/margins": 8.5, | |
| "rewards/rejected": -14.4375, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6384092098377813, | |
| "grad_norm": 30.280173464201145, | |
| "learning_rate": 1.7396415252139288e-07, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.171875, | |
| "logps/chosen": -820.0, | |
| "logps/rejected": -1584.0, | |
| "loss": 0.0848, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -5.28125, | |
| "rewards/margins": 7.78125, | |
| "rewards/rejected": -13.0625, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.6488749345892203, | |
| "grad_norm": 31.809087141096686, | |
| "learning_rate": 1.6530691736402316e-07, | |
| "logits/chosen": -2.515625, | |
| "logits/rejected": -2.25, | |
| "logps/chosen": -828.0, | |
| "logps/rejected": -1576.0, | |
| "loss": 0.0699, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -5.375, | |
| "rewards/margins": 7.6875, | |
| "rewards/rejected": -13.0625, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.6593406593406593, | |
| "grad_norm": 20.492927408780787, | |
| "learning_rate": 1.5676295169786864e-07, | |
| "logits/chosen": -2.515625, | |
| "logits/rejected": -2.203125, | |
| "logps/chosen": -928.0, | |
| "logps/rejected": -1608.0, | |
| "loss": 0.0906, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": -6.1875, | |
| "rewards/margins": 7.15625, | |
| "rewards/rejected": -13.375, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.6698063840920984, | |
| "grad_norm": 24.302749821696764, | |
| "learning_rate": 1.483436823197092e-07, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -916.0, | |
| "logps/rejected": -1680.0, | |
| "loss": 0.0811, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -6.375, | |
| "rewards/margins": 7.78125, | |
| "rewards/rejected": -14.1875, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.6802721088435374, | |
| "grad_norm": 30.640711717907518, | |
| "learning_rate": 1.4006036925609243e-07, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -984.0, | |
| "logps/rejected": -1704.0, | |
| "loss": 0.0736, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -6.8125, | |
| "rewards/margins": 7.53125, | |
| "rewards/rejected": -14.3125, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6907378335949764, | |
| "grad_norm": 28.697044539390294, | |
| "learning_rate": 1.319240907040458e-07, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.125, | |
| "logps/chosen": -960.0, | |
| "logps/rejected": -1752.0, | |
| "loss": 0.0737, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -6.59375, | |
| "rewards/margins": 7.84375, | |
| "rewards/rejected": -14.4375, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.7012035583464155, | |
| "grad_norm": 17.345360984589735, | |
| "learning_rate": 1.2394572821496948e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.03125, | |
| "logps/chosen": -988.0, | |
| "logps/rejected": -1848.0, | |
| "loss": 0.0695, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -7.15625, | |
| "rewards/margins": 8.625, | |
| "rewards/rejected": -15.75, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.7116692830978545, | |
| "grad_norm": 33.23147099676755, | |
| "learning_rate": 1.1613595214152711e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.09375, | |
| "logps/chosen": -1064.0, | |
| "logps/rejected": -1880.0, | |
| "loss": 0.0806, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -7.53125, | |
| "rewards/margins": 8.4375, | |
| "rewards/rejected": -15.9375, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.7221350078492935, | |
| "grad_norm": 20.63657009936635, | |
| "learning_rate": 1.0850520736699362e-07, | |
| "logits/chosen": -2.515625, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -1004.0, | |
| "logps/rejected": -1792.0, | |
| "loss": 0.08, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -6.875, | |
| "rewards/margins": 8.0, | |
| "rewards/rejected": -14.875, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "grad_norm": 18.557890534440453, | |
| "learning_rate": 1.0106369933615042e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.234375, | |
| "logps/chosen": -960.0, | |
| "logps/rejected": -1840.0, | |
| "loss": 0.0614, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -6.5625, | |
| "rewards/margins": 8.75, | |
| "rewards/rejected": -15.25, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7430664573521716, | |
| "grad_norm": 17.0748564532371, | |
| "learning_rate": 9.382138040640714e-08, | |
| "logits/chosen": -2.53125, | |
| "logits/rejected": -2.15625, | |
| "logps/chosen": -960.0, | |
| "logps/rejected": -1888.0, | |
| "loss": 0.0656, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -6.71875, | |
| "rewards/margins": 9.125, | |
| "rewards/rejected": -15.875, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.7535321821036107, | |
| "grad_norm": 28.894806455872907, | |
| "learning_rate": 8.678793653740632e-08, | |
| "logits/chosen": -2.5, | |
| "logits/rejected": -2.203125, | |
| "logps/chosen": -1040.0, | |
| "logps/rejected": -1840.0, | |
| "loss": 0.0632, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -7.0625, | |
| "rewards/margins": 8.4375, | |
| "rewards/rejected": -15.5, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.7639979068550498, | |
| "grad_norm": 26.829195320759613, | |
| "learning_rate": 7.997277433690983e-08, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.15625, | |
| "logps/chosen": -1024.0, | |
| "logps/rejected": -1872.0, | |
| "loss": 0.0744, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -7.46875, | |
| "rewards/margins": 8.5625, | |
| "rewards/rejected": -16.0, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.7744636316064888, | |
| "grad_norm": 19.48203742338746, | |
| "learning_rate": 7.338500848029602e-08, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.0625, | |
| "logps/chosen": -1056.0, | |
| "logps/rejected": -1976.0, | |
| "loss": 0.0672, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -7.53125, | |
| "rewards/margins": 9.4375, | |
| "rewards/rejected": -17.0, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.7849293563579278, | |
| "grad_norm": 14.527749472798101, | |
| "learning_rate": 6.70334495204884e-08, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -1.9453125, | |
| "logps/chosen": -996.0, | |
| "logps/rejected": -1920.0, | |
| "loss": 0.0481, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -7.375, | |
| "rewards/margins": 9.375, | |
| "rewards/rejected": -16.75, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7953950811093669, | |
| "grad_norm": 15.609318913343236, | |
| "learning_rate": 6.092659210462231e-08, | |
| "logits/chosen": -2.40625, | |
| "logits/rejected": -1.9921875, | |
| "logps/chosen": -1056.0, | |
| "logps/rejected": -1976.0, | |
| "loss": 0.0532, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -7.78125, | |
| "rewards/margins": 9.4375, | |
| "rewards/rejected": -17.25, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.8058608058608059, | |
| "grad_norm": 7.809802817809847, | |
| "learning_rate": 5.507260361320737e-08, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -1.96875, | |
| "logps/chosen": -1056.0, | |
| "logps/rejected": -2064.0, | |
| "loss": 0.0419, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": -7.6875, | |
| "rewards/margins": 10.25, | |
| "rewards/rejected": -17.875, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.8163265306122449, | |
| "grad_norm": 22.462937925254167, | |
| "learning_rate": 4.947931323697982e-08, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.015625, | |
| "logps/chosen": -1120.0, | |
| "logps/rejected": -2064.0, | |
| "loss": 0.0616, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -8.125, | |
| "rewards/margins": 9.75, | |
| "rewards/rejected": -17.875, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.826792255363684, | |
| "grad_norm": 25.68320010476297, | |
| "learning_rate": 4.415420150605398e-08, | |
| "logits/chosen": -2.171875, | |
| "logits/rejected": -1.890625, | |
| "logps/chosen": -1048.0, | |
| "logps/rejected": -2040.0, | |
| "loss": 0.0657, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -7.8125, | |
| "rewards/margins": 9.8125, | |
| "rewards/rejected": -17.625, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "grad_norm": 26.750079370530244, | |
| "learning_rate": 3.9104390285376374e-08, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -1.9765625, | |
| "logps/chosen": -1040.0, | |
| "logps/rejected": -2016.0, | |
| "loss": 0.0513, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -7.5, | |
| "rewards/margins": 9.6875, | |
| "rewards/rejected": -17.125, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.847723704866562, | |
| "grad_norm": 15.408792175860983, | |
| "learning_rate": 3.433663324986208e-08, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.015625, | |
| "logps/chosen": -1024.0, | |
| "logps/rejected": -1936.0, | |
| "loss": 0.0679, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -7.28125, | |
| "rewards/margins": 9.0625, | |
| "rewards/rejected": -16.375, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.858189429618001, | |
| "grad_norm": 16.721851816987886, | |
| "learning_rate": 2.9857306851953897e-08, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -1.875, | |
| "logps/chosen": -1032.0, | |
| "logps/rejected": -1960.0, | |
| "loss": 0.0581, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -7.59375, | |
| "rewards/margins": 9.5, | |
| "rewards/rejected": -17.125, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.8686551543694401, | |
| "grad_norm": 25.788818922540266, | |
| "learning_rate": 2.567240179368185e-08, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -1.75, | |
| "logps/chosen": -1072.0, | |
| "logps/rejected": -2064.0, | |
| "loss": 0.052, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -7.6875, | |
| "rewards/margins": 10.3125, | |
| "rewards/rejected": -18.0, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.8791208791208791, | |
| "grad_norm": 14.284086558067754, | |
| "learning_rate": 2.1787515014630357e-08, | |
| "logits/chosen": -2.203125, | |
| "logits/rejected": -1.875, | |
| "logps/chosen": -1040.0, | |
| "logps/rejected": -2128.0, | |
| "loss": 0.0524, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -7.75, | |
| "rewards/margins": 10.75, | |
| "rewards/rejected": -18.5, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.8895866038723181, | |
| "grad_norm": 30.104620767056684, | |
| "learning_rate": 1.820784220652766e-08, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -1.9375, | |
| "logps/chosen": -1088.0, | |
| "logps/rejected": -2112.0, | |
| "loss": 0.0408, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": -7.84375, | |
| "rewards/margins": 10.4375, | |
| "rewards/rejected": -18.25, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9000523286237572, | |
| "grad_norm": 30.017914965835946, | |
| "learning_rate": 1.4938170864468636e-08, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -1.75, | |
| "logps/chosen": -1072.0, | |
| "logps/rejected": -2192.0, | |
| "loss": 0.0575, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -8.0625, | |
| "rewards/margins": 11.25, | |
| "rewards/rejected": -19.375, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.9105180533751962, | |
| "grad_norm": 22.849493479324984, | |
| "learning_rate": 1.1982873884064465e-08, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -1.875, | |
| "logps/chosen": -1120.0, | |
| "logps/rejected": -2080.0, | |
| "loss": 0.0585, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": -8.125, | |
| "rewards/margins": 10.0625, | |
| "rewards/rejected": -18.125, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.9209837781266352, | |
| "grad_norm": 13.839322356316236, | |
| "learning_rate": 9.345903713082304e-09, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -1184.0, | |
| "logps/rejected": -2208.0, | |
| "loss": 0.0595, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -8.5, | |
| "rewards/margins": 10.5625, | |
| "rewards/rejected": -19.125, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.9314495028780743, | |
| "grad_norm": 17.72445198042151, | |
| "learning_rate": 7.030787065396865e-09, | |
| "logits/chosen": -2.21875, | |
| "logits/rejected": -1.7578125, | |
| "logps/chosen": -1080.0, | |
| "logps/rejected": -2176.0, | |
| "loss": 0.0598, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -7.84375, | |
| "rewards/margins": 11.25, | |
| "rewards/rejected": -19.125, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "grad_norm": 7.302557618391421, | |
| "learning_rate": 5.04062020432286e-09, | |
| "logits/chosen": -2.40625, | |
| "logits/rejected": -1.859375, | |
| "logps/chosen": -1112.0, | |
| "logps/rejected": -2128.0, | |
| "loss": 0.0641, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -8.0, | |
| "rewards/margins": 10.4375, | |
| "rewards/rejected": -18.5, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9523809523809523, | |
| "grad_norm": 26.036561376029788, | |
| "learning_rate": 3.3780648016376866e-09, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -1.890625, | |
| "logps/chosen": -1120.0, | |
| "logps/rejected": -2064.0, | |
| "loss": 0.0606, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -8.0625, | |
| "rewards/margins": 9.6875, | |
| "rewards/rejected": -17.75, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.9628466771323915, | |
| "grad_norm": 42.338882753255156, | |
| "learning_rate": 2.0453443778310766e-09, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -1.8203125, | |
| "logps/chosen": -1048.0, | |
| "logps/rejected": -2032.0, | |
| "loss": 0.0514, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -7.59375, | |
| "rewards/margins": 10.0625, | |
| "rewards/rejected": -17.625, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.9733124018838305, | |
| "grad_norm": 25.61689367855155, | |
| "learning_rate": 1.0442413283435758e-09, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -1.7265625, | |
| "logps/chosen": -1040.0, | |
| "logps/rejected": -1992.0, | |
| "loss": 0.048, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -7.78125, | |
| "rewards/margins": 9.625, | |
| "rewards/rejected": -17.375, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.9837781266352695, | |
| "grad_norm": 23.291401299662382, | |
| "learning_rate": 3.760945397705828e-10, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -1.96875, | |
| "logps/chosen": -1064.0, | |
| "logps/rejected": -2096.0, | |
| "loss": 0.0548, | |
| "rewards/accuracies": 0.96875, | |
| "rewards/chosen": -7.5625, | |
| "rewards/margins": 10.375, | |
| "rewards/rejected": -17.875, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.9942438513867086, | |
| "grad_norm": 25.219313507903824, | |
| "learning_rate": 4.17975992204056e-11, | |
| "logits/chosen": -2.21875, | |
| "logits/rejected": -1.8125, | |
| "logps/chosen": -1096.0, | |
| "logps/rejected": -2128.0, | |
| "loss": 0.0409, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": -7.84375, | |
| "rewards/margins": 10.5, | |
| "rewards/rejected": -18.375, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.9994767137624281, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 0.15506600241386453, | |
| "train_runtime": 14070.209, | |
| "train_samples_per_second": 8.69, | |
| "train_steps_per_second": 0.068 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 955, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "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": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |