| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 478, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0020920502092050207, | |
| "grad_norm": 9.06575721766735, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": -3.375, | |
| "logits/rejected": -3.3125, | |
| "logps/chosen": -304.0, | |
| "logps/rejected": -370.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02092050209205021, | |
| "grad_norm": 8.259315535138102, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -3.171875, | |
| "logits/rejected": -3.1875, | |
| "logps/chosen": -268.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.6919, | |
| "rewards/accuracies": 0.2638888955116272, | |
| "rewards/chosen": 0.00160980224609375, | |
| "rewards/margins": 0.0011749267578125, | |
| "rewards/rejected": 0.00043487548828125, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04184100418410042, | |
| "grad_norm": 8.69903088939775, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -3.21875, | |
| "logits/rejected": -3.21875, | |
| "logps/chosen": -292.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.6913, | |
| "rewards/accuracies": 0.4000000059604645, | |
| "rewards/chosen": -0.00118255615234375, | |
| "rewards/margins": 0.0023956298828125, | |
| "rewards/rejected": -0.0035858154296875, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06276150627615062, | |
| "grad_norm": 9.20861373072528, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -3.25, | |
| "logits/rejected": -3.21875, | |
| "logps/chosen": -306.0, | |
| "logps/rejected": -302.0, | |
| "loss": 0.6844, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.00836181640625, | |
| "rewards/margins": 0.01519775390625, | |
| "rewards/rejected": -0.0235595703125, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08368200836820083, | |
| "grad_norm": 9.062167569963929, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -3.09375, | |
| "logits/rejected": -3.1875, | |
| "logps/chosen": -276.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.6644, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -0.03271484375, | |
| "rewards/margins": 0.06884765625, | |
| "rewards/rejected": -0.1015625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10460251046025104, | |
| "grad_norm": 11.43153835170189, | |
| "learning_rate": 4.999733114418725e-07, | |
| "logits/chosen": -3.109375, | |
| "logits/rejected": -3.046875, | |
| "logps/chosen": -298.0, | |
| "logps/rejected": -322.0, | |
| "loss": 0.6197, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.0439453125, | |
| "rewards/margins": 0.1572265625, | |
| "rewards/rejected": -0.201171875, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12552301255230125, | |
| "grad_norm": 17.81032056097967, | |
| "learning_rate": 4.990398100856366e-07, | |
| "logits/chosen": -3.171875, | |
| "logits/rejected": -3.140625, | |
| "logps/chosen": -288.0, | |
| "logps/rejected": -350.0, | |
| "loss": 0.5247, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.1259765625, | |
| "rewards/margins": 0.388671875, | |
| "rewards/rejected": -0.51171875, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14644351464435146, | |
| "grad_norm": 25.618554799822828, | |
| "learning_rate": 4.967775735898179e-07, | |
| "logits/chosen": -2.96875, | |
| "logits/rejected": -2.984375, | |
| "logps/chosen": -304.0, | |
| "logps/rejected": -400.0, | |
| "loss": 0.4244, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": -0.322265625, | |
| "rewards/margins": 0.86328125, | |
| "rewards/rejected": -1.1875, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16736401673640167, | |
| "grad_norm": 25.84470141626761, | |
| "learning_rate": 4.931986719649298e-07, | |
| "logits/chosen": -3.09375, | |
| "logits/rejected": -3.046875, | |
| "logps/chosen": -396.0, | |
| "logps/rejected": -506.0, | |
| "loss": 0.3205, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.60546875, | |
| "rewards/margins": 1.515625, | |
| "rewards/rejected": -2.109375, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18828451882845187, | |
| "grad_norm": 25.014757375528426, | |
| "learning_rate": 4.883222001996351e-07, | |
| "logits/chosen": -3.0, | |
| "logits/rejected": -2.96875, | |
| "logps/chosen": -346.0, | |
| "logps/rejected": -516.0, | |
| "loss": 0.29, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -0.7734375, | |
| "rewards/margins": 1.875, | |
| "rewards/rejected": -2.640625, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20920502092050208, | |
| "grad_norm": 51.461518247004356, | |
| "learning_rate": 4.821741763807186e-07, | |
| "logits/chosen": -2.859375, | |
| "logits/rejected": -2.90625, | |
| "logps/chosen": -350.0, | |
| "logps/rejected": -548.0, | |
| "loss": 0.2889, | |
| "rewards/accuracies": 0.8687499761581421, | |
| "rewards/chosen": -0.77734375, | |
| "rewards/margins": 2.109375, | |
| "rewards/rejected": -2.890625, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2301255230125523, | |
| "grad_norm": 23.843663606669114, | |
| "learning_rate": 4.747874028753375e-07, | |
| "logits/chosen": -2.90625, | |
| "logits/rejected": -2.8125, | |
| "logps/chosen": -414.0, | |
| "logps/rejected": -604.0, | |
| "loss": 0.2908, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": -1.015625, | |
| "rewards/margins": 2.078125, | |
| "rewards/rejected": -3.09375, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.2510460251046025, | |
| "grad_norm": 31.907707809752832, | |
| "learning_rate": 4.662012913161997e-07, | |
| "logits/chosen": -2.796875, | |
| "logits/rejected": -2.71875, | |
| "logps/chosen": -400.0, | |
| "logps/rejected": -592.0, | |
| "loss": 0.2526, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.0703125, | |
| "rewards/margins": 2.0625, | |
| "rewards/rejected": -3.140625, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2719665271966527, | |
| "grad_norm": 30.72791174599876, | |
| "learning_rate": 4.5646165232345103e-07, | |
| "logits/chosen": -2.859375, | |
| "logits/rejected": -2.765625, | |
| "logps/chosen": -370.0, | |
| "logps/rejected": -596.0, | |
| "loss": 0.2332, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -0.84765625, | |
| "rewards/margins": 2.3125, | |
| "rewards/rejected": -3.15625, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.2928870292887029, | |
| "grad_norm": 23.82206952691506, | |
| "learning_rate": 4.456204510851956e-07, | |
| "logits/chosen": -2.765625, | |
| "logits/rejected": -2.703125, | |
| "logps/chosen": -438.0, | |
| "logps/rejected": -676.0, | |
| "loss": 0.2309, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.1796875, | |
| "rewards/margins": 2.671875, | |
| "rewards/rejected": -3.84375, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.3138075313807531, | |
| "grad_norm": 38.02376301991461, | |
| "learning_rate": 4.337355301007335e-07, | |
| "logits/chosen": -2.75, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -418.0, | |
| "logps/rejected": -716.0, | |
| "loss": 0.2268, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -1.265625, | |
| "rewards/margins": 3.046875, | |
| "rewards/rejected": -4.3125, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33472803347280333, | |
| "grad_norm": 35.00032097537842, | |
| "learning_rate": 4.2087030056579986e-07, | |
| "logits/chosen": -2.734375, | |
| "logits/rejected": -2.625, | |
| "logps/chosen": -406.0, | |
| "logps/rejected": -712.0, | |
| "loss": 0.2075, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -1.2734375, | |
| "rewards/margins": 3.078125, | |
| "rewards/rejected": -4.34375, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.35564853556485354, | |
| "grad_norm": 32.10425911100618, | |
| "learning_rate": 4.070934040463998e-07, | |
| "logits/chosen": -2.578125, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -406.0, | |
| "logps/rejected": -700.0, | |
| "loss": 0.2271, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -1.4453125, | |
| "rewards/margins": 2.9375, | |
| "rewards/rejected": -4.375, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.37656903765690375, | |
| "grad_norm": 29.833221800540095, | |
| "learning_rate": 3.9247834624635404e-07, | |
| "logits/chosen": -2.609375, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -392.0, | |
| "logps/rejected": -744.0, | |
| "loss": 0.2362, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -1.265625, | |
| "rewards/margins": 3.515625, | |
| "rewards/rejected": -4.78125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.39748953974895396, | |
| "grad_norm": 21.11238514940775, | |
| "learning_rate": 3.7710310482256523e-07, | |
| "logits/chosen": -2.65625, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -398.0, | |
| "logps/rejected": -700.0, | |
| "loss": 0.2183, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.1171875, | |
| "rewards/margins": 3.03125, | |
| "rewards/rejected": -4.125, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.41841004184100417, | |
| "grad_norm": 35.4524188550725, | |
| "learning_rate": 3.610497133404795e-07, | |
| "logits/chosen": -2.453125, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -430.0, | |
| "logps/rejected": -772.0, | |
| "loss": 0.2113, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.6015625, | |
| "rewards/margins": 3.296875, | |
| "rewards/rejected": -4.875, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4393305439330544, | |
| "grad_norm": 41.104340492508946, | |
| "learning_rate": 3.4440382358952115e-07, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -460.0, | |
| "logps/rejected": -812.0, | |
| "loss": 0.223, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.5703125, | |
| "rewards/margins": 3.5625, | |
| "rewards/rejected": -5.125, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.4602510460251046, | |
| "grad_norm": 34.26638515144042, | |
| "learning_rate": 3.272542485937368e-07, | |
| "logits/chosen": -2.5625, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -378.0, | |
| "logps/rejected": -716.0, | |
| "loss": 0.1988, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.0859375, | |
| "rewards/margins": 3.265625, | |
| "rewards/rejected": -4.34375, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.4811715481171548, | |
| "grad_norm": 53.87875153292311, | |
| "learning_rate": 3.096924887558854e-07, | |
| "logits/chosen": -2.4375, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -402.0, | |
| "logps/rejected": -752.0, | |
| "loss": 0.2061, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.4765625, | |
| "rewards/margins": 3.5, | |
| "rewards/rejected": -4.96875, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.502092050209205, | |
| "grad_norm": 43.34840736906749, | |
| "learning_rate": 2.9181224366319943e-07, | |
| "logits/chosen": -2.546875, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -436.0, | |
| "logps/rejected": -776.0, | |
| "loss": 0.2256, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": -1.5703125, | |
| "rewards/margins": 3.515625, | |
| "rewards/rejected": -5.09375, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5230125523012552, | |
| "grad_norm": 32.74059701980047, | |
| "learning_rate": 2.7370891215954565e-07, | |
| "logits/chosen": -2.546875, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -456.0, | |
| "logps/rejected": -800.0, | |
| "loss": 0.1809, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.4609375, | |
| "rewards/margins": 3.75, | |
| "rewards/rejected": -5.21875, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5439330543933054, | |
| "grad_norm": 29.82969454828055, | |
| "learning_rate": 2.55479083351317e-07, | |
| "logits/chosen": -2.453125, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -496.0, | |
| "logps/rejected": -932.0, | |
| "loss": 0.1868, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": -1.9296875, | |
| "rewards/margins": 4.5, | |
| "rewards/rejected": -6.4375, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5648535564853556, | |
| "grad_norm": 31.97984265842138, | |
| "learning_rate": 2.3722002126275822e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.09375, | |
| "logps/chosen": -512.0, | |
| "logps/rejected": -908.0, | |
| "loss": 0.1886, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -2.234375, | |
| "rewards/margins": 4.0, | |
| "rewards/rejected": -6.25, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5857740585774058, | |
| "grad_norm": 33.842905395793814, | |
| "learning_rate": 2.19029145890313e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.0625, | |
| "logps/chosen": -468.0, | |
| "logps/rejected": -912.0, | |
| "loss": 0.2042, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -1.921875, | |
| "rewards/margins": 4.375, | |
| "rewards/rejected": -6.3125, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.606694560669456, | |
| "grad_norm": 29.808042971238702, | |
| "learning_rate": 2.0100351342479216e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.078125, | |
| "logps/chosen": -504.0, | |
| "logps/rejected": -892.0, | |
| "loss": 0.2046, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -2.40625, | |
| "rewards/margins": 4.0625, | |
| "rewards/rejected": -6.46875, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6276150627615062, | |
| "grad_norm": 24.901363485174162, | |
| "learning_rate": 1.8323929841460178e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.03125, | |
| "logps/chosen": -544.0, | |
| "logps/rejected": -1024.0, | |
| "loss": 0.1633, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.375, | |
| "rewards/margins": 5.125, | |
| "rewards/rejected": -7.5, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6485355648535565, | |
| "grad_norm": 37.43441496699697, | |
| "learning_rate": 1.6583128063291573e-07, | |
| "logits/chosen": -2.109375, | |
| "logits/rejected": -1.9765625, | |
| "logps/chosen": -628.0, | |
| "logps/rejected": -1096.0, | |
| "loss": 0.1643, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -3.265625, | |
| "rewards/margins": 5.0, | |
| "rewards/rejected": -8.25, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6694560669456067, | |
| "grad_norm": 37.48003403608105, | |
| "learning_rate": 1.488723393865766e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -1.8984375, | |
| "logps/chosen": -664.0, | |
| "logps/rejected": -1200.0, | |
| "loss": 0.1704, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -3.578125, | |
| "rewards/margins": 5.46875, | |
| "rewards/rejected": -9.0625, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6903765690376569, | |
| "grad_norm": 28.243117125958133, | |
| "learning_rate": 1.3245295796480788e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -1.9375, | |
| "logps/chosen": -616.0, | |
| "logps/rejected": -1112.0, | |
| "loss": 0.1781, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -3.421875, | |
| "rewards/margins": 5.03125, | |
| "rewards/rejected": -8.4375, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7112970711297071, | |
| "grad_norm": 32.38858546736874, | |
| "learning_rate": 1.1666074087171627e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.09375, | |
| "logps/chosen": -684.0, | |
| "logps/rejected": -1200.0, | |
| "loss": 0.1773, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -3.765625, | |
| "rewards/margins": 5.1875, | |
| "rewards/rejected": -8.9375, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7322175732217573, | |
| "grad_norm": 20.881908991758937, | |
| "learning_rate": 1.0157994641835734e-07, | |
| "logits/chosen": -2.125, | |
| "logits/rejected": -1.8359375, | |
| "logps/chosen": -672.0, | |
| "logps/rejected": -1176.0, | |
| "loss": 0.1391, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -4.0, | |
| "rewards/margins": 5.0, | |
| "rewards/rejected": -9.0, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7531380753138075, | |
| "grad_norm": 47.96096036564937, | |
| "learning_rate": 8.729103716819111e-08, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -1.859375, | |
| "logps/chosen": -772.0, | |
| "logps/rejected": -1328.0, | |
| "loss": 0.1579, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -4.625, | |
| "rewards/margins": 5.71875, | |
| "rewards/rejected": -10.375, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7740585774058577, | |
| "grad_norm": 26.83924066040214, | |
| "learning_rate": 7.387025063449081e-08, | |
| "logits/chosen": -2.15625, | |
| "logits/rejected": -1.8203125, | |
| "logps/chosen": -692.0, | |
| "logps/rejected": -1216.0, | |
| "loss": 0.1937, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -4.09375, | |
| "rewards/margins": 5.6875, | |
| "rewards/rejected": -9.75, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7949790794979079, | |
| "grad_norm": 20.889028958191947, | |
| "learning_rate": 6.138919252022435e-08, | |
| "logits/chosen": -2.078125, | |
| "logits/rejected": -1.828125, | |
| "logps/chosen": -608.0, | |
| "logps/rejected": -1144.0, | |
| "loss": 0.1732, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -3.625, | |
| "rewards/margins": 5.15625, | |
| "rewards/rejected": -8.75, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8158995815899581, | |
| "grad_norm": 34.69509873753847, | |
| "learning_rate": 4.991445467064689e-08, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -1.9609375, | |
| "logps/chosen": -704.0, | |
| "logps/rejected": -1248.0, | |
| "loss": 0.1511, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.890625, | |
| "rewards/margins": 5.625, | |
| "rewards/rejected": -9.5, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8368200836820083, | |
| "grad_norm": 33.52819855991653, | |
| "learning_rate": 3.9507259776993954e-08, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -1.96875, | |
| "logps/chosen": -684.0, | |
| "logps/rejected": -1248.0, | |
| "loss": 0.1726, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -4.0625, | |
| "rewards/margins": 5.53125, | |
| "rewards/rejected": -9.5625, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8577405857740585, | |
| "grad_norm": 26.436772284559964, | |
| "learning_rate": 3.022313472693447e-08, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.015625, | |
| "logps/chosen": -704.0, | |
| "logps/rejected": -1240.0, | |
| "loss": 0.1649, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -4.0625, | |
| "rewards/margins": 5.65625, | |
| "rewards/rejected": -9.75, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8786610878661087, | |
| "grad_norm": 24.368544694437634, | |
| "learning_rate": 2.2111614344599684e-08, | |
| "logits/chosen": -2.1875, | |
| "logits/rejected": -2.03125, | |
| "logps/chosen": -712.0, | |
| "logps/rejected": -1192.0, | |
| "loss": 0.1711, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -3.90625, | |
| "rewards/margins": 5.15625, | |
| "rewards/rejected": -9.0625, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.899581589958159, | |
| "grad_norm": 18.822439290497464, | |
| "learning_rate": 1.521597710086439e-08, | |
| "logits/chosen": -2.15625, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -672.0, | |
| "logps/rejected": -1208.0, | |
| "loss": 0.1708, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -3.8125, | |
| "rewards/margins": 5.84375, | |
| "rewards/rejected": -9.6875, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9205020920502092, | |
| "grad_norm": 37.62001312817063, | |
| "learning_rate": 9.57301420397924e-09, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -1.953125, | |
| "logps/chosen": -680.0, | |
| "logps/rejected": -1304.0, | |
| "loss": 0.1607, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.8125, | |
| "rewards/margins": 6.25, | |
| "rewards/rejected": -10.0625, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9414225941422594, | |
| "grad_norm": 20.4070987924558, | |
| "learning_rate": 5.212833302556258e-09, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -1.9609375, | |
| "logps/chosen": -656.0, | |
| "logps/rejected": -1240.0, | |
| "loss": 0.153, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -3.5625, | |
| "rewards/margins": 5.84375, | |
| "rewards/rejected": -9.4375, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9623430962343096, | |
| "grad_norm": 20.790605983769925, | |
| "learning_rate": 2.158697848236607e-09, | |
| "logits/chosen": -2.171875, | |
| "logits/rejected": -1.8671875, | |
| "logps/chosen": -668.0, | |
| "logps/rejected": -1232.0, | |
| "loss": 0.1462, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -3.875, | |
| "rewards/margins": 5.75, | |
| "rewards/rejected": -9.625, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.9832635983263598, | |
| "grad_norm": 38.726643797088684, | |
| "learning_rate": 4.269029751107489e-10, | |
| "logits/chosen": -2.171875, | |
| "logits/rejected": -1.9140625, | |
| "logps/chosen": -700.0, | |
| "logps/rejected": -1184.0, | |
| "loss": 0.148, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": -4.0625, | |
| "rewards/margins": 5.0625, | |
| "rewards/rejected": -9.125, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 478, | |
| "total_flos": 0.0, | |
| "train_loss": 0.2593778032638039, | |
| "train_runtime": 6907.5185, | |
| "train_samples_per_second": 8.85, | |
| "train_steps_per_second": 0.069 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 478, | |
| "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 | |
| } | |