| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984301412872841, | |
| "eval_steps": 500, | |
| "global_step": 477, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.020931449502878074, | |
| "grad_norm": 0.8589722514152527, | |
| "learning_rate": 3.3333333333333335e-07, | |
| "logits/chosen": -0.86328125, | |
| "logits/rejected": -0.79296875, | |
| "logps/chosen": -290.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.6924, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": 0.0004367828369140625, | |
| "rewards/margins": 0.0003795623779296875, | |
| "rewards/rejected": 6.008148193359375e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04186289900575615, | |
| "grad_norm": 0.926968514919281, | |
| "learning_rate": 6.666666666666667e-07, | |
| "logits/chosen": -0.80078125, | |
| "logits/rejected": -0.77734375, | |
| "logps/chosen": -282.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.4437499940395355, | |
| "rewards/chosen": 0.00124359130859375, | |
| "rewards/margins": 0.002593994140625, | |
| "rewards/rejected": -0.00135040283203125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06279434850863422, | |
| "grad_norm": 0.7986102104187012, | |
| "learning_rate": 7.996537632406924e-07, | |
| "logits/chosen": -0.80859375, | |
| "logits/rejected": -0.70703125, | |
| "logps/chosen": -280.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.6869, | |
| "rewards/accuracies": 0.578125, | |
| "rewards/chosen": 0.007049560546875, | |
| "rewards/margins": 0.0140380859375, | |
| "rewards/rejected": -0.0069580078125, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.0837257980115123, | |
| "grad_norm": 0.7863888144493103, | |
| "learning_rate": 7.975400422088283e-07, | |
| "logits/chosen": -0.89453125, | |
| "logits/rejected": -0.8203125, | |
| "logps/chosen": -320.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.6783, | |
| "rewards/accuracies": 0.659375011920929, | |
| "rewards/chosen": 0.00872802734375, | |
| "rewards/margins": 0.04638671875, | |
| "rewards/rejected": -0.037841796875, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "grad_norm": 0.8727136254310608, | |
| "learning_rate": 7.9351510351176e-07, | |
| "logits/chosen": -0.83984375, | |
| "logits/rejected": -0.7734375, | |
| "logps/chosen": -314.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.6657, | |
| "rewards/accuracies": 0.628125011920929, | |
| "rewards/chosen": -0.043212890625, | |
| "rewards/margins": 0.078125, | |
| "rewards/rejected": -0.12158203125, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12558869701726844, | |
| "grad_norm": 1.0111013650894165, | |
| "learning_rate": 7.875982974868683e-07, | |
| "logits/chosen": -0.87109375, | |
| "logits/rejected": -0.796875, | |
| "logps/chosen": -302.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.6539, | |
| "rewards/accuracies": 0.640625, | |
| "rewards/chosen": -0.1083984375, | |
| "rewards/margins": 0.08837890625, | |
| "rewards/rejected": -0.197265625, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14652014652014653, | |
| "grad_norm": 0.9877094030380249, | |
| "learning_rate": 7.798180698326886e-07, | |
| "logits/chosen": -0.9375, | |
| "logits/rejected": -0.8046875, | |
| "logps/chosen": -310.0, | |
| "logps/rejected": -294.0, | |
| "loss": 0.6521, | |
| "rewards/accuracies": 0.6156250238418579, | |
| "rewards/chosen": -0.177734375, | |
| "rewards/margins": 0.09375, | |
| "rewards/rejected": -0.271484375, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.1674515960230246, | |
| "grad_norm": 1.1160807609558105, | |
| "learning_rate": 7.702118248530725e-07, | |
| "logits/chosen": -1.015625, | |
| "logits/rejected": -0.875, | |
| "logps/chosen": -302.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.639, | |
| "rewards/accuracies": 0.659375011920929, | |
| "rewards/chosen": -0.1962890625, | |
| "rewards/margins": 0.1416015625, | |
| "rewards/rejected": -0.337890625, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18838304552590268, | |
| "grad_norm": 1.193607211112976, | |
| "learning_rate": 7.588257456318656e-07, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -1.0, | |
| "logps/chosen": -324.0, | |
| "logps/rejected": -316.0, | |
| "loss": 0.6385, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.25390625, | |
| "rewards/margins": 0.1552734375, | |
| "rewards/rejected": -0.41015625, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "grad_norm": 1.208680510520935, | |
| "learning_rate": 7.457145720026305e-07, | |
| "logits/chosen": -1.046875, | |
| "logits/rejected": -0.96484375, | |
| "logps/chosen": -306.0, | |
| "logps/rejected": -306.0, | |
| "loss": 0.6233, | |
| "rewards/accuracies": 0.703125, | |
| "rewards/chosen": -0.279296875, | |
| "rewards/margins": 0.2138671875, | |
| "rewards/rejected": -0.4921875, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2302459445316588, | |
| "grad_norm": 1.580940842628479, | |
| "learning_rate": 7.309413373808504e-07, | |
| "logits/chosen": -1.2421875, | |
| "logits/rejected": -1.171875, | |
| "logps/chosen": -318.0, | |
| "logps/rejected": -330.0, | |
| "loss": 0.6237, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -0.365234375, | |
| "rewards/margins": 0.23046875, | |
| "rewards/rejected": -0.59765625, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.25117739403453687, | |
| "grad_norm": 1.2790683507919312, | |
| "learning_rate": 7.145770657238214e-07, | |
| "logits/chosen": -1.3203125, | |
| "logits/rejected": -1.296875, | |
| "logps/chosen": -350.0, | |
| "logps/rejected": -332.0, | |
| "loss": 0.6122, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.333984375, | |
| "rewards/margins": 0.2734375, | |
| "rewards/rejected": -0.60546875, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.272108843537415, | |
| "grad_norm": 1.2920271158218384, | |
| "learning_rate": 6.967004300751329e-07, | |
| "logits/chosen": -1.375, | |
| "logits/rejected": -1.2265625, | |
| "logps/chosen": -350.0, | |
| "logps/rejected": -366.0, | |
| "loss": 0.6083, | |
| "rewards/accuracies": 0.6968749761581421, | |
| "rewards/chosen": -0.392578125, | |
| "rewards/margins": 0.279296875, | |
| "rewards/rejected": -0.671875, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.29304029304029305, | |
| "grad_norm": 1.4771559238433838, | |
| "learning_rate": 6.773973743353193e-07, | |
| "logits/chosen": -1.3828125, | |
| "logits/rejected": -1.28125, | |
| "logps/chosen": -338.0, | |
| "logps/rejected": -338.0, | |
| "loss": 0.6074, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.578125, | |
| "rewards/margins": 0.28125, | |
| "rewards/rejected": -0.859375, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "grad_norm": 1.7459709644317627, | |
| "learning_rate": 6.567607000770662e-07, | |
| "logits/chosen": -1.390625, | |
| "logits/rejected": -1.2421875, | |
| "logps/chosen": -328.0, | |
| "logps/rejected": -344.0, | |
| "loss": 0.6011, | |
| "rewards/accuracies": 0.6656249761581421, | |
| "rewards/chosen": -0.453125, | |
| "rewards/margins": 0.26953125, | |
| "rewards/rejected": -0.72265625, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.3349031920460492, | |
| "grad_norm": 1.526841163635254, | |
| "learning_rate": 6.348896203913977e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -362.0, | |
| "logps/rejected": -330.0, | |
| "loss": 0.5885, | |
| "rewards/accuracies": 0.703125, | |
| "rewards/chosen": -0.419921875, | |
| "rewards/margins": 0.33203125, | |
| "rewards/rejected": -0.75390625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.35583464154892724, | |
| "grad_norm": 1.7275638580322266, | |
| "learning_rate": 6.118892829097829e-07, | |
| "logits/chosen": -1.5234375, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -342.0, | |
| "logps/rejected": -342.0, | |
| "loss": 0.596, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": -0.46875, | |
| "rewards/margins": 0.287109375, | |
| "rewards/rejected": -0.75390625, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.37676609105180536, | |
| "grad_norm": 1.8728772401809692, | |
| "learning_rate": 5.878702642952824e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -372.0, | |
| "logps/rejected": -352.0, | |
| "loss": 0.5859, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.6015625, | |
| "rewards/margins": 0.263671875, | |
| "rewards/rejected": -0.8671875, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.3976975405546834, | |
| "grad_norm": 1.604046106338501, | |
| "learning_rate": 5.629480386330271e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.390625, | |
| "logps/chosen": -336.0, | |
| "logps/rejected": -348.0, | |
| "loss": 0.599, | |
| "rewards/accuracies": 0.671875, | |
| "rewards/chosen": -0.56640625, | |
| "rewards/margins": 0.326171875, | |
| "rewards/rejected": -0.890625, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "grad_norm": 1.5941054821014404, | |
| "learning_rate": 5.372424222757983e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -338.0, | |
| "logps/rejected": -340.0, | |
| "loss": 0.5832, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.63671875, | |
| "rewards/margins": 0.337890625, | |
| "rewards/rejected": -0.97265625, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.43956043956043955, | |
| "grad_norm": 1.9532604217529297, | |
| "learning_rate": 5.108769978136762e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -352.0, | |
| "logps/rejected": -360.0, | |
| "loss": 0.5674, | |
| "rewards/accuracies": 0.721875011920929, | |
| "rewards/chosen": -0.54296875, | |
| "rewards/margins": 0.41796875, | |
| "rewards/rejected": -0.9609375, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.4604918890633176, | |
| "grad_norm": 1.8788700103759766, | |
| "learning_rate": 4.839785199370884e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -370.0, | |
| "logps/rejected": -364.0, | |
| "loss": 0.5674, | |
| "rewards/accuracies": 0.734375, | |
| "rewards/chosen": -0.6171875, | |
| "rewards/margins": 0.435546875, | |
| "rewards/rejected": -1.0546875, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.48142333856619574, | |
| "grad_norm": 2.920095682144165, | |
| "learning_rate": 4.5667630604963293e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -390.0, | |
| "logps/rejected": -404.0, | |
| "loss": 0.5769, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.796875, | |
| "rewards/margins": 0.43359375, | |
| "rewards/rejected": -1.2265625, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.5023547880690737, | |
| "grad_norm": 2.513183832168579, | |
| "learning_rate": 4.291016145603776e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -378.0, | |
| "logps/rejected": -382.0, | |
| "loss": 0.5923, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.765625, | |
| "rewards/margins": 0.369140625, | |
| "rewards/rejected": -1.1328125, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "grad_norm": 1.8569754362106323, | |
| "learning_rate": 4.0138701384455775e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -336.0, | |
| "logps/rejected": -360.0, | |
| "loss": 0.5964, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": -0.5859375, | |
| "rewards/margins": 0.373046875, | |
| "rewards/rejected": -0.9609375, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.54421768707483, | |
| "grad_norm": 1.8124388456344604, | |
| "learning_rate": 3.736657449064707e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -378.0, | |
| "logps/rejected": -384.0, | |
| "loss": 0.5732, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.546875, | |
| "rewards/margins": 0.4609375, | |
| "rewards/rejected": -1.0078125, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.565149136577708, | |
| "grad_norm": 1.7872382402420044, | |
| "learning_rate": 3.4607108080862845e-07, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -344.0, | |
| "logps/rejected": -370.0, | |
| "loss": 0.5571, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.58984375, | |
| "rewards/margins": 0.4609375, | |
| "rewards/rejected": -1.046875, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5860805860805861, | |
| "grad_norm": 2.041353940963745, | |
| "learning_rate": 3.187356859467888e-07, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -374.0, | |
| "logps/rejected": -388.0, | |
| "loss": 0.5822, | |
| "rewards/accuracies": 0.6343749761581421, | |
| "rewards/chosen": -0.65234375, | |
| "rewards/margins": 0.388671875, | |
| "rewards/rejected": -1.0390625, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.6070120355834642, | |
| "grad_norm": 1.7994887828826904, | |
| "learning_rate": 2.917909782512186e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -354.0, | |
| "logps/rejected": -364.0, | |
| "loss": 0.5545, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.640625, | |
| "rewards/margins": 0.49609375, | |
| "rewards/rejected": -1.1328125, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "grad_norm": 2.489745855331421, | |
| "learning_rate": 2.6536649738047756e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -380.0, | |
| "logps/rejected": -388.0, | |
| "loss": 0.5812, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.69140625, | |
| "rewards/margins": 0.41015625, | |
| "rewards/rejected": -1.1015625, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6488749345892203, | |
| "grad_norm": 1.7875962257385254, | |
| "learning_rate": 2.3958928194520076e-07, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.390625, | |
| "logps/chosen": -360.0, | |
| "logps/rejected": -412.0, | |
| "loss": 0.5586, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.6484375, | |
| "rewards/margins": 0.49609375, | |
| "rewards/rejected": -1.1484375, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6698063840920984, | |
| "grad_norm": 2.7529468536376953, | |
| "learning_rate": 2.145832587559451e-07, | |
| "logits/chosen": -1.4296875, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -350.0, | |
| "logps/rejected": -390.0, | |
| "loss": 0.5537, | |
| "rewards/accuracies": 0.753125011920929, | |
| "rewards/chosen": -0.61328125, | |
| "rewards/margins": 0.515625, | |
| "rewards/rejected": -1.125, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6907378335949764, | |
| "grad_norm": 2.0331408977508545, | |
| "learning_rate": 1.9046864703135954e-07, | |
| "logits/chosen": -1.3671875, | |
| "logits/rejected": -1.3203125, | |
| "logps/chosen": -354.0, | |
| "logps/rejected": -380.0, | |
| "loss": 0.5699, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.640625, | |
| "rewards/margins": 0.486328125, | |
| "rewards/rejected": -1.125, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7116692830978545, | |
| "grad_norm": 2.0676398277282715, | |
| "learning_rate": 1.673613804310103e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.3828125, | |
| "logps/chosen": -342.0, | |
| "logps/rejected": -368.0, | |
| "loss": 0.583, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.55859375, | |
| "rewards/margins": 0.51171875, | |
| "rewards/rejected": -1.0703125, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "grad_norm": 2.087122917175293, | |
| "learning_rate": 1.4537254969150808e-07, | |
| "logits/chosen": -1.4140625, | |
| "logits/rejected": -1.34375, | |
| "logps/chosen": -346.0, | |
| "logps/rejected": -368.0, | |
| "loss": 0.5659, | |
| "rewards/accuracies": 0.6781250238418579, | |
| "rewards/chosen": -0.62109375, | |
| "rewards/margins": 0.39453125, | |
| "rewards/rejected": -1.015625, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7535321821036107, | |
| "grad_norm": 1.7774925231933594, | |
| "learning_rate": 1.2460786854552e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -342.0, | |
| "logps/rejected": -366.0, | |
| "loss": 0.5771, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": -0.70703125, | |
| "rewards/margins": 0.431640625, | |
| "rewards/rejected": -1.140625, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7744636316064888, | |
| "grad_norm": 3.015979766845703, | |
| "learning_rate": 1.0516716549132283e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -384.0, | |
| "logps/rejected": -380.0, | |
| "loss": 0.572, | |
| "rewards/accuracies": 0.690625011920929, | |
| "rewards/chosen": -0.6875, | |
| "rewards/margins": 0.48828125, | |
| "rewards/rejected": -1.1796875, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7953950811093669, | |
| "grad_norm": 2.341261625289917, | |
| "learning_rate": 8.714390385627006e-08, | |
| "logits/chosen": -1.4296875, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -364.0, | |
| "logps/rejected": -390.0, | |
| "loss": 0.5823, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.62109375, | |
| "rewards/margins": 0.455078125, | |
| "rewards/rejected": -1.078125, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8163265306122449, | |
| "grad_norm": 2.0791091918945312, | |
| "learning_rate": 7.062473246152229e-08, | |
| "logits/chosen": -1.4296875, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -356.0, | |
| "logps/rejected": -364.0, | |
| "loss": 0.571, | |
| "rewards/accuracies": 0.6968749761581421, | |
| "rewards/chosen": -0.62890625, | |
| "rewards/margins": 0.443359375, | |
| "rewards/rejected": -1.0703125, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "grad_norm": 2.144294023513794, | |
| "learning_rate": 5.568906904826809e-08, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.3828125, | |
| "logps/chosen": -364.0, | |
| "logps/rejected": -404.0, | |
| "loss": 0.579, | |
| "rewards/accuracies": 0.6968749761581421, | |
| "rewards/chosen": -0.66796875, | |
| "rewards/margins": 0.4375, | |
| "rewards/rejected": -1.109375, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.858189429618001, | |
| "grad_norm": 1.9777884483337402, | |
| "learning_rate": 4.2408718468160386e-08, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -376.0, | |
| "loss": 0.56, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.59375, | |
| "rewards/margins": 0.52734375, | |
| "rewards/rejected": -1.1171875, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8791208791208791, | |
| "grad_norm": 1.7942496538162231, | |
| "learning_rate": 3.084752747356116e-08, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -374.0, | |
| "loss": 0.5576, | |
| "rewards/accuracies": 0.721875011920929, | |
| "rewards/chosen": -0.61328125, | |
| "rewards/margins": 0.470703125, | |
| "rewards/rejected": -1.0859375, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.9000523286237572, | |
| "grad_norm": 2.875974178314209, | |
| "learning_rate": 2.1061077767226255e-08, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -364.0, | |
| "logps/rejected": -396.0, | |
| "loss": 0.5589, | |
| "rewards/accuracies": 0.6968749761581421, | |
| "rewards/chosen": -0.64453125, | |
| "rewards/margins": 0.484375, | |
| "rewards/rejected": -1.1328125, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9209837781266352, | |
| "grad_norm": 1.7986969947814941, | |
| "learning_rate": 1.309641878713048e-08, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -368.0, | |
| "loss": 0.569, | |
| "rewards/accuracies": 0.7093750238418579, | |
| "rewards/chosen": -0.65234375, | |
| "rewards/margins": 0.5078125, | |
| "rewards/rejected": -1.1640625, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "grad_norm": 2.22747540473938, | |
| "learning_rate": 6.991841511097174e-09, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -376.0, | |
| "logps/rejected": -376.0, | |
| "loss": 0.5611, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.59765625, | |
| "rewards/margins": 0.4296875, | |
| "rewards/rejected": -1.03125, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9628466771323915, | |
| "grad_norm": 1.762345314025879, | |
| "learning_rate": 2.776694368692656e-09, | |
| "logits/chosen": -1.5234375, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -348.0, | |
| "logps/rejected": -368.0, | |
| "loss": 0.5689, | |
| "rewards/accuracies": 0.7281249761581421, | |
| "rewards/chosen": -0.6484375, | |
| "rewards/margins": 0.48828125, | |
| "rewards/rejected": -1.140625, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.9837781266352695, | |
| "grad_norm": 1.8886781930923462, | |
| "learning_rate": 4.712421454093007e-10, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -374.0, | |
| "logps/rejected": -400.0, | |
| "loss": 0.5792, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.65234375, | |
| "rewards/margins": 0.5234375, | |
| "rewards/rejected": -1.171875, | |
| "step": 470 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 477, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "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 | |
| } | |