| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.998430141287284, | |
| "eval_steps": 100, | |
| "global_step": 2865, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0010465724751439038, | |
| "grad_norm": 2.5192613284347782, | |
| "learning_rate": 1.7421602787456448e-07, | |
| "logits/chosen": -10.4375, | |
| "logits/rejected": -9.875, | |
| "logps/chosen": -470.0, | |
| "logps/rejected": -576.0, | |
| "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": 2.4665353383267132, | |
| "learning_rate": 1.7421602787456445e-06, | |
| "logits/chosen": -10.5, | |
| "logits/rejected": -9.625, | |
| "logps/chosen": -410.0, | |
| "logps/rejected": -524.0, | |
| "loss": 0.691, | |
| "rewards/accuracies": 0.2083333283662796, | |
| "rewards/chosen": -0.000408172607421875, | |
| "rewards/margins": 0.00201416015625, | |
| "rewards/rejected": -0.002410888671875, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.020931449502878074, | |
| "grad_norm": 2.4986198604012717, | |
| "learning_rate": 3.484320557491289e-06, | |
| "logits/chosen": -10.875, | |
| "logits/rejected": -9.875, | |
| "logps/chosen": -450.0, | |
| "logps/rejected": -540.0, | |
| "loss": 0.6754, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.004638671875, | |
| "rewards/margins": 0.03369140625, | |
| "rewards/rejected": -0.038330078125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03139717425431711, | |
| "grad_norm": 2.3167489602247455, | |
| "learning_rate": 5.226480836236934e-06, | |
| "logits/chosen": -11.5625, | |
| "logits/rejected": -10.375, | |
| "logps/chosen": -462.0, | |
| "logps/rejected": -588.0, | |
| "loss": 0.5855, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.1220703125, | |
| "rewards/margins": 0.2421875, | |
| "rewards/rejected": -0.36328125, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04186289900575615, | |
| "grad_norm": 1.6654146301365214, | |
| "learning_rate": 6.968641114982578e-06, | |
| "logits/chosen": -10.0625, | |
| "logits/rejected": -9.0, | |
| "logps/chosen": -532.0, | |
| "logps/rejected": -760.0, | |
| "loss": 0.3444, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -0.73828125, | |
| "rewards/margins": 1.0390625, | |
| "rewards/rejected": -1.7734375, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.052328623757195186, | |
| "grad_norm": 1.291290445354996, | |
| "learning_rate": 8.710801393728225e-06, | |
| "logits/chosen": -8.875, | |
| "logits/rejected": -8.25, | |
| "logps/chosen": -636.0, | |
| "logps/rejected": -1088.0, | |
| "loss": 0.1331, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.1875, | |
| "rewards/margins": 3.3125, | |
| "rewards/rejected": -5.5, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06279434850863422, | |
| "grad_norm": 0.37955778767137793, | |
| "learning_rate": 1.0452961672473868e-05, | |
| "logits/chosen": -9.1875, | |
| "logits/rejected": -10.375, | |
| "logps/chosen": -644.0, | |
| "logps/rejected": -1688.0, | |
| "loss": 0.0248, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.140625, | |
| "rewards/margins": 9.375, | |
| "rewards/rejected": -11.5, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07326007326007326, | |
| "grad_norm": 0.2506597831495488, | |
| "learning_rate": 1.2195121951219513e-05, | |
| "logits/chosen": -10.375, | |
| "logits/rejected": -11.9375, | |
| "logps/chosen": -704.0, | |
| "logps/rejected": -2912.0, | |
| "loss": 0.0113, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.28125, | |
| "rewards/margins": 21.375, | |
| "rewards/rejected": -23.625, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.0837257980115123, | |
| "grad_norm": 0.7138296973571668, | |
| "learning_rate": 1.3937282229965156e-05, | |
| "logits/chosen": -12.8125, | |
| "logits/rejected": -14.875, | |
| "logps/chosen": -688.0, | |
| "logps/rejected": -3824.0, | |
| "loss": 0.0101, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.578125, | |
| "rewards/margins": 30.0, | |
| "rewards/rejected": -32.5, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.09419152276295134, | |
| "grad_norm": 0.13528124292191762, | |
| "learning_rate": 1.56794425087108e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -13.375, | |
| "logps/chosen": -660.0, | |
| "logps/rejected": -4032.0, | |
| "loss": 0.0061, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.21875, | |
| "rewards/margins": 32.25, | |
| "rewards/rejected": -34.5, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "grad_norm": 0.013904500654373389, | |
| "learning_rate": 1.742160278745645e-05, | |
| "logits/chosen": -10.5, | |
| "logits/rejected": -12.25, | |
| "logps/chosen": -744.0, | |
| "logps/rejected": -4016.0, | |
| "loss": 0.0021, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.0, | |
| "rewards/margins": 32.0, | |
| "rewards/rejected": -35.0, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "eval_logits/chosen": -10.25, | |
| "eval_logits/rejected": -9.6875, | |
| "eval_logps/chosen": -644.0, | |
| "eval_logps/rejected": -628.0, | |
| "eval_loss": 0.9012656211853027, | |
| "eval_rewards/accuracies": 0.48828125, | |
| "eval_rewards/chosen": -3.25, | |
| "eval_rewards/margins": 0.1416015625, | |
| "eval_rewards/rejected": -3.390625, | |
| "eval_runtime": 48.9435, | |
| "eval_samples_per_second": 40.863, | |
| "eval_steps_per_second": 0.654, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1151229722658294, | |
| "grad_norm": 0.3021381235568857, | |
| "learning_rate": 1.9163763066202093e-05, | |
| "logits/chosen": -10.625, | |
| "logits/rejected": -11.375, | |
| "logps/chosen": -840.0, | |
| "logps/rejected": -3584.0, | |
| "loss": 0.006, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -4.03125, | |
| "rewards/margins": 26.75, | |
| "rewards/rejected": -30.75, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12558869701726844, | |
| "grad_norm": 0.05849187362698848, | |
| "learning_rate": 2.0905923344947736e-05, | |
| "logits/chosen": -10.9375, | |
| "logits/rejected": -11.4375, | |
| "logps/chosen": -736.0, | |
| "logps/rejected": -4064.0, | |
| "loss": 0.0061, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.609375, | |
| "rewards/margins": 32.0, | |
| "rewards/rejected": -34.5, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.1360544217687075, | |
| "grad_norm": 0.019008934010510316, | |
| "learning_rate": 2.264808362369338e-05, | |
| "logits/chosen": -11.5625, | |
| "logits/rejected": -12.3125, | |
| "logps/chosen": -736.0, | |
| "logps/rejected": -5472.0, | |
| "loss": 0.0024, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.328125, | |
| "rewards/margins": 45.75, | |
| "rewards/rejected": -49.25, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.14652014652014653, | |
| "grad_norm": 0.631197401279319, | |
| "learning_rate": 2.4390243902439026e-05, | |
| "logits/chosen": -9.9375, | |
| "logits/rejected": -10.625, | |
| "logps/chosen": -680.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0028, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.59375, | |
| "rewards/margins": 37.5, | |
| "rewards/rejected": -40.0, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15698587127158556, | |
| "grad_norm": 0.1904595584932117, | |
| "learning_rate": 2.6132404181184672e-05, | |
| "logits/chosen": -9.8125, | |
| "logits/rejected": -10.75, | |
| "logps/chosen": -632.0, | |
| "logps/rejected": -4128.0, | |
| "loss": 0.002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.296875, | |
| "rewards/margins": 33.75, | |
| "rewards/rejected": -36.0, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.1674515960230246, | |
| "grad_norm": 10.799019911046502, | |
| "learning_rate": 2.7874564459930312e-05, | |
| "logits/chosen": -9.25, | |
| "logits/rejected": -9.75, | |
| "logps/chosen": -700.0, | |
| "logps/rejected": -4192.0, | |
| "loss": 0.0042, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.625, | |
| "rewards/margins": 34.0, | |
| "rewards/rejected": -36.5, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.17791732077446362, | |
| "grad_norm": 8.148466425401029, | |
| "learning_rate": 2.9616724738675962e-05, | |
| "logits/chosen": -9.875, | |
| "logits/rejected": -10.875, | |
| "logps/chosen": -596.0, | |
| "logps/rejected": -4288.0, | |
| "loss": 0.01, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -1.7109375, | |
| "rewards/margins": 36.0, | |
| "rewards/rejected": -37.75, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.18838304552590268, | |
| "grad_norm": 0.03691271074178826, | |
| "learning_rate": 3.13588850174216e-05, | |
| "logits/chosen": -10.3125, | |
| "logits/rejected": -10.375, | |
| "logps/chosen": -632.0, | |
| "logps/rejected": -4640.0, | |
| "loss": 0.0016, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.890625, | |
| "rewards/margins": 38.75, | |
| "rewards/rejected": -40.75, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.1988487702773417, | |
| "grad_norm": 1.0564597417354518, | |
| "learning_rate": 3.310104529616725e-05, | |
| "logits/chosen": -9.875, | |
| "logits/rejected": -10.25, | |
| "logps/chosen": -772.0, | |
| "logps/rejected": -5312.0, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.3125, | |
| "rewards/margins": 44.5, | |
| "rewards/rejected": -47.75, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "grad_norm": 0.34710483575481166, | |
| "learning_rate": 3.48432055749129e-05, | |
| "logits/chosen": -8.75, | |
| "logits/rejected": -9.4375, | |
| "logps/chosen": -572.0, | |
| "logps/rejected": -3888.0, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.296875, | |
| "rewards/margins": 31.75, | |
| "rewards/rejected": -33.25, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "eval_logits/chosen": -8.0, | |
| "eval_logits/rejected": -7.25, | |
| "eval_logps/chosen": -430.0, | |
| "eval_logps/rejected": -408.0, | |
| "eval_loss": 0.781906247138977, | |
| "eval_rewards/accuracies": 0.5078125, | |
| "eval_rewards/chosen": -1.109375, | |
| "eval_rewards/margins": 0.0849609375, | |
| "eval_rewards/rejected": -1.1953125, | |
| "eval_runtime": 46.5571, | |
| "eval_samples_per_second": 42.958, | |
| "eval_steps_per_second": 0.687, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21978021978021978, | |
| "grad_norm": 0.20333662978795222, | |
| "learning_rate": 3.6585365853658535e-05, | |
| "logits/chosen": -7.5625, | |
| "logits/rejected": -8.125, | |
| "logps/chosen": -580.0, | |
| "logps/rejected": -3392.0, | |
| "loss": 0.0045, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.4140625, | |
| "rewards/margins": 27.25, | |
| "rewards/rejected": -28.75, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.2302459445316588, | |
| "grad_norm": 0.0692589299768928, | |
| "learning_rate": 3.8327526132404185e-05, | |
| "logits/chosen": -9.125, | |
| "logits/rejected": -9.875, | |
| "logps/chosen": -636.0, | |
| "logps/rejected": -3472.0, | |
| "loss": 0.0023, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.125, | |
| "rewards/margins": 27.375, | |
| "rewards/rejected": -29.375, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.24071166928309787, | |
| "grad_norm": 0.011953743371116337, | |
| "learning_rate": 4.006968641114983e-05, | |
| "logits/chosen": -10.6875, | |
| "logits/rejected": -11.1875, | |
| "logps/chosen": -656.0, | |
| "logps/rejected": -3008.0, | |
| "loss": 0.0025, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.109375, | |
| "rewards/margins": 22.25, | |
| "rewards/rejected": -24.375, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.25117739403453687, | |
| "grad_norm": 0.058712478513155926, | |
| "learning_rate": 4.181184668989547e-05, | |
| "logits/chosen": -8.25, | |
| "logits/rejected": -8.4375, | |
| "logps/chosen": -688.0, | |
| "logps/rejected": -3024.0, | |
| "loss": 0.0021, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.828125, | |
| "rewards/margins": 22.625, | |
| "rewards/rejected": -25.375, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.2616431187859759, | |
| "grad_norm": 0.09168338443092798, | |
| "learning_rate": 4.3554006968641115e-05, | |
| "logits/chosen": -9.125, | |
| "logits/rejected": -9.4375, | |
| "logps/chosen": -724.0, | |
| "logps/rejected": -3216.0, | |
| "loss": 0.0111, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -3.03125, | |
| "rewards/margins": 23.875, | |
| "rewards/rejected": -26.875, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.272108843537415, | |
| "grad_norm": 0.010925537570642083, | |
| "learning_rate": 4.529616724738676e-05, | |
| "logits/chosen": -9.1875, | |
| "logits/rejected": -9.75, | |
| "logps/chosen": -684.0, | |
| "logps/rejected": -3600.0, | |
| "loss": 0.0092, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.65625, | |
| "rewards/margins": 28.0, | |
| "rewards/rejected": -30.625, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.282574568288854, | |
| "grad_norm": 0.009810575313643462, | |
| "learning_rate": 4.703832752613241e-05, | |
| "logits/chosen": -8.8125, | |
| "logits/rejected": -10.4375, | |
| "logps/chosen": -572.0, | |
| "logps/rejected": -3296.0, | |
| "loss": 0.0021, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.375, | |
| "rewards/margins": 26.125, | |
| "rewards/rejected": -27.5, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.29304029304029305, | |
| "grad_norm": 0.05390608217681133, | |
| "learning_rate": 4.878048780487805e-05, | |
| "logits/chosen": -10.4375, | |
| "logits/rejected": -12.0, | |
| "logps/chosen": -568.0, | |
| "logps/rejected": -3040.0, | |
| "loss": 0.0046, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.390625, | |
| "rewards/margins": 23.375, | |
| "rewards/rejected": -24.875, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.3035060177917321, | |
| "grad_norm": 0.26800883798976244, | |
| "learning_rate": 4.999983293481418e-05, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -11.6875, | |
| "logps/chosen": -616.0, | |
| "logps/rejected": -3392.0, | |
| "loss": 0.0043, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -1.890625, | |
| "rewards/margins": 26.5, | |
| "rewards/rejected": -28.5, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "grad_norm": 0.04974100891728773, | |
| "learning_rate": 4.9996862949181595e-05, | |
| "logits/chosen": -7.21875, | |
| "logits/rejected": -5.6875, | |
| "logps/chosen": -676.0, | |
| "logps/rejected": -3472.0, | |
| "loss": 0.0056, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -2.453125, | |
| "rewards/margins": 27.125, | |
| "rewards/rejected": -29.5, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "eval_logits/chosen": -4.6875, | |
| "eval_logits/rejected": -3.734375, | |
| "eval_logps/chosen": -716.0, | |
| "eval_logps/rejected": -720.0, | |
| "eval_loss": 0.8232969045639038, | |
| "eval_rewards/accuracies": 0.5390625, | |
| "eval_rewards/chosen": -3.984375, | |
| "eval_rewards/margins": 0.33984375, | |
| "eval_rewards/rejected": -4.3125, | |
| "eval_runtime": 46.5518, | |
| "eval_samples_per_second": 42.963, | |
| "eval_steps_per_second": 0.687, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32443746729461015, | |
| "grad_norm": 0.05023971722538096, | |
| "learning_rate": 4.9990180911526896e-05, | |
| "logits/chosen": -5.78125, | |
| "logits/rejected": -4.375, | |
| "logps/chosen": -784.0, | |
| "logps/rejected": -4016.0, | |
| "loss": 0.0017, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.6875, | |
| "rewards/margins": 31.125, | |
| "rewards/rejected": -34.75, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.3349031920460492, | |
| "grad_norm": 0.053152668302615586, | |
| "learning_rate": 4.997978781413745e-05, | |
| "logits/chosen": -7.21875, | |
| "logits/rejected": -5.4375, | |
| "logps/chosen": -900.0, | |
| "logps/rejected": -4224.0, | |
| "loss": 0.0033, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.625, | |
| "rewards/margins": 32.0, | |
| "rewards/rejected": -36.75, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.3453689167974882, | |
| "grad_norm": 0.0032197212600996055, | |
| "learning_rate": 4.996568520039565e-05, | |
| "logits/chosen": -8.3125, | |
| "logits/rejected": -5.53125, | |
| "logps/chosen": -692.0, | |
| "logps/rejected": -2944.0, | |
| "loss": 0.0033, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.65625, | |
| "rewards/margins": 21.625, | |
| "rewards/rejected": -24.25, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.35583464154892724, | |
| "grad_norm": 0.04260850605903489, | |
| "learning_rate": 4.994787516454973e-05, | |
| "logits/chosen": -7.71875, | |
| "logits/rejected": -4.96875, | |
| "logps/chosen": -596.0, | |
| "logps/rejected": -2640.0, | |
| "loss": 0.0073, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.0625, | |
| "rewards/margins": 19.125, | |
| "rewards/rejected": -21.125, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.3663003663003663, | |
| "grad_norm": 0.025053430782429897, | |
| "learning_rate": 4.992636035140273e-05, | |
| "logits/chosen": -9.3125, | |
| "logits/rejected": -6.1875, | |
| "logps/chosen": -712.0, | |
| "logps/rejected": -3040.0, | |
| "loss": 0.0035, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.609375, | |
| "rewards/margins": 22.0, | |
| "rewards/rejected": -24.625, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.37676609105180536, | |
| "grad_norm": 0.000619333389746619, | |
| "learning_rate": 4.99011439559198e-05, | |
| "logits/chosen": -10.1875, | |
| "logits/rejected": -6.1875, | |
| "logps/chosen": -892.0, | |
| "logps/rejected": -4160.0, | |
| "loss": 0.0029, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 32.0, | |
| "rewards/rejected": -36.25, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.3872318158032444, | |
| "grad_norm": 0.8086485639533979, | |
| "learning_rate": 4.987222972275369e-05, | |
| "logits/chosen": -9.8125, | |
| "logits/rejected": -7.625, | |
| "logps/chosen": -700.0, | |
| "logps/rejected": -3488.0, | |
| "loss": 0.007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.109375, | |
| "rewards/margins": 26.625, | |
| "rewards/rejected": -29.75, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.3976975405546834, | |
| "grad_norm": 0.036739349084641895, | |
| "learning_rate": 4.983962194568872e-05, | |
| "logits/chosen": -8.9375, | |
| "logits/rejected": -5.03125, | |
| "logps/chosen": -772.0, | |
| "logps/rejected": -3616.0, | |
| "loss": 0.0075, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.578125, | |
| "rewards/margins": 27.25, | |
| "rewards/rejected": -30.875, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.40816326530612246, | |
| "grad_norm": 0.0034155971218821357, | |
| "learning_rate": 4.9803325467003084e-05, | |
| "logits/chosen": -10.0, | |
| "logits/rejected": -4.625, | |
| "logps/chosen": -792.0, | |
| "logps/rejected": -5088.0, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.6875, | |
| "rewards/margins": 42.0, | |
| "rewards/rejected": -45.75, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "grad_norm": 0.04415452617677344, | |
| "learning_rate": 4.9763345676749816e-05, | |
| "logits/chosen": -9.5625, | |
| "logits/rejected": -4.90625, | |
| "logps/chosen": -852.0, | |
| "logps/rejected": -7104.0, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.734375, | |
| "rewards/margins": 61.25, | |
| "rewards/rejected": -65.0, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "eval_logits/chosen": -8.125, | |
| "eval_logits/rejected": -7.5, | |
| "eval_logps/chosen": -856.0, | |
| "eval_logps/rejected": -868.0, | |
| "eval_loss": 1.2957812547683716, | |
| "eval_rewards/accuracies": 0.515625, | |
| "eval_rewards/chosen": -5.375, | |
| "eval_rewards/margins": 0.4140625, | |
| "eval_rewards/rejected": -5.78125, | |
| "eval_runtime": 46.551, | |
| "eval_samples_per_second": 42.964, | |
| "eval_steps_per_second": 0.687, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4290947148090005, | |
| "grad_norm": 0.0014740017840330338, | |
| "learning_rate": 4.9719688511956374e-05, | |
| "logits/chosen": -9.9375, | |
| "logits/rejected": -5.5625, | |
| "logps/chosen": -816.0, | |
| "logps/rejected": -5600.0, | |
| "loss": 0.0131, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.21875, | |
| "rewards/margins": 46.5, | |
| "rewards/rejected": -50.75, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.43956043956043955, | |
| "grad_norm": 0.08589315438382522, | |
| "learning_rate": 4.967236045574294e-05, | |
| "logits/chosen": -8.1875, | |
| "logits/rejected": -4.375, | |
| "logps/chosen": -1012.0, | |
| "logps/rejected": -4192.0, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.84375, | |
| "rewards/margins": 30.75, | |
| "rewards/rejected": -36.5, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.4500261643118786, | |
| "grad_norm": 0.000984314369031844, | |
| "learning_rate": 4.96213685363597e-05, | |
| "logits/chosen": -9.125, | |
| "logits/rejected": -5.59375, | |
| "logps/chosen": -1056.0, | |
| "logps/rejected": -4160.0, | |
| "loss": 0.0044, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -6.0625, | |
| "rewards/margins": 30.125, | |
| "rewards/rejected": -36.25, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.4604918890633176, | |
| "grad_norm": 0.1319461525289956, | |
| "learning_rate": 4.956672032614315e-05, | |
| "logits/chosen": -8.0625, | |
| "logits/rejected": -5.625, | |
| "logps/chosen": -972.0, | |
| "logps/rejected": -3216.0, | |
| "loss": 0.063, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.4375, | |
| "rewards/margins": 21.5, | |
| "rewards/rejected": -26.875, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.47095761381475665, | |
| "grad_norm": 0.0008895978316205002, | |
| "learning_rate": 4.950842394039157e-05, | |
| "logits/chosen": -9.0625, | |
| "logits/rejected": -6.4375, | |
| "logps/chosen": -1112.0, | |
| "logps/rejected": -3536.0, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.8125, | |
| "rewards/margins": 23.25, | |
| "rewards/rejected": -30.0, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.48142333856619574, | |
| "grad_norm": 5.8429232435058585e-05, | |
| "learning_rate": 4.944648803615994e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -8.125, | |
| "logps/chosen": -932.0, | |
| "logps/rejected": -4224.0, | |
| "loss": 0.0222, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 31.75, | |
| "rewards/rejected": -36.5, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.49188906331763477, | |
| "grad_norm": 0.005101644588821262, | |
| "learning_rate": 4.938092181097433e-05, | |
| "logits/chosen": -8.5, | |
| "logits/rejected": -4.78125, | |
| "logps/chosen": -1272.0, | |
| "logps/rejected": -5664.0, | |
| "loss": 0.0306, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -8.125, | |
| "rewards/margins": 43.0, | |
| "rewards/rejected": -51.0, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5023547880690737, | |
| "grad_norm": 0.0006639191730265733, | |
| "learning_rate": 4.931173500146607e-05, | |
| "logits/chosen": -7.15625, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -1048.0, | |
| "logps/rejected": -4352.0, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.96875, | |
| "rewards/margins": 31.875, | |
| "rewards/rejected": -37.75, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.5128205128205128, | |
| "grad_norm": 2.668412040730523, | |
| "learning_rate": 4.923893788192582e-05, | |
| "logits/chosen": -7.53125, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -1032.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0191, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.09375, | |
| "rewards/margins": 32.25, | |
| "rewards/rejected": -38.5, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "grad_norm": 1.0168979183641662e-06, | |
| "learning_rate": 4.9162541262777906e-05, | |
| "logits/chosen": -8.625, | |
| "logits/rejected": -3.578125, | |
| "logps/chosen": -1016.0, | |
| "logps/rejected": -4992.0, | |
| "loss": 0.0089, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.0, | |
| "rewards/margins": 39.0, | |
| "rewards/rejected": -44.75, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "eval_logits/chosen": -8.375, | |
| "eval_logits/rejected": -7.65625, | |
| "eval_logps/chosen": -1160.0, | |
| "eval_logps/rejected": -1176.0, | |
| "eval_loss": 1.5850000381469727, | |
| "eval_rewards/accuracies": 0.52734375, | |
| "eval_rewards/chosen": -8.4375, | |
| "eval_rewards/margins": 0.46875, | |
| "eval_rewards/rejected": -8.875, | |
| "eval_runtime": 46.5423, | |
| "eval_samples_per_second": 42.972, | |
| "eval_steps_per_second": 0.688, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.533751962323391, | |
| "grad_norm": 0.006778710850875391, | |
| "learning_rate": 4.90825564889749e-05, | |
| "logits/chosen": -10.625, | |
| "logits/rejected": -7.96875, | |
| "logps/chosen": -856.0, | |
| "logps/rejected": -4768.0, | |
| "loss": 0.0336, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 37.75, | |
| "rewards/rejected": -42.0, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.54421768707483, | |
| "grad_norm": 0.08763393728515022, | |
| "learning_rate": 4.899899543831288e-05, | |
| "logits/chosen": -10.3125, | |
| "logits/rejected": -7.34375, | |
| "logps/chosen": -836.0, | |
| "logps/rejected": -3728.0, | |
| "loss": 0.0084, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -4.0625, | |
| "rewards/margins": 28.0, | |
| "rewards/rejected": -32.0, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.554683411826269, | |
| "grad_norm": 0.16313057171585002, | |
| "learning_rate": 4.891187051966765e-05, | |
| "logits/chosen": -7.84375, | |
| "logits/rejected": -4.5625, | |
| "logps/chosen": -1088.0, | |
| "logps/rejected": -4128.0, | |
| "loss": 0.0031, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.53125, | |
| "rewards/margins": 29.375, | |
| "rewards/rejected": -35.75, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.565149136577708, | |
| "grad_norm": 0.018223514851259795, | |
| "learning_rate": 4.882119467115191e-05, | |
| "logits/chosen": -7.46875, | |
| "logits/rejected": -3.9375, | |
| "logps/chosen": -1352.0, | |
| "logps/rejected": -4048.0, | |
| "loss": 0.023, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -9.4375, | |
| "rewards/margins": 25.75, | |
| "rewards/rejected": -35.25, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.5756148613291471, | |
| "grad_norm": 0.037616019348923946, | |
| "learning_rate": 4.872698135819402e-05, | |
| "logits/chosen": -12.5, | |
| "logits/rejected": -10.375, | |
| "logps/chosen": -800.0, | |
| "logps/rejected": -2464.0, | |
| "loss": 0.0043, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.484375, | |
| "rewards/margins": 15.5625, | |
| "rewards/rejected": -19.0, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5860805860805861, | |
| "grad_norm": 0.015118877796574037, | |
| "learning_rate": 4.862924457153831e-05, | |
| "logits/chosen": -11.9375, | |
| "logits/rejected": -8.625, | |
| "logps/chosen": -924.0, | |
| "logps/rejected": -3808.0, | |
| "loss": 0.0215, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 27.75, | |
| "rewards/rejected": -32.25, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.5965463108320251, | |
| "grad_norm": 0.5989417262104058, | |
| "learning_rate": 4.8527998825167484e-05, | |
| "logits/chosen": -10.5, | |
| "logits/rejected": -6.71875, | |
| "logps/chosen": -876.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0038, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.34375, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -39.75, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.6070120355834642, | |
| "grad_norm": 0.04169938757830077, | |
| "learning_rate": 4.842325915414728e-05, | |
| "logits/chosen": -10.8125, | |
| "logits/rejected": -6.625, | |
| "logps/chosen": -792.0, | |
| "logps/rejected": -4704.0, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.53125, | |
| "rewards/margins": 38.0, | |
| "rewards/rejected": -41.5, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.6174777603349032, | |
| "grad_norm": 0.00017558567671798871, | |
| "learning_rate": 4.831504111239373e-05, | |
| "logits/chosen": -10.5, | |
| "logits/rejected": -5.9375, | |
| "logps/chosen": -736.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.140625, | |
| "rewards/margins": 37.5, | |
| "rewards/rejected": -40.5, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "grad_norm": 1.978509901663985, | |
| "learning_rate": 4.8203360770363425e-05, | |
| "logits/chosen": -12.1875, | |
| "logits/rejected": -8.5625, | |
| "logps/chosen": -804.0, | |
| "logps/rejected": -4768.0, | |
| "loss": 0.0037, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -3.59375, | |
| "rewards/margins": 38.75, | |
| "rewards/rejected": -42.25, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "eval_logits/chosen": -12.3125, | |
| "eval_logits/rejected": -12.0, | |
| "eval_logps/chosen": -720.0, | |
| "eval_logps/rejected": -720.0, | |
| "eval_loss": 0.9524999856948853, | |
| "eval_rewards/accuracies": 0.521484375, | |
| "eval_rewards/chosen": -4.03125, | |
| "eval_rewards/margins": 0.302734375, | |
| "eval_rewards/rejected": -4.34375, | |
| "eval_runtime": 46.5433, | |
| "eval_samples_per_second": 42.971, | |
| "eval_steps_per_second": 0.688, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6384092098377813, | |
| "grad_norm": 0.043121005256497584, | |
| "learning_rate": 4.808823471266702e-05, | |
| "logits/chosen": -12.0, | |
| "logits/rejected": -10.8125, | |
| "logps/chosen": -716.0, | |
| "logps/rejected": -3424.0, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.6875, | |
| "rewards/margins": 26.125, | |
| "rewards/rejected": -28.75, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.6488749345892203, | |
| "grad_norm": 0.2502728284395065, | |
| "learning_rate": 4.796968003560639e-05, | |
| "logits/chosen": -11.625, | |
| "logits/rejected": -9.1875, | |
| "logps/chosen": -760.0, | |
| "logps/rejected": -2848.0, | |
| "loss": 0.0087, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.859375, | |
| "rewards/margins": 20.25, | |
| "rewards/rejected": -23.125, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.6593406593406593, | |
| "grad_norm": 0.001007321119730767, | |
| "learning_rate": 4.7847714344635864e-05, | |
| "logits/chosen": -10.875, | |
| "logits/rejected": -7.25, | |
| "logps/chosen": -816.0, | |
| "logps/rejected": -2736.0, | |
| "loss": 0.002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.0625, | |
| "rewards/margins": 18.0, | |
| "rewards/rejected": -22.0, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.6698063840920984, | |
| "grad_norm": 0.20831872329492016, | |
| "learning_rate": 4.772235575174776e-05, | |
| "logits/chosen": -8.9375, | |
| "logits/rejected": -4.78125, | |
| "logps/chosen": -816.0, | |
| "logps/rejected": -2256.0, | |
| "loss": 0.0014, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.03125, | |
| "rewards/margins": 13.0625, | |
| "rewards/rejected": -17.125, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.6802721088435374, | |
| "grad_norm": 0.019717028707114094, | |
| "learning_rate": 4.7593622872782754e-05, | |
| "logits/chosen": -8.9375, | |
| "logits/rejected": -4.4375, | |
| "logps/chosen": -860.0, | |
| "logps/rejected": -3232.0, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.25, | |
| "rewards/margins": 22.625, | |
| "rewards/rejected": -26.875, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6907378335949764, | |
| "grad_norm": 8.748764957843357e-06, | |
| "learning_rate": 4.746153482466542e-05, | |
| "logits/chosen": -8.625, | |
| "logits/rejected": -3.046875, | |
| "logps/chosen": -1024.0, | |
| "logps/rejected": -4256.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.0625, | |
| "rewards/margins": 31.375, | |
| "rewards/rejected": -37.5, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.7012035583464155, | |
| "grad_norm": 0.00981450013888215, | |
| "learning_rate": 4.732611122256537e-05, | |
| "logits/chosen": -8.875, | |
| "logits/rejected": -3.609375, | |
| "logps/chosen": -956.0, | |
| "logps/rejected": -4384.0, | |
| "loss": 0.0034, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.5625, | |
| "rewards/margins": 33.0, | |
| "rewards/rejected": -38.75, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.7116692830978545, | |
| "grad_norm": 1.6508606378659035e-05, | |
| "learning_rate": 4.718737217698435e-05, | |
| "logits/chosen": -9.25, | |
| "logits/rejected": -4.75, | |
| "logps/chosen": -1012.0, | |
| "logps/rejected": -4288.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.59375, | |
| "rewards/margins": 31.375, | |
| "rewards/rejected": -37.0, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.7221350078492935, | |
| "grad_norm": 0.08024287621840012, | |
| "learning_rate": 4.704533829076982e-05, | |
| "logits/chosen": -9.5625, | |
| "logits/rejected": -4.6875, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -4016.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 30.25, | |
| "rewards/rejected": -34.75, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "grad_norm": 0.012359221847515261, | |
| "learning_rate": 4.6900030656055455e-05, | |
| "logits/chosen": -7.84375, | |
| "logits/rejected": -2.015625, | |
| "logps/chosen": -1080.0, | |
| "logps/rejected": -4192.0, | |
| "loss": 0.003, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -6.4375, | |
| "rewards/margins": 30.25, | |
| "rewards/rejected": -36.75, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "eval_logits/chosen": -4.875, | |
| "eval_logits/rejected": -3.6875, | |
| "eval_logps/chosen": -1256.0, | |
| "eval_logps/rejected": -1344.0, | |
| "eval_loss": 2.1297969818115234, | |
| "eval_rewards/accuracies": 0.537109375, | |
| "eval_rewards/chosen": -9.4375, | |
| "eval_rewards/margins": 1.1015625, | |
| "eval_rewards/rejected": -10.5, | |
| "eval_runtime": 46.5282, | |
| "eval_samples_per_second": 42.985, | |
| "eval_steps_per_second": 0.688, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7430664573521716, | |
| "grad_norm": 0.002510002699957797, | |
| "learning_rate": 4.6751470851128906e-05, | |
| "logits/chosen": -8.5, | |
| "logits/rejected": -3.40625, | |
| "logps/chosen": -1032.0, | |
| "logps/rejected": -4384.0, | |
| "loss": 0.055, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -5.59375, | |
| "rewards/margins": 32.25, | |
| "rewards/rejected": -38.0, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.7535321821036107, | |
| "grad_norm": 1.9431838844114737, | |
| "learning_rate": 4.6599680937227454e-05, | |
| "logits/chosen": -9.0, | |
| "logits/rejected": -1.9375, | |
| "logps/chosen": -964.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0037, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -5.1875, | |
| "rewards/margins": 33.5, | |
| "rewards/rejected": -38.75, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.7639979068550498, | |
| "grad_norm": 0.00650278196567405, | |
| "learning_rate": 4.644468345526183e-05, | |
| "logits/chosen": -9.6875, | |
| "logits/rejected": -2.765625, | |
| "logps/chosen": -1032.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.28125, | |
| "rewards/margins": 34.25, | |
| "rewards/rejected": -39.5, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.7744636316064888, | |
| "grad_norm": 0.0015257500652966817, | |
| "learning_rate": 4.628650142246894e-05, | |
| "logits/chosen": -9.9375, | |
| "logits/rejected": -2.515625, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.875, | |
| "rewards/margins": 34.25, | |
| "rewards/rejected": -39.0, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.7849293563579278, | |
| "grad_norm": 0.12347475400230522, | |
| "learning_rate": 4.6125158328993765e-05, | |
| "logits/chosen": -10.5, | |
| "logits/rejected": -3.140625, | |
| "logps/chosen": -956.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.0625, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -40.25, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7953950811093669, | |
| "grad_norm": 0.0017630442295318556, | |
| "learning_rate": 4.5960678134401064e-05, | |
| "logits/chosen": -9.8125, | |
| "logits/rejected": -2.671875, | |
| "logps/chosen": -952.0, | |
| "logps/rejected": -4768.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.25, | |
| "rewards/margins": 36.5, | |
| "rewards/rejected": -41.75, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.8058608058608059, | |
| "grad_norm": 0.02084007597688944, | |
| "learning_rate": 4.579308526411734e-05, | |
| "logits/chosen": -10.0625, | |
| "logits/rejected": -2.765625, | |
| "logps/chosen": -960.0, | |
| "logps/rejected": -4768.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.09375, | |
| "rewards/margins": 37.25, | |
| "rewards/rejected": -42.25, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.8163265306122449, | |
| "grad_norm": 0.0002352505397217453, | |
| "learning_rate": 4.562240460580367e-05, | |
| "logits/chosen": -11.0625, | |
| "logits/rejected": -4.65625, | |
| "logps/chosen": -928.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -40.0, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.826792255363684, | |
| "grad_norm": 0.002995146243877048, | |
| "learning_rate": 4.544866150565985e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -5.03125, | |
| "logps/chosen": -936.0, | |
| "logps/rejected": -4864.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.0625, | |
| "rewards/margins": 38.0, | |
| "rewards/rejected": -43.25, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "grad_norm": 7.203434690522864e-07, | |
| "learning_rate": 4.527188176466051e-05, | |
| "logits/chosen": -11.8125, | |
| "logits/rejected": -4.46875, | |
| "logps/chosen": -1240.0, | |
| "logps/rejected": -5536.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -7.96875, | |
| "rewards/margins": 41.5, | |
| "rewards/rejected": -49.5, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "eval_logits/chosen": -9.5, | |
| "eval_logits/rejected": -8.6875, | |
| "eval_logps/chosen": -1264.0, | |
| "eval_logps/rejected": -1336.0, | |
| "eval_loss": 2.112093687057495, | |
| "eval_rewards/accuracies": 0.53125, | |
| "eval_rewards/chosen": -9.4375, | |
| "eval_rewards/margins": 1.0546875, | |
| "eval_rewards/rejected": -10.4375, | |
| "eval_runtime": 46.5341, | |
| "eval_samples_per_second": 42.979, | |
| "eval_steps_per_second": 0.688, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.847723704866562, | |
| "grad_norm": 0.010681170382533217, | |
| "learning_rate": 4.509209163472357e-05, | |
| "logits/chosen": -11.5625, | |
| "logits/rejected": -4.25, | |
| "logps/chosen": -1296.0, | |
| "logps/rejected": -5984.0, | |
| "loss": 0.0427, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -8.5, | |
| "rewards/margins": 45.5, | |
| "rewards/rejected": -54.0, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.858189429618001, | |
| "grad_norm": 0.0013400406572175883, | |
| "learning_rate": 4.4909317814811896e-05, | |
| "logits/chosen": -10.5625, | |
| "logits/rejected": -2.8125, | |
| "logps/chosen": -1168.0, | |
| "logps/rejected": -5024.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -7.53125, | |
| "rewards/margins": 37.25, | |
| "rewards/rejected": -44.75, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.8686551543694401, | |
| "grad_norm": 0.014026889755695321, | |
| "learning_rate": 4.472358744696845e-05, | |
| "logits/chosen": -9.875, | |
| "logits/rejected": -2.515625, | |
| "logps/chosen": -844.0, | |
| "logps/rejected": -3680.0, | |
| "loss": 0.0046, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -4.21875, | |
| "rewards/margins": 27.5, | |
| "rewards/rejected": -31.75, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.8791208791208791, | |
| "grad_norm": 0.0015751584750640687, | |
| "learning_rate": 4.4534928112285645e-05, | |
| "logits/chosen": -7.84375, | |
| "logits/rejected": 0.12109375, | |
| "logps/chosen": -896.0, | |
| "logps/rejected": -3264.0, | |
| "loss": 0.002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.3125, | |
| "rewards/margins": 22.625, | |
| "rewards/rejected": -27.0, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.8895866038723181, | |
| "grad_norm": 0.03288048960543516, | |
| "learning_rate": 4.43433678268096e-05, | |
| "logits/chosen": -8.875, | |
| "logits/rejected": -2.03125, | |
| "logps/chosen": -856.0, | |
| "logps/rejected": -3200.0, | |
| "loss": 0.0123, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.09375, | |
| "rewards/margins": 22.25, | |
| "rewards/rejected": -26.375, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9000523286237572, | |
| "grad_norm": 0.06948267541196752, | |
| "learning_rate": 4.414893503737968e-05, | |
| "logits/chosen": -10.75, | |
| "logits/rejected": -8.3125, | |
| "logps/chosen": -676.0, | |
| "logps/rejected": -2336.0, | |
| "loss": 0.0426, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.5625, | |
| "rewards/margins": 15.5, | |
| "rewards/rejected": -18.125, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.9105180533751962, | |
| "grad_norm": 0.011693288696649684, | |
| "learning_rate": 4.395165861740414e-05, | |
| "logits/chosen": -10.75, | |
| "logits/rejected": -8.1875, | |
| "logps/chosen": -860.0, | |
| "logps/rejected": -3088.0, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.984375, | |
| "rewards/margins": 21.5, | |
| "rewards/rejected": -25.5, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.9209837781266352, | |
| "grad_norm": 0.027460697890334507, | |
| "learning_rate": 4.375156786257241e-05, | |
| "logits/chosen": -10.8125, | |
| "logits/rejected": -6.46875, | |
| "logps/chosen": -1072.0, | |
| "logps/rejected": -4320.0, | |
| "loss": 0.0067, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.40625, | |
| "rewards/margins": 31.25, | |
| "rewards/rejected": -37.75, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.9314495028780743, | |
| "grad_norm": 0.005759252454555949, | |
| "learning_rate": 4.354869248650466e-05, | |
| "logits/chosen": -10.5625, | |
| "logits/rejected": -7.90625, | |
| "logps/chosen": -960.0, | |
| "logps/rejected": -4288.0, | |
| "loss": 0.0032, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.5, | |
| "rewards/margins": 32.25, | |
| "rewards/rejected": -37.75, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "grad_norm": 0.09081079379754264, | |
| "learning_rate": 4.334306261633931e-05, | |
| "logits/chosen": -10.0, | |
| "logits/rejected": -9.25, | |
| "logps/chosen": -792.0, | |
| "logps/rejected": -3936.0, | |
| "loss": 0.0037, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -3.484375, | |
| "rewards/margins": 30.875, | |
| "rewards/rejected": -34.5, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "eval_logits/chosen": -9.9375, | |
| "eval_logits/rejected": -9.625, | |
| "eval_logps/chosen": -1016.0, | |
| "eval_logps/rejected": -996.0, | |
| "eval_loss": 1.3021249771118164, | |
| "eval_rewards/accuracies": 0.515625, | |
| "eval_rewards/chosen": -7.0, | |
| "eval_rewards/margins": 0.09228515625, | |
| "eval_rewards/rejected": -7.09375, | |
| "eval_runtime": 46.5254, | |
| "eval_samples_per_second": 42.987, | |
| "eval_steps_per_second": 0.688, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9523809523809523, | |
| "grad_norm": 0.02091272354950702, | |
| "learning_rate": 4.313470878825911e-05, | |
| "logits/chosen": -10.0, | |
| "logits/rejected": -9.0, | |
| "logps/chosen": -1064.0, | |
| "logps/rejected": -5184.0, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.1875, | |
| "rewards/margins": 40.25, | |
| "rewards/rejected": -46.5, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.9628466771323915, | |
| "grad_norm": 0.01707817572030035, | |
| "learning_rate": 4.292366194295652e-05, | |
| "logits/chosen": -10.0625, | |
| "logits/rejected": -8.6875, | |
| "logps/chosen": -1000.0, | |
| "logps/rejected": -5440.0, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.0, | |
| "rewards/margins": 43.0, | |
| "rewards/rejected": -49.0, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.9733124018838305, | |
| "grad_norm": 0.013174813872298772, | |
| "learning_rate": 4.2709953421038964e-05, | |
| "logits/chosen": -9.875, | |
| "logits/rejected": -8.1875, | |
| "logps/chosen": -1032.0, | |
| "logps/rejected": -5760.0, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.875, | |
| "rewards/margins": 46.5, | |
| "rewards/rejected": -52.5, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.9837781266352695, | |
| "grad_norm": 0.0005257155002367302, | |
| "learning_rate": 4.249361495837478e-05, | |
| "logits/chosen": -10.0625, | |
| "logits/rejected": -8.0, | |
| "logps/chosen": -1032.0, | |
| "logps/rejected": -5920.0, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.25, | |
| "rewards/margins": 48.25, | |
| "rewards/rejected": -54.5, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.9942438513867086, | |
| "grad_norm": 0.000727303659916889, | |
| "learning_rate": 4.227467868138035e-05, | |
| "logits/chosen": -9.625, | |
| "logits/rejected": -7.59375, | |
| "logps/chosen": -1048.0, | |
| "logps/rejected": -6240.0, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.3125, | |
| "rewards/margins": 51.25, | |
| "rewards/rejected": -57.5, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.0047095761381475, | |
| "grad_norm": 0.003795784122476434, | |
| "learning_rate": 4.205317710224938e-05, | |
| "logits/chosen": -9.3125, | |
| "logits/rejected": -7.625, | |
| "logps/chosen": -1048.0, | |
| "logps/rejected": -5696.0, | |
| "loss": 0.0149, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -6.28125, | |
| "rewards/margins": 45.25, | |
| "rewards/rejected": -51.5, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 1.0151753008895865, | |
| "grad_norm": 0.02951928170024009, | |
| "learning_rate": 4.1829143114124733e-05, | |
| "logits/chosen": -9.6875, | |
| "logits/rejected": -8.5, | |
| "logps/chosen": -804.0, | |
| "logps/rejected": -2288.0, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.09375, | |
| "rewards/margins": 13.4375, | |
| "rewards/rejected": -17.5, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 1.0256410256410255, | |
| "grad_norm": 0.0008167443538212538, | |
| "learning_rate": 4.160260998621387e-05, | |
| "logits/chosen": -10.8125, | |
| "logits/rejected": -9.8125, | |
| "logps/chosen": -840.0, | |
| "logps/rejected": -2320.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.96875, | |
| "rewards/margins": 13.5625, | |
| "rewards/rejected": -17.5, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 1.0361067503924646, | |
| "grad_norm": 0.000990869884197823, | |
| "learning_rate": 4.137361135884824e-05, | |
| "logits/chosen": -10.8125, | |
| "logits/rejected": -9.8125, | |
| "logps/chosen": -824.0, | |
| "logps/rejected": -2720.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.09375, | |
| "rewards/margins": 18.0, | |
| "rewards/rejected": -22.125, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 1.0465724751439036, | |
| "grad_norm": 0.0031786008610621705, | |
| "learning_rate": 4.114218123848777e-05, | |
| "logits/chosen": -10.9375, | |
| "logits/rejected": -10.0625, | |
| "logps/chosen": -864.0, | |
| "logps/rejected": -2880.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.25, | |
| "rewards/margins": 19.5, | |
| "rewards/rejected": -23.75, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0465724751439036, | |
| "eval_logits/chosen": -10.75, | |
| "eval_logits/rejected": -10.375, | |
| "eval_logps/chosen": -888.0, | |
| "eval_logps/rejected": -880.0, | |
| "eval_loss": 1.0153437852859497, | |
| "eval_rewards/accuracies": 0.54296875, | |
| "eval_rewards/chosen": -5.71875, | |
| "eval_rewards/margins": 0.208984375, | |
| "eval_rewards/rejected": -5.90625, | |
| "eval_runtime": 46.5378, | |
| "eval_samples_per_second": 42.976, | |
| "eval_steps_per_second": 0.688, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0570381998953426, | |
| "grad_norm": 0.023248161929238204, | |
| "learning_rate": 4.0908353992670824e-05, | |
| "logits/chosen": -10.9375, | |
| "logits/rejected": -10.0, | |
| "logps/chosen": -916.0, | |
| "logps/rejected": -3296.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.0, | |
| "rewards/margins": 22.375, | |
| "rewards/rejected": -27.375, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 1.0675039246467817, | |
| "grad_norm": 0.00047012374625297746, | |
| "learning_rate": 4.067216434491059e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -10.3125, | |
| "logps/chosen": -968.0, | |
| "logps/rejected": -3536.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.28125, | |
| "rewards/margins": 24.5, | |
| "rewards/rejected": -29.875, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 1.077969649398221, | |
| "grad_norm": 0.0011367798820355963, | |
| "learning_rate": 4.043364736953866e-05, | |
| "logits/chosen": -11.0, | |
| "logits/rejected": -10.1875, | |
| "logps/chosen": -920.0, | |
| "logps/rejected": -3472.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.21875, | |
| "rewards/margins": 24.125, | |
| "rewards/rejected": -29.375, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 1.08843537414966, | |
| "grad_norm": 0.00876291909528573, | |
| "learning_rate": 4.019283848649643e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -10.25, | |
| "logps/chosen": -952.0, | |
| "logps/rejected": -3568.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.40625, | |
| "rewards/margins": 24.875, | |
| "rewards/rejected": -30.375, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 1.098901098901099, | |
| "grad_norm": 0.0354271849117392, | |
| "learning_rate": 3.9949773456075215e-05, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -10.0625, | |
| "logps/chosen": -924.0, | |
| "logps/rejected": -3520.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.21875, | |
| "rewards/margins": 24.75, | |
| "rewards/rejected": -30.0, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.109366823652538, | |
| "grad_norm": 0.006569880026363552, | |
| "learning_rate": 3.970448837360585e-05, | |
| "logits/chosen": -11.0625, | |
| "logits/rejected": -9.875, | |
| "logps/chosen": -992.0, | |
| "logps/rejected": -3616.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.71875, | |
| "rewards/margins": 25.0, | |
| "rewards/rejected": -30.75, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 1.119832548403977, | |
| "grad_norm": 0.001495047842097171, | |
| "learning_rate": 3.945701966409846e-05, | |
| "logits/chosen": -11.75, | |
| "logits/rejected": -10.3125, | |
| "logps/chosen": -1012.0, | |
| "logps/rejected": -3664.0, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.5625, | |
| "rewards/margins": 25.375, | |
| "rewards/rejected": -30.875, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 1.130298273155416, | |
| "grad_norm": 0.024631775182688477, | |
| "learning_rate": 3.9207404076833374e-05, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -9.5625, | |
| "logps/chosen": -936.0, | |
| "logps/rejected": -3488.0, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.9375, | |
| "rewards/margins": 24.25, | |
| "rewards/rejected": -29.125, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 1.1407639979068551, | |
| "grad_norm": 0.001820480402381163, | |
| "learning_rate": 3.895567867990379e-05, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -8.4375, | |
| "logps/chosen": -936.0, | |
| "logps/rejected": -3504.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.0625, | |
| "rewards/margins": 24.75, | |
| "rewards/rejected": -29.75, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 1.1512297226582942, | |
| "grad_norm": 2.563474019139197e-05, | |
| "learning_rate": 3.870188085471114e-05, | |
| "logits/chosen": -11.75, | |
| "logits/rejected": -8.375, | |
| "logps/chosen": -956.0, | |
| "logps/rejected": -3552.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.28125, | |
| "rewards/margins": 25.0, | |
| "rewards/rejected": -30.25, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.1512297226582942, | |
| "eval_logits/chosen": -10.5, | |
| "eval_logits/rejected": -9.9375, | |
| "eval_logps/chosen": -972.0, | |
| "eval_logps/rejected": -968.0, | |
| "eval_loss": 1.1537187099456787, | |
| "eval_rewards/accuracies": 0.52734375, | |
| "eval_rewards/chosen": -6.53125, | |
| "eval_rewards/margins": 0.2734375, | |
| "eval_rewards/rejected": -6.8125, | |
| "eval_runtime": 46.5336, | |
| "eval_samples_per_second": 42.98, | |
| "eval_steps_per_second": 0.688, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.1616954474097332, | |
| "grad_norm": 0.00010616798633833599, | |
| "learning_rate": 3.844604829041395e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -9.0625, | |
| "logps/chosen": -892.0, | |
| "logps/rejected": -3504.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.875, | |
| "rewards/margins": 25.125, | |
| "rewards/rejected": -30.0, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 1.1721611721611722, | |
| "grad_norm": 0.0007772764355653845, | |
| "learning_rate": 3.818821897833095e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -10.125, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -3520.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 25.5, | |
| "rewards/rejected": -30.125, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 1.1826268969126112, | |
| "grad_norm": 0.00013227983778573078, | |
| "learning_rate": 3.792843120629935e-05, | |
| "logits/chosen": -11.9375, | |
| "logits/rejected": -10.9375, | |
| "logps/chosen": -884.0, | |
| "logps/rejected": -3696.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 27.0, | |
| "rewards/rejected": -31.75, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 1.1930926216640503, | |
| "grad_norm": 1.783557204401042e-05, | |
| "learning_rate": 3.766672355298907e-05, | |
| "logits/chosen": -11.875, | |
| "logits/rejected": -10.625, | |
| "logps/chosen": -864.0, | |
| "logps/rejected": -3616.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 26.625, | |
| "rewards/rejected": -31.25, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 1.2035583464154893, | |
| "grad_norm": 0.0005379724442896183, | |
| "learning_rate": 3.740313488217373e-05, | |
| "logits/chosen": -11.625, | |
| "logits/rejected": -10.6875, | |
| "logps/chosen": -916.0, | |
| "logps/rejected": -3648.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.59375, | |
| "rewards/margins": 26.25, | |
| "rewards/rejected": -30.75, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.2140240711669283, | |
| "grad_norm": 0.0014939394314237252, | |
| "learning_rate": 3.7137704336959464e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -10.375, | |
| "logps/chosen": -840.0, | |
| "logps/rejected": -3488.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.375, | |
| "rewards/margins": 25.5, | |
| "rewards/rejected": -29.875, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 1.2244897959183674, | |
| "grad_norm": 0.037958633437854467, | |
| "learning_rate": 3.687047133397201e-05, | |
| "logits/chosen": -11.4375, | |
| "logits/rejected": -10.125, | |
| "logps/chosen": -884.0, | |
| "logps/rejected": -3680.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 26.75, | |
| "rewards/rejected": -31.5, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 1.2349555206698064, | |
| "grad_norm": 0.004505565333615856, | |
| "learning_rate": 3.6601475557503415e-05, | |
| "logits/chosen": -11.875, | |
| "logits/rejected": -10.5625, | |
| "logps/chosen": -900.0, | |
| "logps/rejected": -3776.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 27.625, | |
| "rewards/rejected": -32.5, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 1.2454212454212454, | |
| "grad_norm": 0.0005466604045641071, | |
| "learning_rate": 3.633075695361882e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -9.9375, | |
| "logps/chosen": -996.0, | |
| "logps/rejected": -3808.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.1875, | |
| "rewards/margins": 27.125, | |
| "rewards/rejected": -32.5, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 1.2558869701726845, | |
| "grad_norm": 6.136922271865644e-06, | |
| "learning_rate": 3.6058355724224477e-05, | |
| "logits/chosen": -12.125, | |
| "logits/rejected": -10.1875, | |
| "logps/chosen": -944.0, | |
| "logps/rejected": -3984.0, | |
| "loss": 0.0016, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.1875, | |
| "rewards/margins": 29.375, | |
| "rewards/rejected": -34.5, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.2558869701726845, | |
| "eval_logits/chosen": -11.5, | |
| "eval_logits/rejected": -11.125, | |
| "eval_logps/chosen": -1016.0, | |
| "eval_logps/rejected": -1012.0, | |
| "eval_loss": 1.2422187328338623, | |
| "eval_rewards/accuracies": 0.53125, | |
| "eval_rewards/chosen": -6.96875, | |
| "eval_rewards/margins": 0.27734375, | |
| "eval_rewards/rejected": -7.25, | |
| "eval_runtime": 46.5679, | |
| "eval_samples_per_second": 42.948, | |
| "eval_steps_per_second": 0.687, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.2663526949241235, | |
| "grad_norm": 0.0007641794569993429, | |
| "learning_rate": 3.5784312321097744e-05, | |
| "logits/chosen": -11.875, | |
| "logits/rejected": -10.125, | |
| "logps/chosen": -988.0, | |
| "logps/rejected": -4032.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.5, | |
| "rewards/margins": 29.25, | |
| "rewards/rejected": -34.75, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 1.2768184196755625, | |
| "grad_norm": 3.317651627061065e-06, | |
| "learning_rate": 3.550866743987994e-05, | |
| "logits/chosen": -12.25, | |
| "logits/rejected": -10.4375, | |
| "logps/chosen": -1016.0, | |
| "logps/rejected": -4048.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.71875, | |
| "rewards/margins": 29.375, | |
| "rewards/rejected": -35.0, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 1.2872841444270016, | |
| "grad_norm": 1.2500630556567215e-05, | |
| "learning_rate": 3.5231462014033024e-05, | |
| "logits/chosen": -12.1875, | |
| "logits/rejected": -10.0625, | |
| "logps/chosen": -1020.0, | |
| "logps/rejected": -4032.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.65625, | |
| "rewards/margins": 29.125, | |
| "rewards/rejected": -34.75, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 1.2977498691784406, | |
| "grad_norm": 0.0004590841919038584, | |
| "learning_rate": 3.495273720876095e-05, | |
| "logits/chosen": -11.9375, | |
| "logits/rejected": -9.8125, | |
| "logps/chosen": -976.0, | |
| "logps/rejected": -3984.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.71875, | |
| "rewards/margins": 28.875, | |
| "rewards/rejected": -34.5, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 1.3082155939298796, | |
| "grad_norm": 0.003187762256538688, | |
| "learning_rate": 3.4672534414896616e-05, | |
| "logits/chosen": -12.0625, | |
| "logits/rejected": -9.8125, | |
| "logps/chosen": -1064.0, | |
| "logps/rejected": -4192.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.90625, | |
| "rewards/margins": 30.25, | |
| "rewards/rejected": -36.25, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.3186813186813187, | |
| "grad_norm": 0.0008497582138128509, | |
| "learning_rate": 3.439089524275526e-05, | |
| "logits/chosen": -12.25, | |
| "logits/rejected": -10.4375, | |
| "logps/chosen": -1056.0, | |
| "logps/rejected": -4080.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.84375, | |
| "rewards/margins": 29.25, | |
| "rewards/rejected": -35.0, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 1.3291470434327577, | |
| "grad_norm": 5.5808816018181193e-05, | |
| "learning_rate": 3.410786151595533e-05, | |
| "logits/chosen": -12.3125, | |
| "logits/rejected": -9.875, | |
| "logps/chosen": -996.0, | |
| "logps/rejected": -4096.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.59375, | |
| "rewards/margins": 29.875, | |
| "rewards/rejected": -35.5, | |
| "step": 1270 | |
| }, | |
| { | |
| "epoch": 1.3396127681841967, | |
| "grad_norm": 0.0006105307003125895, | |
| "learning_rate": 3.382347526520765e-05, | |
| "logits/chosen": -12.0625, | |
| "logits/rejected": -9.75, | |
| "logps/chosen": -1008.0, | |
| "logps/rejected": -3984.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.8125, | |
| "rewards/margins": 29.0, | |
| "rewards/rejected": -34.75, | |
| "step": 1280 | |
| }, | |
| { | |
| "epoch": 1.3500784929356358, | |
| "grad_norm": 0.04124806042289618, | |
| "learning_rate": 3.353777872207381e-05, | |
| "logits/chosen": -12.125, | |
| "logits/rejected": -9.8125, | |
| "logps/chosen": -956.0, | |
| "logps/rejected": -4016.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.59375, | |
| "rewards/margins": 29.375, | |
| "rewards/rejected": -35.0, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 1.3605442176870748, | |
| "grad_norm": 0.00028278807215880315, | |
| "learning_rate": 3.325081431269475e-05, | |
| "logits/chosen": -12.0625, | |
| "logits/rejected": -10.0, | |
| "logps/chosen": -996.0, | |
| "logps/rejected": -4128.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.4375, | |
| "rewards/margins": 30.25, | |
| "rewards/rejected": -35.75, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.3605442176870748, | |
| "eval_logits/chosen": -11.5625, | |
| "eval_logits/rejected": -11.25, | |
| "eval_logps/chosen": -1056.0, | |
| "eval_logps/rejected": -1056.0, | |
| "eval_loss": 1.2745000123977661, | |
| "eval_rewards/accuracies": 0.521484375, | |
| "eval_rewards/chosen": -7.40625, | |
| "eval_rewards/margins": 0.296875, | |
| "eval_rewards/rejected": -7.6875, | |
| "eval_runtime": 46.5576, | |
| "eval_samples_per_second": 42.958, | |
| "eval_steps_per_second": 0.687, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.3710099424385138, | |
| "grad_norm": 0.35907688338523, | |
| "learning_rate": 3.296262465149046e-05, | |
| "logits/chosen": -12.0625, | |
| "logits/rejected": -9.9375, | |
| "logps/chosen": -960.0, | |
| "logps/rejected": -4048.0, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.3125, | |
| "rewards/margins": 29.5, | |
| "rewards/rejected": -35.0, | |
| "step": 1310 | |
| }, | |
| { | |
| "epoch": 1.3814756671899528, | |
| "grad_norm": 0.0002966913727874566, | |
| "learning_rate": 3.2673252534831687e-05, | |
| "logits/chosen": -12.5, | |
| "logits/rejected": -10.9375, | |
| "logps/chosen": -872.0, | |
| "logps/rejected": -3648.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.4375, | |
| "rewards/margins": 26.625, | |
| "rewards/rejected": -31.125, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 1.3919413919413919, | |
| "grad_norm": 0.0001702542524251842, | |
| "learning_rate": 3.2382740934684696e-05, | |
| "logits/chosen": -12.5, | |
| "logits/rejected": -11.6875, | |
| "logps/chosen": -812.0, | |
| "logps/rejected": -3728.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.625, | |
| "rewards/margins": 27.875, | |
| "rewards/rejected": -31.5, | |
| "step": 1330 | |
| }, | |
| { | |
| "epoch": 1.402407116692831, | |
| "grad_norm": 0.008402687980743987, | |
| "learning_rate": 3.209113299222982e-05, | |
| "logits/chosen": -12.5, | |
| "logits/rejected": -11.75, | |
| "logps/chosen": -808.0, | |
| "logps/rejected": -3632.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.609375, | |
| "rewards/margins": 27.25, | |
| "rewards/rejected": -30.875, | |
| "step": 1340 | |
| }, | |
| { | |
| "epoch": 1.41287284144427, | |
| "grad_norm": 0.005207977848846674, | |
| "learning_rate": 3.179847201145505e-05, | |
| "logits/chosen": -12.5, | |
| "logits/rejected": -11.5625, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -3904.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 28.875, | |
| "rewards/rejected": -33.0, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.423338566195709, | |
| "grad_norm": 5.948544608208322e-05, | |
| "learning_rate": 3.150480145272528e-05, | |
| "logits/chosen": -12.5, | |
| "logits/rejected": -11.1875, | |
| "logps/chosen": -828.0, | |
| "logps/rejected": -3808.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.25, | |
| "rewards/margins": 29.0, | |
| "rewards/rejected": -33.25, | |
| "step": 1360 | |
| }, | |
| { | |
| "epoch": 1.433804290947148, | |
| "grad_norm": 0.00027024899916237453, | |
| "learning_rate": 3.121016492632848e-05, | |
| "logits/chosen": -12.125, | |
| "logits/rejected": -10.9375, | |
| "logps/chosen": -860.0, | |
| "logps/rejected": -3792.0, | |
| "loss": 0.0026, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.15625, | |
| "rewards/margins": 28.25, | |
| "rewards/rejected": -32.5, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 1.4442700156985873, | |
| "grad_norm": 0.0036263869968399605, | |
| "learning_rate": 3.091460618599951e-05, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -10.3125, | |
| "logps/chosen": -660.0, | |
| "logps/rejected": -3056.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.6875, | |
| "rewards/margins": 22.75, | |
| "rewards/rejected": -25.5, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 1.454735740450026, | |
| "grad_norm": 0.004896822883151655, | |
| "learning_rate": 3.061816912242265e-05, | |
| "logits/chosen": -10.875, | |
| "logits/rejected": -9.9375, | |
| "logps/chosen": -760.0, | |
| "logps/rejected": -3200.0, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.0, | |
| "rewards/margins": 23.125, | |
| "rewards/rejected": -26.125, | |
| "step": 1390 | |
| }, | |
| { | |
| "epoch": 1.4652014652014653, | |
| "grad_norm": 0.004462128873719239, | |
| "learning_rate": 3.0320897756713785e-05, | |
| "logits/chosen": -11.0625, | |
| "logits/rejected": -9.8125, | |
| "logps/chosen": -792.0, | |
| "logps/rejected": -3408.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.4375, | |
| "rewards/margins": 25.375, | |
| "rewards/rejected": -28.875, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.4652014652014653, | |
| "eval_logits/chosen": -10.8125, | |
| "eval_logits/rejected": -10.5, | |
| "eval_logps/chosen": -732.0, | |
| "eval_logps/rejected": -724.0, | |
| "eval_loss": 0.9128749966621399, | |
| "eval_rewards/accuracies": 0.533203125, | |
| "eval_rewards/chosen": -4.15625, | |
| "eval_rewards/margins": 0.216796875, | |
| "eval_rewards/rejected": -4.375, | |
| "eval_runtime": 46.548, | |
| "eval_samples_per_second": 42.966, | |
| "eval_steps_per_second": 0.687, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.4756671899529041, | |
| "grad_norm": 0.0020583665205300378, | |
| "learning_rate": 3.002283623388332e-05, | |
| "logits/chosen": -10.875, | |
| "logits/rejected": -9.5, | |
| "logps/chosen": -812.0, | |
| "logps/rejected": -3488.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.609375, | |
| "rewards/margins": 25.625, | |
| "rewards/rejected": -29.25, | |
| "step": 1410 | |
| }, | |
| { | |
| "epoch": 1.4861329147043434, | |
| "grad_norm": 0.003004299624124082, | |
| "learning_rate": 2.9724028816280507e-05, | |
| "logits/chosen": -11.1875, | |
| "logits/rejected": -9.75, | |
| "logps/chosen": -880.0, | |
| "logps/rejected": -3712.0, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.21875, | |
| "rewards/margins": 27.25, | |
| "rewards/rejected": -31.375, | |
| "step": 1420 | |
| }, | |
| { | |
| "epoch": 1.4965986394557822, | |
| "grad_norm": 1.982507460483849e-05, | |
| "learning_rate": 2.9424519877020524e-05, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -9.3125, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -3840.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.90625, | |
| "rewards/margins": 28.0, | |
| "rewards/rejected": -33.0, | |
| "step": 1430 | |
| }, | |
| { | |
| "epoch": 1.5070643642072215, | |
| "grad_norm": 0.0067003724488342445, | |
| "learning_rate": 2.912435389339504e-05, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -9.625, | |
| "logps/chosen": -1032.0, | |
| "logps/rejected": -4048.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.625, | |
| "rewards/margins": 29.125, | |
| "rewards/rejected": -34.75, | |
| "step": 1440 | |
| }, | |
| { | |
| "epoch": 1.5175300889586603, | |
| "grad_norm": 0.0027638743449127138, | |
| "learning_rate": 2.8823575440267226e-05, | |
| "logits/chosen": -11.5625, | |
| "logits/rejected": -9.25, | |
| "logps/chosen": -964.0, | |
| "logps/rejected": -3968.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.5625, | |
| "rewards/margins": 28.75, | |
| "rewards/rejected": -34.5, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.5279958137100995, | |
| "grad_norm": 0.0006602676320385186, | |
| "learning_rate": 2.8522229183452465e-05, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -8.6875, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -3856.0, | |
| "loss": 0.002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 28.25, | |
| "rewards/rejected": -33.0, | |
| "step": 1460 | |
| }, | |
| { | |
| "epoch": 1.5384615384615383, | |
| "grad_norm": 0.001656705862130333, | |
| "learning_rate": 2.8220359873085373e-05, | |
| "logits/chosen": -11.0, | |
| "logits/rejected": -7.71875, | |
| "logps/chosen": -764.0, | |
| "logps/rejected": -3424.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.328125, | |
| "rewards/margins": 25.375, | |
| "rewards/rejected": -28.75, | |
| "step": 1470 | |
| }, | |
| { | |
| "epoch": 1.5489272632129776, | |
| "grad_norm": 0.001330318003981793, | |
| "learning_rate": 2.7918012336974388e-05, | |
| "logits/chosen": -10.8125, | |
| "logits/rejected": -5.40625, | |
| "logps/chosen": -744.0, | |
| "logps/rejected": -3264.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.015625, | |
| "rewards/margins": 24.5, | |
| "rewards/rejected": -27.375, | |
| "step": 1480 | |
| }, | |
| { | |
| "epoch": 1.5593929879644164, | |
| "grad_norm": 0.0003775179544172728, | |
| "learning_rate": 2.7615231473944815e-05, | |
| "logits/chosen": -10.75, | |
| "logits/rejected": -3.671875, | |
| "logps/chosen": -896.0, | |
| "logps/rejected": -3712.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 27.125, | |
| "rewards/rejected": -31.625, | |
| "step": 1490 | |
| }, | |
| { | |
| "epoch": 1.5698587127158556, | |
| "grad_norm": 0.0014295447451050046, | |
| "learning_rate": 2.7312062247171327e-05, | |
| "logits/chosen": -10.625, | |
| "logits/rejected": -1.59375, | |
| "logps/chosen": -948.0, | |
| "logps/rejected": -3872.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.15625, | |
| "rewards/margins": 28.375, | |
| "rewards/rejected": -33.5, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.5698587127158556, | |
| "eval_logits/chosen": -7.84375, | |
| "eval_logits/rejected": -7.0625, | |
| "eval_logps/chosen": -932.0, | |
| "eval_logps/rejected": -948.0, | |
| "eval_loss": 1.199874997138977, | |
| "eval_rewards/accuracies": 0.544921875, | |
| "eval_rewards/chosen": -6.15625, | |
| "eval_rewards/margins": 0.447265625, | |
| "eval_rewards/rejected": -6.59375, | |
| "eval_runtime": 46.5473, | |
| "eval_samples_per_second": 42.967, | |
| "eval_steps_per_second": 0.687, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.5803244374672945, | |
| "grad_norm": 9.165153900654413e-05, | |
| "learning_rate": 2.7008549677500876e-05, | |
| "logits/chosen": -10.6875, | |
| "logits/rejected": -2.84375, | |
| "logps/chosen": -972.0, | |
| "logps/rejected": -3904.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.3125, | |
| "rewards/margins": 28.25, | |
| "rewards/rejected": -33.5, | |
| "step": 1510 | |
| }, | |
| { | |
| "epoch": 1.5907901622187337, | |
| "grad_norm": 0.0006948860111818887, | |
| "learning_rate": 2.670473883676709e-05, | |
| "logits/chosen": -10.375, | |
| "logits/rejected": -2.734375, | |
| "logps/chosen": -1008.0, | |
| "logps/rejected": -4128.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.59375, | |
| "rewards/margins": 29.875, | |
| "rewards/rejected": -35.5, | |
| "step": 1520 | |
| }, | |
| { | |
| "epoch": 1.6012558869701727, | |
| "grad_norm": 0.00047834848500504076, | |
| "learning_rate": 2.6400674841097074e-05, | |
| "logits/chosen": -10.125, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -900.0, | |
| "logps/rejected": -4064.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.9375, | |
| "rewards/margins": 30.125, | |
| "rewards/rejected": -35.0, | |
| "step": 1530 | |
| }, | |
| { | |
| "epoch": 1.6117216117216118, | |
| "grad_norm": 0.0026260959653123615, | |
| "learning_rate": 2.609640284421158e-05, | |
| "logits/chosen": -10.5625, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -936.0, | |
| "logps/rejected": -3936.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.34375, | |
| "rewards/margins": 28.875, | |
| "rewards/rejected": -34.25, | |
| "step": 1540 | |
| }, | |
| { | |
| "epoch": 1.6221873364730508, | |
| "grad_norm": 0.0012268885261220224, | |
| "learning_rate": 2.5791968030719726e-05, | |
| "logits/chosen": -10.5, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -924.0, | |
| "logps/rejected": -4000.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.03125, | |
| "rewards/margins": 29.5, | |
| "rewards/rejected": -34.5, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.6326530612244898, | |
| "grad_norm": 2.067360267030323e-05, | |
| "learning_rate": 2.548741560940902e-05, | |
| "logits/chosen": -11.1875, | |
| "logits/rejected": -2.203125, | |
| "logps/chosen": -1000.0, | |
| "logps/rejected": -4192.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.71875, | |
| "rewards/margins": 30.875, | |
| "rewards/rejected": -36.5, | |
| "step": 1560 | |
| }, | |
| { | |
| "epoch": 1.6431187859759289, | |
| "grad_norm": 0.00042918312326544427, | |
| "learning_rate": 2.5182790806531787e-05, | |
| "logits/chosen": -10.875, | |
| "logits/rejected": -3.09375, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -3968.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.90625, | |
| "rewards/margins": 29.625, | |
| "rewards/rejected": -34.5, | |
| "step": 1570 | |
| }, | |
| { | |
| "epoch": 1.653584510727368, | |
| "grad_norm": 0.008585642178845426, | |
| "learning_rate": 2.4878138859089074e-05, | |
| "logits/chosen": -11.0625, | |
| "logits/rejected": -3.984375, | |
| "logps/chosen": -856.0, | |
| "logps/rejected": -3968.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 30.0, | |
| "rewards/rejected": -34.5, | |
| "step": 1580 | |
| }, | |
| { | |
| "epoch": 1.664050235478807, | |
| "grad_norm": 0.0007071124042562713, | |
| "learning_rate": 2.4573505008112924e-05, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -3.671875, | |
| "logps/chosen": -956.0, | |
| "logps/rejected": -4160.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.03125, | |
| "rewards/margins": 31.0, | |
| "rewards/rejected": -36.0, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 1.674515960230246, | |
| "grad_norm": 0.0037337809244873334, | |
| "learning_rate": 2.4268934491948032e-05, | |
| "logits/chosen": -11.0, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -868.0, | |
| "logps/rejected": -3872.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 28.75, | |
| "rewards/rejected": -33.5, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.674515960230246, | |
| "eval_logits/chosen": -9.5, | |
| "eval_logits/rejected": -8.75, | |
| "eval_logps/chosen": -892.0, | |
| "eval_logps/rejected": -908.0, | |
| "eval_loss": 1.2006562948226929, | |
| "eval_rewards/accuracies": 0.537109375, | |
| "eval_rewards/chosen": -5.75, | |
| "eval_rewards/margins": 0.443359375, | |
| "eval_rewards/rejected": -6.1875, | |
| "eval_runtime": 46.5715, | |
| "eval_samples_per_second": 42.945, | |
| "eval_steps_per_second": 0.687, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.684981684981685, | |
| "grad_norm": 0.21158098841598244, | |
| "learning_rate": 2.396447253953385e-05, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -3.921875, | |
| "logps/chosen": -944.0, | |
| "logps/rejected": -4128.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.9375, | |
| "rewards/margins": 31.125, | |
| "rewards/rejected": -36.0, | |
| "step": 1610 | |
| }, | |
| { | |
| "epoch": 1.695447409733124, | |
| "grad_norm": 1.1572621980868541e-06, | |
| "learning_rate": 2.3660164363687998e-05, | |
| "logits/chosen": -10.75, | |
| "logits/rejected": -2.796875, | |
| "logps/chosen": -856.0, | |
| "logps/rejected": -4096.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 31.375, | |
| "rewards/rejected": -36.0, | |
| "step": 1620 | |
| }, | |
| { | |
| "epoch": 1.705913134484563, | |
| "grad_norm": 0.002547548810105444, | |
| "learning_rate": 2.33560551543922e-05, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -2.859375, | |
| "logps/chosen": -1000.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.6875, | |
| "rewards/margins": 33.0, | |
| "rewards/rejected": -38.75, | |
| "step": 1630 | |
| }, | |
| { | |
| "epoch": 1.716378859236002, | |
| "grad_norm": 0.005536042521941976, | |
| "learning_rate": 2.3052190072081493e-05, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -2.96875, | |
| "logps/chosen": -1012.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.8125, | |
| "rewards/margins": 33.0, | |
| "rewards/rejected": -38.75, | |
| "step": 1640 | |
| }, | |
| { | |
| "epoch": 1.7268445839874411, | |
| "grad_norm": 5.121219765136539e-06, | |
| "learning_rate": 2.2748614240937864e-05, | |
| "logits/chosen": -10.5625, | |
| "logits/rejected": -0.7421875, | |
| "logps/chosen": -1048.0, | |
| "logps/rejected": -4640.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.21875, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -41.0, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.7373103087388801, | |
| "grad_norm": 3.8121758270966755e-06, | |
| "learning_rate": 2.2445372742189334e-05, | |
| "logits/chosen": -10.9375, | |
| "logits/rejected": -1.125, | |
| "logps/chosen": -1024.0, | |
| "logps/rejected": -4352.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.8125, | |
| "rewards/margins": 32.5, | |
| "rewards/rejected": -38.5, | |
| "step": 1660 | |
| }, | |
| { | |
| "epoch": 1.7477760334903192, | |
| "grad_norm": 0.0004625082060615881, | |
| "learning_rate": 2.214251060741528e-05, | |
| "logits/chosen": -10.5625, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -1152.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.53125, | |
| "rewards/margins": 33.0, | |
| "rewards/rejected": -39.5, | |
| "step": 1670 | |
| }, | |
| { | |
| "epoch": 1.7582417582417582, | |
| "grad_norm": 0.002147631541751008, | |
| "learning_rate": 2.1840072811859302e-05, | |
| "logits/chosen": -11.0, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -1136.0, | |
| "logps/rejected": -4736.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.84375, | |
| "rewards/margins": 35.0, | |
| "rewards/rejected": -41.75, | |
| "step": 1680 | |
| }, | |
| { | |
| "epoch": 1.7687074829931972, | |
| "grad_norm": 0.00026607497101248346, | |
| "learning_rate": 2.1538104267750287e-05, | |
| "logits/chosen": -10.8125, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -1144.0, | |
| "logps/rejected": -4992.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.65625, | |
| "rewards/margins": 37.5, | |
| "rewards/rejected": -44.25, | |
| "step": 1690 | |
| }, | |
| { | |
| "epoch": 1.7791732077446363, | |
| "grad_norm": 9.795478298966919e-05, | |
| "learning_rate": 2.1236649817632952e-05, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -2.03125, | |
| "logps/chosen": -1048.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.125, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -41.0, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.7791732077446363, | |
| "eval_logits/chosen": -8.8125, | |
| "eval_logits/rejected": -8.0625, | |
| "eval_logps/chosen": -1056.0, | |
| "eval_logps/rejected": -1080.0, | |
| "eval_loss": 1.3752343654632568, | |
| "eval_rewards/accuracies": 0.537109375, | |
| "eval_rewards/chosen": -7.34375, | |
| "eval_rewards/margins": 0.56640625, | |
| "eval_rewards/rejected": -7.90625, | |
| "eval_runtime": 46.5567, | |
| "eval_samples_per_second": 42.958, | |
| "eval_steps_per_second": 0.687, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.7896389324960753, | |
| "grad_norm": 0.0013843186105031414, | |
| "learning_rate": 2.093575422770872e-05, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -2.765625, | |
| "logps/chosen": -1064.0, | |
| "logps/rejected": -4704.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.09375, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -41.5, | |
| "step": 1710 | |
| }, | |
| { | |
| "epoch": 1.8001046572475143, | |
| "grad_norm": 0.002905622017159022, | |
| "learning_rate": 2.0635462181187827e-05, | |
| "logits/chosen": -10.8125, | |
| "logits/rejected": -2.6875, | |
| "logps/chosen": -1056.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.25, | |
| "rewards/margins": 33.75, | |
| "rewards/rejected": -40.0, | |
| "step": 1720 | |
| }, | |
| { | |
| "epoch": 1.8105703819989536, | |
| "grad_norm": 4.7350301163257546e-05, | |
| "learning_rate": 2.0335818271653917e-05, | |
| "logits/chosen": -10.9375, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -1012.0, | |
| "logps/rejected": -4672.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.03125, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -41.25, | |
| "step": 1730 | |
| }, | |
| { | |
| "epoch": 1.8210361067503924, | |
| "grad_norm": 3.38378971960805e-05, | |
| "learning_rate": 2.0036866996441815e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -2.921875, | |
| "logps/chosen": -1040.0, | |
| "logps/rejected": -4736.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.0625, | |
| "rewards/margins": 36.0, | |
| "rewards/rejected": -42.0, | |
| "step": 1740 | |
| }, | |
| { | |
| "epoch": 1.8315018315018317, | |
| "grad_norm": 0.052277678924924235, | |
| "learning_rate": 1.9738652750029677e-05, | |
| "logits/chosen": -11.4375, | |
| "logits/rejected": -2.046875, | |
| "logps/chosen": -1048.0, | |
| "logps/rejected": -4672.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.03125, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -41.75, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.8419675562532705, | |
| "grad_norm": 3.28661269673869e-05, | |
| "learning_rate": 1.944121981744631e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -2.765625, | |
| "logps/chosen": -948.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.4375, | |
| "rewards/margins": 34.0, | |
| "rewards/rejected": -39.5, | |
| "step": 1760 | |
| }, | |
| { | |
| "epoch": 1.8524332810047097, | |
| "grad_norm": 0.004398881896200284, | |
| "learning_rate": 1.9144612367694887e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -4.03125, | |
| "logps/chosen": -968.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.4375, | |
| "rewards/margins": 34.25, | |
| "rewards/rejected": -39.75, | |
| "step": 1770 | |
| }, | |
| { | |
| "epoch": 1.8628990057561485, | |
| "grad_norm": 3.030559136631043e-06, | |
| "learning_rate": 1.8848874447193803e-05, | |
| "logits/chosen": -11.625, | |
| "logits/rejected": -3.09375, | |
| "logps/chosen": -996.0, | |
| "logps/rejected": -4640.0, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.5625, | |
| "rewards/margins": 35.75, | |
| "rewards/rejected": -41.25, | |
| "step": 1780 | |
| }, | |
| { | |
| "epoch": 1.8733647305075878, | |
| "grad_norm": 3.533833902136687e-05, | |
| "learning_rate": 1.855404997323571e-05, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -3.984375, | |
| "logps/chosen": -948.0, | |
| "logps/rejected": -4352.0, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.96875, | |
| "rewards/margins": 33.25, | |
| "rewards/rejected": -38.25, | |
| "step": 1790 | |
| }, | |
| { | |
| "epoch": 1.8838304552590266, | |
| "grad_norm": 0.03132437985031672, | |
| "learning_rate": 1.82601827274658e-05, | |
| "logits/chosen": -11.75, | |
| "logits/rejected": -3.21875, | |
| "logps/chosen": -996.0, | |
| "logps/rejected": -4640.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.75, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -41.25, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.8838304552590266, | |
| "eval_logits/chosen": -9.75, | |
| "eval_logits/rejected": -9.0625, | |
| "eval_logps/chosen": -976.0, | |
| "eval_logps/rejected": -1000.0, | |
| "eval_loss": 1.273687481880188, | |
| "eval_rewards/accuracies": 0.546875, | |
| "eval_rewards/chosen": -6.5625, | |
| "eval_rewards/margins": 0.55078125, | |
| "eval_rewards/rejected": -7.125, | |
| "eval_runtime": 46.554, | |
| "eval_samples_per_second": 42.961, | |
| "eval_steps_per_second": 0.687, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.8942961800104658, | |
| "grad_norm": 0.012605198825709478, | |
| "learning_rate": 1.7967316349380213e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -3.09375, | |
| "logps/chosen": -1056.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.78125, | |
| "rewards/margins": 33.25, | |
| "rewards/rejected": -39.0, | |
| "step": 1810 | |
| }, | |
| { | |
| "epoch": 1.9047619047619047, | |
| "grad_norm": 0.00012143358410047413, | |
| "learning_rate": 1.7675494329845516e-05, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -4.125, | |
| "logps/chosen": -1056.0, | |
| "logps/rejected": -4928.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.96875, | |
| "rewards/margins": 37.0, | |
| "rewards/rejected": -43.0, | |
| "step": 1820 | |
| }, | |
| { | |
| "epoch": 1.915227629513344, | |
| "grad_norm": 1.616432794551237e-06, | |
| "learning_rate": 1.738476000464026e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -3.046875, | |
| "logps/chosen": -1064.0, | |
| "logps/rejected": -4800.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.0, | |
| "rewards/margins": 36.5, | |
| "rewards/rejected": -42.5, | |
| "step": 1830 | |
| }, | |
| { | |
| "epoch": 1.9256933542647827, | |
| "grad_norm": 0.0006497808004301234, | |
| "learning_rate": 1.7095156548019646e-05, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -1.7734375, | |
| "logps/chosen": -980.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.71875, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -40.5, | |
| "step": 1840 | |
| }, | |
| { | |
| "epoch": 1.936159079016222, | |
| "grad_norm": 0.0006291463992247915, | |
| "learning_rate": 1.6806726966304064e-05, | |
| "logits/chosen": -11.5625, | |
| "logits/rejected": -2.703125, | |
| "logps/chosen": -1032.0, | |
| "logps/rejected": -4832.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.90625, | |
| "rewards/margins": 37.0, | |
| "rewards/rejected": -43.0, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 1.9466248037676608, | |
| "grad_norm": 5.564010651180756e-07, | |
| "learning_rate": 1.6519514091492625e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -2.640625, | |
| "logps/chosen": -1120.0, | |
| "logps/rejected": -4864.0, | |
| "loss": 0.0002, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -6.59375, | |
| "rewards/margins": 36.25, | |
| "rewards/rejected": -42.75, | |
| "step": 1860 | |
| }, | |
| { | |
| "epoch": 1.9570905285191, | |
| "grad_norm": 9.808795647131748e-07, | |
| "learning_rate": 1.6233560574902627e-05, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -1000.0, | |
| "logps/rejected": -4640.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.71875, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -41.0, | |
| "step": 1870 | |
| }, | |
| { | |
| "epoch": 1.9675562532705388, | |
| "grad_norm": 0.0010134580631731241, | |
| "learning_rate": 1.5948908880835753e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -3.015625, | |
| "logps/chosen": -1004.0, | |
| "logps/rejected": -4640.0, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.53125, | |
| "rewards/margins": 35.0, | |
| "rewards/rejected": -40.5, | |
| "step": 1880 | |
| }, | |
| { | |
| "epoch": 1.978021978021978, | |
| "grad_norm": 0.0023072751959159333, | |
| "learning_rate": 1.5665601280272124e-05, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -4.375, | |
| "logps/chosen": -868.0, | |
| "logps/rejected": -4256.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.21875, | |
| "rewards/margins": 33.0, | |
| "rewards/rejected": -37.25, | |
| "step": 1890 | |
| }, | |
| { | |
| "epoch": 1.988487702773417, | |
| "grad_norm": 1.6369002270159292e-05, | |
| "learning_rate": 1.538367984459298e-05, | |
| "logits/chosen": -11.625, | |
| "logits/rejected": -5.25, | |
| "logps/chosen": -792.0, | |
| "logps/rejected": -4096.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.65625, | |
| "rewards/margins": 31.875, | |
| "rewards/rejected": -35.5, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.988487702773417, | |
| "eval_logits/chosen": -10.8125, | |
| "eval_logits/rejected": -10.25, | |
| "eval_logps/chosen": -780.0, | |
| "eval_logps/rejected": -784.0, | |
| "eval_loss": 1.0199531316757202, | |
| "eval_rewards/accuracies": 0.5390625, | |
| "eval_rewards/chosen": -4.625, | |
| "eval_rewards/margins": 0.296875, | |
| "eval_rewards/rejected": -4.9375, | |
| "eval_runtime": 46.5587, | |
| "eval_samples_per_second": 42.957, | |
| "eval_steps_per_second": 0.687, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.9989534275248562, | |
| "grad_norm": 0.00012666517057991159, | |
| "learning_rate": 1.5103186439333133e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -5.125, | |
| "logps/chosen": -836.0, | |
| "logps/rejected": -4160.0, | |
| "loss": 0.0003, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.9375, | |
| "rewards/margins": 32.5, | |
| "rewards/rejected": -36.5, | |
| "step": 1910 | |
| }, | |
| { | |
| "epoch": 2.009419152276295, | |
| "grad_norm": 0.000585596591254348, | |
| "learning_rate": 1.4824162717963829e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -4.90625, | |
| "logps/chosen": -768.0, | |
| "logps/rejected": -4096.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.640625, | |
| "rewards/margins": 32.0, | |
| "rewards/rejected": -35.75, | |
| "step": 1920 | |
| }, | |
| { | |
| "epoch": 2.0198848770277342, | |
| "grad_norm": 5.946201249043237e-05, | |
| "learning_rate": 1.4546650115707192e-05, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -4.71875, | |
| "logps/chosen": -864.0, | |
| "logps/rejected": -4080.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.09375, | |
| "rewards/margins": 31.5, | |
| "rewards/rejected": -35.5, | |
| "step": 1930 | |
| }, | |
| { | |
| "epoch": 2.030350601779173, | |
| "grad_norm": 9.153183176406476e-06, | |
| "learning_rate": 1.4270689843383111e-05, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -4.34375, | |
| "logps/chosen": -900.0, | |
| "logps/rejected": -4320.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 33.0, | |
| "rewards/rejected": -37.75, | |
| "step": 1940 | |
| }, | |
| { | |
| "epoch": 2.0408163265306123, | |
| "grad_norm": 8.501992922069672e-06, | |
| "learning_rate": 1.3996322881289347e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -5.0625, | |
| "logps/chosen": -876.0, | |
| "logps/rejected": -4384.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 33.75, | |
| "rewards/rejected": -38.0, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 2.051282051282051, | |
| "grad_norm": 1.2025628151003164e-05, | |
| "learning_rate": 1.3723589973115963e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -5.3125, | |
| "logps/chosen": -872.0, | |
| "logps/rejected": -4288.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 33.0, | |
| "rewards/rejected": -37.25, | |
| "step": 1960 | |
| }, | |
| { | |
| "epoch": 2.0617477760334904, | |
| "grad_norm": 0.0014306770131796796, | |
| "learning_rate": 1.3452531619894837e-05, | |
| "logits/chosen": -11.75, | |
| "logits/rejected": -4.84375, | |
| "logps/chosen": -900.0, | |
| "logps/rejected": -4384.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.34375, | |
| "rewards/margins": 33.75, | |
| "rewards/rejected": -38.0, | |
| "step": 1970 | |
| }, | |
| { | |
| "epoch": 2.072213500784929, | |
| "grad_norm": 0.00047624732455137683, | |
| "learning_rate": 1.3183188073985245e-05, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -5.4375, | |
| "logps/chosen": -844.0, | |
| "logps/rejected": -4224.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 32.5, | |
| "rewards/rejected": -36.75, | |
| "step": 1980 | |
| }, | |
| { | |
| "epoch": 2.0826792255363684, | |
| "grad_norm": 2.504797049865482e-05, | |
| "learning_rate": 1.2915599333096351e-05, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -5.15625, | |
| "logps/chosen": -872.0, | |
| "logps/rejected": -4384.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 34.25, | |
| "rewards/rejected": -38.5, | |
| "step": 1990 | |
| }, | |
| { | |
| "epoch": 2.0931449502878072, | |
| "grad_norm": 0.00011623551249806416, | |
| "learning_rate": 1.264980513434752e-05, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -876.0, | |
| "logps/rejected": -4320.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.40625, | |
| "rewards/margins": 33.25, | |
| "rewards/rejected": -37.75, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.0931449502878072, | |
| "eval_logits/chosen": -10.375, | |
| "eval_logits/rejected": -9.75, | |
| "eval_logps/chosen": -844.0, | |
| "eval_logps/rejected": -852.0, | |
| "eval_loss": 1.0691250562667847, | |
| "eval_rewards/accuracies": 0.544921875, | |
| "eval_rewards/chosen": -5.25, | |
| "eval_rewards/margins": 0.392578125, | |
| "eval_rewards/rejected": -5.65625, | |
| "eval_runtime": 46.5585, | |
| "eval_samples_per_second": 42.957, | |
| "eval_steps_per_second": 0.687, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.1036106750392465, | |
| "grad_norm": 1.960546048198272e-06, | |
| "learning_rate": 1.2385844948367324e-05, | |
| "logits/chosen": -11.4375, | |
| "logits/rejected": -4.5625, | |
| "logps/chosen": -812.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 34.5, | |
| "rewards/rejected": -38.5, | |
| "step": 2010 | |
| }, | |
| { | |
| "epoch": 2.1140763997906853, | |
| "grad_norm": 2.2954329059859816e-05, | |
| "learning_rate": 1.2123757973432114e-05, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -4.15625, | |
| "logps/chosen": -888.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 34.5, | |
| "rewards/rejected": -39.0, | |
| "step": 2020 | |
| }, | |
| { | |
| "epoch": 2.1245421245421245, | |
| "grad_norm": 0.005708422002054625, | |
| "learning_rate": 1.1863583129645106e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -3.78125, | |
| "logps/chosen": -920.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 34.0, | |
| "rewards/rejected": -38.75, | |
| "step": 2030 | |
| }, | |
| { | |
| "epoch": 2.1350078492935634, | |
| "grad_norm": 0.0003138079506406525, | |
| "learning_rate": 1.1605359053156607e-05, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -4.96875, | |
| "logps/chosen": -864.0, | |
| "logps/rejected": -4384.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.3125, | |
| "rewards/margins": 33.75, | |
| "rewards/rejected": -38.0, | |
| "step": 2040 | |
| }, | |
| { | |
| "epoch": 2.1454735740450026, | |
| "grad_norm": 0.0012141848676057905, | |
| "learning_rate": 1.1349124090426644e-05, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -3.625, | |
| "logps/chosen": -864.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.34375, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -39.0, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 2.155939298796442, | |
| "grad_norm": 0.0012437294678055535, | |
| "learning_rate": 1.1094916292530406e-05, | |
| "logits/chosen": -11.75, | |
| "logits/rejected": -4.3125, | |
| "logps/chosen": -856.0, | |
| "logps/rejected": -4352.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.375, | |
| "rewards/margins": 33.75, | |
| "rewards/rejected": -38.0, | |
| "step": 2060 | |
| }, | |
| { | |
| "epoch": 2.1664050235478807, | |
| "grad_norm": 0.05976160753233462, | |
| "learning_rate": 1.0842773409507623e-05, | |
| "logits/chosen": -11.1875, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -912.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -39.0, | |
| "step": 2070 | |
| }, | |
| { | |
| "epoch": 2.17687074829932, | |
| "grad_norm": 0.0001455265096636152, | |
| "learning_rate": 1.0592732884756754e-05, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -4.28125, | |
| "logps/chosen": -896.0, | |
| "logps/rejected": -4384.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.34375, | |
| "rewards/margins": 34.0, | |
| "rewards/rejected": -38.5, | |
| "step": 2080 | |
| }, | |
| { | |
| "epoch": 2.1873364730507587, | |
| "grad_norm": 0.0006494428551710046, | |
| "learning_rate": 1.0344831849474507e-05, | |
| "logits/chosen": -11.5625, | |
| "logits/rejected": -4.0, | |
| "logps/chosen": -932.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.625, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -39.75, | |
| "step": 2090 | |
| }, | |
| { | |
| "epoch": 2.197802197802198, | |
| "grad_norm": 9.7942225082266e-06, | |
| "learning_rate": 1.0099107117141879e-05, | |
| "logits/chosen": -11.0625, | |
| "logits/rejected": -3.578125, | |
| "logps/chosen": -812.0, | |
| "logps/rejected": -4256.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 33.25, | |
| "rewards/rejected": -37.5, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.197802197802198, | |
| "eval_logits/chosen": -10.0, | |
| "eval_logits/rejected": -9.375, | |
| "eval_logps/chosen": -884.0, | |
| "eval_logps/rejected": -896.0, | |
| "eval_loss": 1.114484429359436, | |
| "eval_rewards/accuracies": 0.546875, | |
| "eval_rewards/chosen": -5.625, | |
| "eval_rewards/margins": 0.453125, | |
| "eval_rewards/rejected": -6.09375, | |
| "eval_runtime": 46.5516, | |
| "eval_samples_per_second": 42.963, | |
| "eval_steps_per_second": 0.687, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.208267922553637, | |
| "grad_norm": 9.805975439873259e-05, | |
| "learning_rate": 9.855595178057334e-06, | |
| "logits/chosen": -11.1875, | |
| "logits/rejected": -2.984375, | |
| "logps/chosen": -876.0, | |
| "logps/rejected": -4352.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 33.25, | |
| "rewards/rejected": -38.25, | |
| "step": 2110 | |
| }, | |
| { | |
| "epoch": 2.218733647305076, | |
| "grad_norm": 1.4384859516953603e-06, | |
| "learning_rate": 9.614332193917949e-06, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -3.03125, | |
| "logps/chosen": -880.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 34.25, | |
| "rewards/rejected": -39.0, | |
| "step": 2120 | |
| }, | |
| { | |
| "epoch": 2.229199372056515, | |
| "grad_norm": 1.5284785114349792e-05, | |
| "learning_rate": 9.375353992449385e-06, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -4.5, | |
| "logps/chosen": -936.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.71875, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -40.0, | |
| "step": 2130 | |
| }, | |
| { | |
| "epoch": 2.239665096807954, | |
| "grad_norm": 1.4701709528243693e-06, | |
| "learning_rate": 9.138696062085442e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -900.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 34.0, | |
| "rewards/rejected": -38.75, | |
| "step": 2140 | |
| }, | |
| { | |
| "epoch": 2.250130821559393, | |
| "grad_norm": 6.291270296692501e-05, | |
| "learning_rate": 8.904393546698007e-06, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -3.125, | |
| "logps/chosen": -884.0, | |
| "logps/rejected": -4384.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 33.75, | |
| "rewards/rejected": -38.5, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 2.260596546310832, | |
| "grad_norm": 8.733672614413915e-05, | |
| "learning_rate": 8.672481240378142e-06, | |
| "logits/chosen": -11.625, | |
| "logits/rejected": -3.265625, | |
| "logps/chosen": -928.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -40.0, | |
| "step": 2160 | |
| }, | |
| { | |
| "epoch": 2.271062271062271, | |
| "grad_norm": 3.2513697897518936e-07, | |
| "learning_rate": 8.44299358226919e-06, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -880.0, | |
| "logps/rejected": -4320.0, | |
| "loss": 0.0008, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 33.75, | |
| "rewards/rejected": -38.5, | |
| "step": 2170 | |
| }, | |
| { | |
| "epoch": 2.2815279958137102, | |
| "grad_norm": 3.557581070325729e-06, | |
| "learning_rate": 8.215964651452457e-06, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -2.796875, | |
| "logps/chosen": -872.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -39.5, | |
| "step": 2180 | |
| }, | |
| { | |
| "epoch": 2.291993720565149, | |
| "grad_norm": 0.003862782284187376, | |
| "learning_rate": 7.991428161886502e-06, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -3.046875, | |
| "logps/chosen": -912.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 35.75, | |
| "rewards/rejected": -40.25, | |
| "step": 2190 | |
| }, | |
| { | |
| "epoch": 2.3024594453165883, | |
| "grad_norm": 6.602391329335526e-05, | |
| "learning_rate": 7.76941745740061e-06, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -900.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.78125, | |
| "rewards/margins": 34.25, | |
| "rewards/rejected": -39.25, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.3024594453165883, | |
| "eval_logits/chosen": -9.8125, | |
| "eval_logits/rejected": -9.125, | |
| "eval_logps/chosen": -900.0, | |
| "eval_logps/rejected": -920.0, | |
| "eval_loss": 1.135671854019165, | |
| "eval_rewards/accuracies": 0.552734375, | |
| "eval_rewards/chosen": -5.8125, | |
| "eval_rewards/margins": 0.4765625, | |
| "eval_rewards/rejected": -6.3125, | |
| "eval_runtime": 46.5523, | |
| "eval_samples_per_second": 42.962, | |
| "eval_steps_per_second": 0.687, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.312925170068027, | |
| "grad_norm": 1.252013711916386e-07, | |
| "learning_rate": 7.549965506743123e-06, | |
| "logits/chosen": -11.0625, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -968.0, | |
| "logps/rejected": -4768.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.0, | |
| "rewards/margins": 36.75, | |
| "rewards/rejected": -41.75, | |
| "step": 2210 | |
| }, | |
| { | |
| "epoch": 2.3233908948194664, | |
| "grad_norm": 1.0130321299142165e-05, | |
| "learning_rate": 7.333104898685631e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -2.515625, | |
| "logps/chosen": -928.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -39.5, | |
| "step": 2220 | |
| }, | |
| { | |
| "epoch": 2.333856619570905, | |
| "grad_norm": 0.0008903750504928692, | |
| "learning_rate": 7.118867837183518e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -3.359375, | |
| "logps/chosen": -940.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.84375, | |
| "rewards/margins": 34.5, | |
| "rewards/rejected": -39.5, | |
| "step": 2230 | |
| }, | |
| { | |
| "epoch": 2.3443223443223444, | |
| "grad_norm": 5.636643522137541e-06, | |
| "learning_rate": 6.9072861365936045e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -2.90625, | |
| "logps/chosen": -916.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.78125, | |
| "rewards/margins": 33.75, | |
| "rewards/rejected": -38.75, | |
| "step": 2240 | |
| }, | |
| { | |
| "epoch": 2.3547880690737832, | |
| "grad_norm": 0.0017077186052231546, | |
| "learning_rate": 6.6983912169497005e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -2.796875, | |
| "logps/chosen": -940.0, | |
| "logps/rejected": -4640.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.96875, | |
| "rewards/margins": 35.75, | |
| "rewards/rejected": -40.75, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 2.3652537938252225, | |
| "grad_norm": 2.591780632012775e-05, | |
| "learning_rate": 6.49221409929677e-06, | |
| "logits/chosen": -11.75, | |
| "logits/rejected": -3.3125, | |
| "logps/chosen": -892.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.625, | |
| "rewards/margins": 35.0, | |
| "rewards/rejected": -39.75, | |
| "step": 2260 | |
| }, | |
| { | |
| "epoch": 2.3757195185766613, | |
| "grad_norm": 2.453915929068766e-05, | |
| "learning_rate": 6.2887854010842075e-06, | |
| "logits/chosen": -11.0, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -868.0, | |
| "logps/rejected": -4320.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.59375, | |
| "rewards/margins": 33.5, | |
| "rewards/rejected": -38.0, | |
| "step": 2270 | |
| }, | |
| { | |
| "epoch": 2.3861852433281006, | |
| "grad_norm": 0.0009595694975992897, | |
| "learning_rate": 6.0881353316191385e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -3.46875, | |
| "logps/chosen": -912.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -40.0, | |
| "step": 2280 | |
| }, | |
| { | |
| "epoch": 2.3966509680795394, | |
| "grad_norm": 0.0007022262841534255, | |
| "learning_rate": 5.890293687580381e-06, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -940.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.9375, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -40.25, | |
| "step": 2290 | |
| }, | |
| { | |
| "epoch": 2.4071166928309786, | |
| "grad_norm": 0.0028605969196156523, | |
| "learning_rate": 5.695289848593532e-06, | |
| "logits/chosen": -11.0, | |
| "logits/rejected": -3.515625, | |
| "logps/chosen": -904.0, | |
| "logps/rejected": -4640.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.84375, | |
| "rewards/margins": 35.75, | |
| "rewards/rejected": -40.75, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.4071166928309786, | |
| "eval_logits/chosen": -9.8125, | |
| "eval_logits/rejected": -9.1875, | |
| "eval_logps/chosen": -900.0, | |
| "eval_logps/rejected": -916.0, | |
| "eval_loss": 1.136171817779541, | |
| "eval_rewards/accuracies": 0.546875, | |
| "eval_rewards/chosen": -5.8125, | |
| "eval_rewards/margins": 0.4765625, | |
| "eval_rewards/rejected": -6.28125, | |
| "eval_runtime": 46.5545, | |
| "eval_samples_per_second": 42.96, | |
| "eval_steps_per_second": 0.687, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.4175824175824174, | |
| "grad_norm": 0.0012142728887248831, | |
| "learning_rate": 5.50315277286812e-06, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -3.296875, | |
| "logps/chosen": -900.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 36.0, | |
| "rewards/rejected": -40.75, | |
| "step": 2310 | |
| }, | |
| { | |
| "epoch": 2.4280481423338567, | |
| "grad_norm": 1.2994182727012979e-05, | |
| "learning_rate": 5.313910992897281e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -2.875, | |
| "logps/chosen": -872.0, | |
| "logps/rejected": -4288.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 33.5, | |
| "rewards/rejected": -38.0, | |
| "step": 2320 | |
| }, | |
| { | |
| "epoch": 2.4385138670852955, | |
| "grad_norm": 2.535510494647404e-05, | |
| "learning_rate": 5.127592611220658e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -2.890625, | |
| "logps/chosen": -916.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 34.0, | |
| "rewards/rejected": -38.75, | |
| "step": 2330 | |
| }, | |
| { | |
| "epoch": 2.4489795918367347, | |
| "grad_norm": 2.5436397997093715e-05, | |
| "learning_rate": 4.944225296251159e-06, | |
| "logits/chosen": -10.9375, | |
| "logits/rejected": -2.09375, | |
| "logps/chosen": -864.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.59375, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -40.0, | |
| "step": 2340 | |
| }, | |
| { | |
| "epoch": 2.4594453165881736, | |
| "grad_norm": 0.000497992376966425, | |
| "learning_rate": 4.763836278166172e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -3.171875, | |
| "logps/chosen": -944.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.09375, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -39.75, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 2.469911041339613, | |
| "grad_norm": 6.544326959877835e-05, | |
| "learning_rate": 4.586452344863876e-06, | |
| "logits/chosen": -10.9375, | |
| "logits/rejected": -3.1875, | |
| "logps/chosen": -888.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.625, | |
| "rewards/margins": 35.0, | |
| "rewards/rejected": -39.5, | |
| "step": 2360 | |
| }, | |
| { | |
| "epoch": 2.4803767660910516, | |
| "grad_norm": 0.0005882616755124241, | |
| "learning_rate": 4.412099837985192e-06, | |
| "logits/chosen": -11.875, | |
| "logits/rejected": -4.53125, | |
| "logps/chosen": -880.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.59375, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -39.75, | |
| "step": 2370 | |
| }, | |
| { | |
| "epoch": 2.490842490842491, | |
| "grad_norm": 0.0016919065831421927, | |
| "learning_rate": 4.240804649002089e-06, | |
| "logits/chosen": -11.5625, | |
| "logits/rejected": -4.0625, | |
| "logps/chosen": -912.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -39.5, | |
| "step": 2380 | |
| }, | |
| { | |
| "epoch": 2.50130821559393, | |
| "grad_norm": 0.003009874916161143, | |
| "learning_rate": 4.072592215372598e-06, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -3.4375, | |
| "logps/chosen": -940.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.90625, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -40.25, | |
| "step": 2390 | |
| }, | |
| { | |
| "epoch": 2.511773940345369, | |
| "grad_norm": 0.001457127466152037, | |
| "learning_rate": 3.907487516763389e-06, | |
| "logits/chosen": -11.4375, | |
| "logits/rejected": -3.5625, | |
| "logps/chosen": -884.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.375, | |
| "rewards/margins": 35.0, | |
| "rewards/rejected": -39.5, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.511773940345369, | |
| "eval_logits/chosen": -10.0625, | |
| "eval_logits/rejected": -9.4375, | |
| "eval_logps/chosen": -892.0, | |
| "eval_logps/rejected": -908.0, | |
| "eval_loss": 1.135312557220459, | |
| "eval_rewards/accuracies": 0.54296875, | |
| "eval_rewards/chosen": -5.71875, | |
| "eval_rewards/margins": 0.46875, | |
| "eval_rewards/rejected": -6.1875, | |
| "eval_runtime": 46.5508, | |
| "eval_samples_per_second": 42.964, | |
| "eval_steps_per_second": 0.687, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.5222396650968077, | |
| "grad_norm": 2.2924777468764287e-05, | |
| "learning_rate": 3.745515071340286e-06, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -2.9375, | |
| "logps/chosen": -888.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 35.0, | |
| "rewards/rejected": -39.5, | |
| "step": 2410 | |
| }, | |
| { | |
| "epoch": 2.532705389848247, | |
| "grad_norm": 5.3594521442316194e-05, | |
| "learning_rate": 3.586698932127236e-06, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -3.671875, | |
| "logps/chosen": -900.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.59375, | |
| "rewards/margins": 35.75, | |
| "rewards/rejected": -40.25, | |
| "step": 2420 | |
| }, | |
| { | |
| "epoch": 2.5431711145996863, | |
| "grad_norm": 1.4473343352749253e-05, | |
| "learning_rate": 3.4310626834344744e-06, | |
| "logits/chosen": -11.5625, | |
| "logits/rejected": -3.53125, | |
| "logps/chosen": -892.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 36.0, | |
| "rewards/rejected": -40.5, | |
| "step": 2430 | |
| }, | |
| { | |
| "epoch": 2.553636839351125, | |
| "grad_norm": 1.0155732198196766e-05, | |
| "learning_rate": 3.2786294373562343e-06, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -3.0, | |
| "logps/chosen": -860.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.34375, | |
| "rewards/margins": 36.25, | |
| "rewards/rejected": -40.5, | |
| "step": 2440 | |
| }, | |
| { | |
| "epoch": 2.564102564102564, | |
| "grad_norm": 0.0006268784876083323, | |
| "learning_rate": 3.1294218303385526e-06, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -3.609375, | |
| "logps/chosen": -924.0, | |
| "logps/rejected": -4384.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 34.0, | |
| "rewards/rejected": -38.5, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 2.574568288854003, | |
| "grad_norm": 0.00012573484995757477, | |
| "learning_rate": 2.9834620198177483e-06, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -3.015625, | |
| "logps/chosen": -916.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.59375, | |
| "rewards/margins": 35.0, | |
| "rewards/rejected": -39.75, | |
| "step": 2460 | |
| }, | |
| { | |
| "epoch": 2.5850340136054424, | |
| "grad_norm": 0.0005877799861677143, | |
| "learning_rate": 2.840771680930068e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -4.34375, | |
| "logps/chosen": -860.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -40.0, | |
| "step": 2470 | |
| }, | |
| { | |
| "epoch": 2.595499738356881, | |
| "grad_norm": 0.019708445434804028, | |
| "learning_rate": 2.701372003292868e-06, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -2.765625, | |
| "logps/chosen": -920.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -39.5, | |
| "step": 2480 | |
| }, | |
| { | |
| "epoch": 2.60596546310832, | |
| "grad_norm": 0.00039224001765805105, | |
| "learning_rate": 2.56528368785795e-06, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -3.421875, | |
| "logps/chosen": -960.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.96875, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -40.25, | |
| "step": 2490 | |
| }, | |
| { | |
| "epoch": 2.6164311878597593, | |
| "grad_norm": 8.294110409370851e-05, | |
| "learning_rate": 2.432526943837471e-06, | |
| "logits/chosen": -11.625, | |
| "logits/rejected": -3.4375, | |
| "logps/chosen": -884.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.625, | |
| "rewards/margins": 35.0, | |
| "rewards/rejected": -39.75, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.6164311878597593, | |
| "eval_logits/chosen": -10.1875, | |
| "eval_logits/rejected": -9.5625, | |
| "eval_logps/chosen": -888.0, | |
| "eval_logps/rejected": -904.0, | |
| "eval_loss": 1.1318438053131104, | |
| "eval_rewards/accuracies": 0.5390625, | |
| "eval_rewards/chosen": -5.6875, | |
| "eval_rewards/margins": 0.462890625, | |
| "eval_rewards/rejected": -6.15625, | |
| "eval_runtime": 46.5453, | |
| "eval_samples_per_second": 42.969, | |
| "eval_steps_per_second": 0.688, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.6268969126111985, | |
| "grad_norm": 0.0002614179739901724, | |
| "learning_rate": 2.3031214857028446e-06, | |
| "logits/chosen": -11.75, | |
| "logits/rejected": -3.65625, | |
| "logps/chosen": -904.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -39.25, | |
| "step": 2510 | |
| }, | |
| { | |
| "epoch": 2.6373626373626373, | |
| "grad_norm": 0.002290461011665828, | |
| "learning_rate": 2.177086530257139e-06, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -3.484375, | |
| "logps/chosen": -860.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -40.0, | |
| "step": 2520 | |
| }, | |
| { | |
| "epoch": 2.647828362114076, | |
| "grad_norm": 1.608673212544125e-05, | |
| "learning_rate": 2.054440793781354e-06, | |
| "logits/chosen": -11.5625, | |
| "logits/rejected": -3.90625, | |
| "logps/chosen": -920.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.84375, | |
| "rewards/margins": 34.0, | |
| "rewards/rejected": -38.75, | |
| "step": 2530 | |
| }, | |
| { | |
| "epoch": 2.6582940868655154, | |
| "grad_norm": 0.0012610875214715503, | |
| "learning_rate": 1.9352024892550587e-06, | |
| "logits/chosen": -11.75, | |
| "logits/rejected": -3.265625, | |
| "logps/chosen": -924.0, | |
| "logps/rejected": -4736.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.875, | |
| "rewards/margins": 36.75, | |
| "rewards/rejected": -41.75, | |
| "step": 2540 | |
| }, | |
| { | |
| "epoch": 2.6687598116169546, | |
| "grad_norm": 2.195115992670894e-06, | |
| "learning_rate": 1.8193893236517379e-06, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -4.34375, | |
| "logps/chosen": -956.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.84375, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -40.0, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 2.6792255363683934, | |
| "grad_norm": 1.858040983703488e-05, | |
| "learning_rate": 1.7070184953092956e-06, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -3.515625, | |
| "logps/chosen": -932.0, | |
| "logps/rejected": -4672.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 36.5, | |
| "rewards/rejected": -41.0, | |
| "step": 2560 | |
| }, | |
| { | |
| "epoch": 2.6896912611198327, | |
| "grad_norm": 8.011981053250268e-06, | |
| "learning_rate": 1.5981066913760918e-06, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -4.125, | |
| "logps/chosen": -860.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.3125, | |
| "rewards/margins": 35.75, | |
| "rewards/rejected": -40.0, | |
| "step": 2570 | |
| }, | |
| { | |
| "epoch": 2.7001569858712715, | |
| "grad_norm": 6.018937999488517e-06, | |
| "learning_rate": 1.492670085332884e-06, | |
| "logits/chosen": -11.8125, | |
| "logits/rejected": -3.328125, | |
| "logps/chosen": -912.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 34.5, | |
| "rewards/rejected": -39.25, | |
| "step": 2580 | |
| }, | |
| { | |
| "epoch": 2.7106227106227108, | |
| "grad_norm": 3.06985736347809e-05, | |
| "learning_rate": 1.3907243345910787e-06, | |
| "logits/chosen": -11.3125, | |
| "logits/rejected": -3.28125, | |
| "logps/chosen": -920.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 35.75, | |
| "rewards/rejected": -40.25, | |
| "step": 2590 | |
| }, | |
| { | |
| "epoch": 2.7210884353741496, | |
| "grad_norm": 0.0007132640154626052, | |
| "learning_rate": 1.292284578167538e-06, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -4.0, | |
| "logps/chosen": -880.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 34.5, | |
| "rewards/rejected": -39.0, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.7210884353741496, | |
| "eval_logits/chosen": -10.1875, | |
| "eval_logits/rejected": -9.5625, | |
| "eval_logps/chosen": -888.0, | |
| "eval_logps/rejected": -904.0, | |
| "eval_loss": 1.1338906288146973, | |
| "eval_rewards/accuracies": 0.54296875, | |
| "eval_rewards/chosen": -5.6875, | |
| "eval_rewards/margins": 0.46875, | |
| "eval_rewards/rejected": -6.15625, | |
| "eval_runtime": 46.5551, | |
| "eval_samples_per_second": 42.96, | |
| "eval_steps_per_second": 0.687, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.731554160125589, | |
| "grad_norm": 0.0014216720551250281, | |
| "learning_rate": 1.1973654344364925e-06, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -3.546875, | |
| "logps/chosen": -896.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.84375, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -40.25, | |
| "step": 2610 | |
| }, | |
| { | |
| "epoch": 2.7420198848770276, | |
| "grad_norm": 7.361582657585765e-05, | |
| "learning_rate": 1.1059809989586583e-06, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -3.640625, | |
| "logps/chosen": -844.0, | |
| "logps/rejected": -4544.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.34375, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -39.75, | |
| "step": 2620 | |
| }, | |
| { | |
| "epoch": 2.752485609628467, | |
| "grad_norm": 2.389794645315607e-08, | |
| "learning_rate": 1.018144842388044e-06, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -3.140625, | |
| "logps/chosen": -852.0, | |
| "logps/rejected": -4352.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 34.25, | |
| "rewards/rejected": -38.75, | |
| "step": 2630 | |
| }, | |
| { | |
| "epoch": 2.7629513343799057, | |
| "grad_norm": 0.0010702642297902379, | |
| "learning_rate": 9.338700084567109e-07, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -924.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.84375, | |
| "rewards/margins": 34.5, | |
| "rewards/rejected": -39.25, | |
| "step": 2640 | |
| }, | |
| { | |
| "epoch": 2.773417059131345, | |
| "grad_norm": 0.0002894005806639565, | |
| "learning_rate": 8.531690120377606e-07, | |
| "logits/chosen": -11.4375, | |
| "logits/rejected": -3.078125, | |
| "logps/chosen": -912.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -40.0, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 2.7838827838827838, | |
| "grad_norm": 0.000246369141052676, | |
| "learning_rate": 7.760538372868636e-07, | |
| "logits/chosen": -11.4375, | |
| "logits/rejected": -2.796875, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -4768.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.8125, | |
| "rewards/margins": 37.5, | |
| "rewards/rejected": -42.25, | |
| "step": 2660 | |
| }, | |
| { | |
| "epoch": 2.794348508634223, | |
| "grad_norm": 4.562458236534399e-05, | |
| "learning_rate": 7.025359358626166e-07, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -2.734375, | |
| "logps/chosen": -888.0, | |
| "logps/rejected": -4352.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 34.0, | |
| "rewards/rejected": -38.5, | |
| "step": 2670 | |
| }, | |
| { | |
| "epoch": 2.804814233385662, | |
| "grad_norm": 0.00011216890586012522, | |
| "learning_rate": 6.326262252259463e-07, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -3.265625, | |
| "logps/chosen": -892.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 35.75, | |
| "rewards/rejected": -40.5, | |
| "step": 2680 | |
| }, | |
| { | |
| "epoch": 2.815279958137101, | |
| "grad_norm": 0.00015204436046517818, | |
| "learning_rate": 5.66335087018871e-07, | |
| "logits/chosen": -11.0625, | |
| "logits/rejected": -2.921875, | |
| "logps/chosen": -892.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.71875, | |
| "rewards/margins": 34.75, | |
| "rewards/rejected": -39.5, | |
| "step": 2690 | |
| }, | |
| { | |
| "epoch": 2.82574568288854, | |
| "grad_norm": 0.00019067999362415615, | |
| "learning_rate": 5.036723655228226e-07, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -2.828125, | |
| "logps/chosen": -860.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.59375, | |
| "rewards/margins": 36.25, | |
| "rewards/rejected": -40.75, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.82574568288854, | |
| "eval_logits/chosen": -10.1875, | |
| "eval_logits/rejected": -9.5625, | |
| "eval_logps/chosen": -888.0, | |
| "eval_logps/rejected": -904.0, | |
| "eval_loss": 1.1358593702316284, | |
| "eval_rewards/accuracies": 0.546875, | |
| "eval_rewards/chosen": -5.6875, | |
| "eval_rewards/margins": 0.466796875, | |
| "eval_rewards/rejected": -6.15625, | |
| "eval_runtime": 46.5574, | |
| "eval_samples_per_second": 42.958, | |
| "eval_steps_per_second": 0.687, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.836211407639979, | |
| "grad_norm": 3.8117397925683664e-07, | |
| "learning_rate": 4.446473661967432e-07, | |
| "logits/chosen": -11.25, | |
| "logits/rejected": -3.109375, | |
| "logps/chosen": -964.0, | |
| "logps/rejected": -4736.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.96875, | |
| "rewards/margins": 36.5, | |
| "rewards/rejected": -41.5, | |
| "step": 2710 | |
| }, | |
| { | |
| "epoch": 2.846677132391418, | |
| "grad_norm": 0.00028704237530662865, | |
| "learning_rate": 3.892688542952411e-07, | |
| "logits/chosen": -11.5, | |
| "logits/rejected": -3.25, | |
| "logps/chosen": -872.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.46875, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -39.5, | |
| "step": 2720 | |
| }, | |
| { | |
| "epoch": 2.857142857142857, | |
| "grad_norm": 2.4322652759084715e-05, | |
| "learning_rate": 3.3754505356694e-07, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -3.640625, | |
| "logps/chosen": -928.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 35.75, | |
| "rewards/rejected": -40.5, | |
| "step": 2730 | |
| }, | |
| { | |
| "epoch": 2.867608581894296, | |
| "grad_norm": 1.2338905432901236e-05, | |
| "learning_rate": 2.8948364503322276e-07, | |
| "logits/chosen": -11.625, | |
| "logits/rejected": -3.65625, | |
| "logps/chosen": -848.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.25, | |
| "rewards/margins": 36.0, | |
| "rewards/rejected": -40.25, | |
| "step": 2740 | |
| }, | |
| { | |
| "epoch": 2.8780743066457353, | |
| "grad_norm": 0.0020374773115683187, | |
| "learning_rate": 2.4509176584762706e-07, | |
| "logits/chosen": -11.625, | |
| "logits/rejected": -3.6875, | |
| "logps/chosen": -908.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 34.25, | |
| "rewards/rejected": -39.0, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 2.8885400313971745, | |
| "grad_norm": 4.2925947700522475e-05, | |
| "learning_rate": 2.0437600823595694e-07, | |
| "logits/chosen": -11.4375, | |
| "logits/rejected": -3.546875, | |
| "logps/chosen": -920.0, | |
| "logps/rejected": -4512.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -40.0, | |
| "step": 2760 | |
| }, | |
| { | |
| "epoch": 2.8990057561486133, | |
| "grad_norm": 4.290083017522985e-06, | |
| "learning_rate": 1.6734241851733e-07, | |
| "logits/chosen": -11.5625, | |
| "logits/rejected": -3.71875, | |
| "logps/chosen": -948.0, | |
| "logps/rejected": -4640.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.78125, | |
| "rewards/margins": 35.75, | |
| "rewards/rejected": -40.75, | |
| "step": 2770 | |
| }, | |
| { | |
| "epoch": 2.909471480900052, | |
| "grad_norm": 1.4714618475940593e-07, | |
| "learning_rate": 1.339964962062984e-07, | |
| "logits/chosen": -11.9375, | |
| "logits/rejected": -3.65625, | |
| "logps/chosen": -896.0, | |
| "logps/rejected": -4672.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.625, | |
| "rewards/margins": 36.75, | |
| "rewards/rejected": -41.5, | |
| "step": 2780 | |
| }, | |
| { | |
| "epoch": 2.9199372056514914, | |
| "grad_norm": 7.994013518549422e-07, | |
| "learning_rate": 1.0434319319617136e-07, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -3.84375, | |
| "logps/chosen": -864.0, | |
| "logps/rejected": -4320.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.46875, | |
| "rewards/margins": 33.75, | |
| "rewards/rejected": -38.25, | |
| "step": 2790 | |
| }, | |
| { | |
| "epoch": 2.9304029304029307, | |
| "grad_norm": 6.713770886125641e-05, | |
| "learning_rate": 7.838691302365363e-08, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -3.203125, | |
| "logps/chosen": -868.0, | |
| "logps/rejected": -4416.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 34.5, | |
| "rewards/rejected": -39.0, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.9304029304029307, | |
| "eval_logits/chosen": -10.1875, | |
| "eval_logits/rejected": -9.5625, | |
| "eval_logps/chosen": -888.0, | |
| "eval_logps/rejected": -904.0, | |
| "eval_loss": 1.1359844207763672, | |
| "eval_rewards/accuracies": 0.546875, | |
| "eval_rewards/chosen": -5.6875, | |
| "eval_rewards/margins": 0.466796875, | |
| "eval_rewards/rejected": -6.15625, | |
| "eval_runtime": 46.566, | |
| "eval_samples_per_second": 42.95, | |
| "eval_steps_per_second": 0.687, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.9408686551543695, | |
| "grad_norm": 0.0004137595981974209, | |
| "learning_rate": 5.613151021490459e-08, | |
| "logits/chosen": -11.6875, | |
| "logits/rejected": -4.40625, | |
| "logps/chosen": -896.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 34.5, | |
| "rewards/rejected": -39.0, | |
| "step": 2810 | |
| }, | |
| { | |
| "epoch": 2.9513343799058083, | |
| "grad_norm": 0.0005732873569584567, | |
| "learning_rate": 3.7580289713159965e-08, | |
| "logits/chosen": -11.4375, | |
| "logits/rejected": -4.34375, | |
| "logps/chosen": -872.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 34.5, | |
| "rewards/rejected": -39.0, | |
| "step": 2820 | |
| }, | |
| { | |
| "epoch": 2.9618001046572475, | |
| "grad_norm": 4.7311671677056815e-07, | |
| "learning_rate": 2.2736006387927277e-08, | |
| "logits/chosen": -11.75, | |
| "logits/rejected": -3.5, | |
| "logps/chosen": -904.0, | |
| "logps/rejected": -4576.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.90625, | |
| "rewards/margins": 35.5, | |
| "rewards/rejected": -40.5, | |
| "step": 2830 | |
| }, | |
| { | |
| "epoch": 2.9722658294086868, | |
| "grad_norm": 0.00013849900919812822, | |
| "learning_rate": 1.1600864625893604e-08, | |
| "logits/chosen": -11.4375, | |
| "logits/rejected": -4.1875, | |
| "logps/chosen": -928.0, | |
| "logps/rejected": -4608.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.84375, | |
| "rewards/margins": 35.25, | |
| "rewards/rejected": -40.25, | |
| "step": 2840 | |
| }, | |
| { | |
| "epoch": 2.9827315541601256, | |
| "grad_norm": 0.0015963235921299032, | |
| "learning_rate": 4.176518003573549e-09, | |
| "logits/chosen": -11.375, | |
| "logits/rejected": -3.546875, | |
| "logps/chosen": -888.0, | |
| "logps/rejected": -4448.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.4375, | |
| "rewards/margins": 34.5, | |
| "rewards/rejected": -38.75, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 2.9931972789115644, | |
| "grad_norm": 3.7822560334789216e-06, | |
| "learning_rate": 4.640690417528903e-10, | |
| "logits/chosen": -11.125, | |
| "logits/rejected": -2.9375, | |
| "logps/chosen": -900.0, | |
| "logps/rejected": -4480.0, | |
| "loss": 0.0, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 35.0, | |
| "rewards/rejected": -39.5, | |
| "step": 2860 | |
| }, | |
| { | |
| "epoch": 2.998430141287284, | |
| "step": 2865, | |
| "total_flos": 0.0, | |
| "train_loss": 0.01091122463360086, | |
| "train_runtime": 12070.0709, | |
| "train_samples_per_second": 15.195, | |
| "train_steps_per_second": 0.237 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 2865, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |