| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.0, | |
| "eval_steps": 500, | |
| "global_step": 1668, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.001199040767386091, | |
| "grad_norm": 25.642412537629156, | |
| "learning_rate": 1.99203187250996e-09, | |
| "logits/chosen": -2.515625, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -251.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.011990407673860911, | |
| "grad_norm": 24.290562967872415, | |
| "learning_rate": 1.99203187250996e-08, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -217.0, | |
| "logps/rejected": -210.0, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.2361111044883728, | |
| "rewards/chosen": 0.0242919921875, | |
| "rewards/margins": 0.009033203125, | |
| "rewards/rejected": 0.01531982421875, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.023980815347721823, | |
| "grad_norm": 24.03209890621009, | |
| "learning_rate": 3.98406374501992e-08, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -207.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": 0.0224609375, | |
| "rewards/margins": 0.015625, | |
| "rewards/rejected": 0.006866455078125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03597122302158273, | |
| "grad_norm": 24.100678886411075, | |
| "learning_rate": 5.97609561752988e-08, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -200.0, | |
| "loss": 0.697, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": 0.0081787109375, | |
| "rewards/margins": -0.0087890625, | |
| "rewards/rejected": 0.016845703125, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.047961630695443645, | |
| "grad_norm": 22.19662351600711, | |
| "learning_rate": 7.96812749003984e-08, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -205.0, | |
| "logps/rejected": -192.0, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.2750000059604645, | |
| "rewards/chosen": 0.0081787109375, | |
| "rewards/margins": 0.0018768310546875, | |
| "rewards/rejected": 0.006256103515625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.05995203836930456, | |
| "grad_norm": 22.91225992894245, | |
| "learning_rate": 9.9601593625498e-08, | |
| "logits/chosen": -2.4375, | |
| "logits/rejected": -2.453125, | |
| "logps/chosen": -217.0, | |
| "logps/rejected": -213.0, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": 0.0238037109375, | |
| "rewards/margins": -0.0031280517578125, | |
| "rewards/rejected": 0.02685546875, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.07194244604316546, | |
| "grad_norm": 22.915913925151774, | |
| "learning_rate": 1.195219123505976e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -202.0, | |
| "logps/rejected": -211.0, | |
| "loss": 0.6965, | |
| "rewards/accuracies": 0.16249999403953552, | |
| "rewards/chosen": 0.0037384033203125, | |
| "rewards/margins": -0.0037384033203125, | |
| "rewards/rejected": 0.00750732421875, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.08393285371702638, | |
| "grad_norm": 24.635308516543407, | |
| "learning_rate": 1.394422310756972e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -204.0, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.01129150390625, | |
| "rewards/margins": -0.0118408203125, | |
| "rewards/rejected": 0.0230712890625, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.09592326139088729, | |
| "grad_norm": 23.422749201250696, | |
| "learning_rate": 1.593625498007968e-07, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.453125, | |
| "logps/chosen": -211.0, | |
| "logps/rejected": -224.0, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.01251220703125, | |
| "rewards/margins": -0.006256103515625, | |
| "rewards/rejected": 0.018798828125, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.1079136690647482, | |
| "grad_norm": 25.550050248970773, | |
| "learning_rate": 1.7928286852589642e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -223.0, | |
| "loss": 0.6919, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": 0.0281982421875, | |
| "rewards/margins": 0.001251220703125, | |
| "rewards/rejected": 0.0269775390625, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.11990407673860912, | |
| "grad_norm": 25.761780120884907, | |
| "learning_rate": 1.99203187250996e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.453125, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -208.0, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": 0.010009765625, | |
| "rewards/margins": -0.00250244140625, | |
| "rewards/rejected": 0.0125732421875, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.13189448441247004, | |
| "grad_norm": 22.84334462455036, | |
| "learning_rate": 2.191235059760956e-07, | |
| "logits/chosen": -2.453125, | |
| "logits/rejected": -2.46875, | |
| "logps/chosen": -202.0, | |
| "logps/rejected": -218.0, | |
| "loss": 0.687, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": 0.021240234375, | |
| "rewards/margins": 0.0050048828125, | |
| "rewards/rejected": 0.016357421875, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.14388489208633093, | |
| "grad_norm": 24.300426924539867, | |
| "learning_rate": 2.390438247011952e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -207.0, | |
| "logps/rejected": -189.0, | |
| "loss": 0.6896, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": 0.006256103515625, | |
| "rewards/margins": 0.0106201171875, | |
| "rewards/rejected": -0.004425048828125, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.15587529976019185, | |
| "grad_norm": 25.941365174437085, | |
| "learning_rate": 2.589641434262948e-07, | |
| "logits/chosen": -2.40625, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -212.0, | |
| "logps/rejected": -204.0, | |
| "loss": 0.6807, | |
| "rewards/accuracies": 0.38749998807907104, | |
| "rewards/chosen": 0.036376953125, | |
| "rewards/margins": 0.0194091796875, | |
| "rewards/rejected": 0.016845703125, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.16786570743405277, | |
| "grad_norm": 22.46827042788133, | |
| "learning_rate": 2.788844621513944e-07, | |
| "logits/chosen": -2.40625, | |
| "logits/rejected": -2.46875, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -208.0, | |
| "loss": 0.6813, | |
| "rewards/accuracies": 0.375, | |
| "rewards/chosen": 0.0281982421875, | |
| "rewards/margins": 0.016845703125, | |
| "rewards/rejected": 0.01129150390625, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.17985611510791366, | |
| "grad_norm": 22.828346012492343, | |
| "learning_rate": 2.98804780876494e-07, | |
| "logits/chosen": -2.40625, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -217.0, | |
| "loss": 0.6788, | |
| "rewards/accuracies": 0.4000000059604645, | |
| "rewards/chosen": 0.03759765625, | |
| "rewards/margins": 0.0238037109375, | |
| "rewards/rejected": 0.01373291015625, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.19184652278177458, | |
| "grad_norm": 25.091544876831573, | |
| "learning_rate": 3.187250996015936e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -215.0, | |
| "loss": 0.6829, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.056396484375, | |
| "rewards/margins": 0.0400390625, | |
| "rewards/rejected": 0.0162353515625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.2038369304556355, | |
| "grad_norm": 22.476144325987256, | |
| "learning_rate": 3.386454183266932e-07, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -222.0, | |
| "logps/rejected": -220.0, | |
| "loss": 0.6707, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.0576171875, | |
| "rewards/margins": 0.07275390625, | |
| "rewards/rejected": -0.0150146484375, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.2158273381294964, | |
| "grad_norm": 22.738418341833743, | |
| "learning_rate": 3.5856573705179284e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -211.0, | |
| "logps/rejected": -214.0, | |
| "loss": 0.6701, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": 0.06640625, | |
| "rewards/margins": 0.05322265625, | |
| "rewards/rejected": 0.01318359375, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.2278177458033573, | |
| "grad_norm": 22.162554634157406, | |
| "learning_rate": 3.784860557768924e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -200.0, | |
| "logps/rejected": -188.0, | |
| "loss": 0.6715, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.08203125, | |
| "rewards/margins": 0.0654296875, | |
| "rewards/rejected": 0.0162353515625, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.23980815347721823, | |
| "grad_norm": 24.649798845932942, | |
| "learning_rate": 3.98406374501992e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -212.0, | |
| "logps/rejected": -207.0, | |
| "loss": 0.6712, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.033203125, | |
| "rewards/margins": 0.050048828125, | |
| "rewards/rejected": -0.016845703125, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2517985611510791, | |
| "grad_norm": 19.334172790085294, | |
| "learning_rate": 4.1832669322709163e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.453125, | |
| "logps/chosen": -205.0, | |
| "logps/rejected": -213.0, | |
| "loss": 0.6612, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.048828125, | |
| "rewards/margins": 0.07275390625, | |
| "rewards/rejected": -0.0238037109375, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.2637889688249401, | |
| "grad_norm": 20.773278941923554, | |
| "learning_rate": 4.382470119521912e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -217.0, | |
| "logps/rejected": -210.0, | |
| "loss": 0.6522, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.0162353515625, | |
| "rewards/margins": 0.08251953125, | |
| "rewards/rejected": -0.0986328125, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.27577937649880097, | |
| "grad_norm": 23.13141722026618, | |
| "learning_rate": 4.581673306772908e-07, | |
| "logits/chosen": -2.40625, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -207.0, | |
| "logps/rejected": -210.0, | |
| "loss": 0.641, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.050048828125, | |
| "rewards/margins": 0.1259765625, | |
| "rewards/rejected": -0.1767578125, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.28776978417266186, | |
| "grad_norm": 25.417525122060646, | |
| "learning_rate": 4.780876494023904e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -205.0, | |
| "loss": 0.6402, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.06201171875, | |
| "rewards/margins": 0.154296875, | |
| "rewards/rejected": -0.2158203125, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.2997601918465228, | |
| "grad_norm": 22.97557618510118, | |
| "learning_rate": 4.9800796812749e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -205.0, | |
| "loss": 0.6303, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.251953125, | |
| "rewards/margins": 0.1259765625, | |
| "rewards/rejected": -0.37890625, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.3117505995203837, | |
| "grad_norm": 22.380902540882072, | |
| "learning_rate": 4.980008884940026e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -206.0, | |
| "loss": 0.6271, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.203125, | |
| "rewards/margins": 0.224609375, | |
| "rewards/rejected": -0.427734375, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.3237410071942446, | |
| "grad_norm": 20.410713037919127, | |
| "learning_rate": 4.957796534873389e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -211.0, | |
| "logps/rejected": -206.0, | |
| "loss": 0.6168, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.462890625, | |
| "rewards/margins": 0.234375, | |
| "rewards/rejected": -0.69921875, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.33573141486810554, | |
| "grad_norm": 19.858477875493026, | |
| "learning_rate": 4.935584184806753e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -213.0, | |
| "logps/rejected": -216.0, | |
| "loss": 0.5864, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.8046875, | |
| "rewards/margins": 0.310546875, | |
| "rewards/rejected": -1.109375, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.34772182254196643, | |
| "grad_norm": 25.75186279410947, | |
| "learning_rate": 4.913371834740116e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -213.0, | |
| "logps/rejected": -219.0, | |
| "loss": 0.6046, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -1.015625, | |
| "rewards/margins": 0.3046875, | |
| "rewards/rejected": -1.328125, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.3597122302158273, | |
| "grad_norm": 20.582004189162635, | |
| "learning_rate": 4.891159484673478e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -223.0, | |
| "loss": 0.6033, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.66796875, | |
| "rewards/margins": 0.26171875, | |
| "rewards/rejected": -0.9296875, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.37170263788968827, | |
| "grad_norm": 17.87508243629168, | |
| "learning_rate": 4.868947134606841e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -233.0, | |
| "loss": 0.5365, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.87890625, | |
| "rewards/margins": 0.6171875, | |
| "rewards/rejected": -1.4921875, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.38369304556354916, | |
| "grad_norm": 22.10087172308854, | |
| "learning_rate": 4.846734784540204e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -231.0, | |
| "logps/rejected": -238.0, | |
| "loss": 0.5703, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.8125, | |
| "rewards/margins": 0.43359375, | |
| "rewards/rejected": -1.25, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.39568345323741005, | |
| "grad_norm": 27.704111990126563, | |
| "learning_rate": 4.824522434473567e-07, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -235.0, | |
| "loss": 0.543, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.15625, | |
| "rewards/margins": 0.490234375, | |
| "rewards/rejected": -1.6484375, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.407673860911271, | |
| "grad_norm": 19.609485654347775, | |
| "learning_rate": 4.80231008440693e-07, | |
| "logits/chosen": -2.40625, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -223.0, | |
| "logps/rejected": -228.0, | |
| "loss": 0.5288, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.8046875, | |
| "rewards/margins": 0.61328125, | |
| "rewards/rejected": -1.4140625, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.4196642685851319, | |
| "grad_norm": 25.341062357182967, | |
| "learning_rate": 4.780097734340293e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.5653, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.8515625, | |
| "rewards/margins": 0.546875, | |
| "rewards/rejected": -1.3984375, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.4316546762589928, | |
| "grad_norm": 24.609756364283665, | |
| "learning_rate": 4.757885384273656e-07, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -213.0, | |
| "logps/rejected": -226.0, | |
| "loss": 0.5599, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.70703125, | |
| "rewards/margins": 0.453125, | |
| "rewards/rejected": -1.15625, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.44364508393285373, | |
| "grad_norm": 20.448105072112956, | |
| "learning_rate": 4.7356730342070187e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -226.0, | |
| "logps/rejected": -231.0, | |
| "loss": 0.5373, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.171875, | |
| "rewards/margins": 0.58984375, | |
| "rewards/rejected": -1.7578125, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.4556354916067146, | |
| "grad_norm": 20.00137526211279, | |
| "learning_rate": 4.713460684140382e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -225.0, | |
| "logps/rejected": -228.0, | |
| "loss": 0.5482, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.4609375, | |
| "rewards/margins": 0.474609375, | |
| "rewards/rejected": -1.9375, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.4676258992805755, | |
| "grad_norm": 20.08046772545977, | |
| "learning_rate": 4.691248334073745e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -222.0, | |
| "logps/rejected": -232.0, | |
| "loss": 0.5292, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.4453125, | |
| "rewards/margins": 0.4609375, | |
| "rewards/rejected": -1.90625, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.47961630695443647, | |
| "grad_norm": 24.656493214054557, | |
| "learning_rate": 4.6690359840071075e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -203.0, | |
| "logps/rejected": -220.0, | |
| "loss": 0.4912, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.72265625, | |
| "rewards/margins": 0.69140625, | |
| "rewards/rejected": -1.4140625, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.49160671462829736, | |
| "grad_norm": 20.967973696578976, | |
| "learning_rate": 4.646823633940471e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.453125, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.5305, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.74609375, | |
| "rewards/margins": 0.58203125, | |
| "rewards/rejected": -1.328125, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.5035971223021583, | |
| "grad_norm": 25.657618682397636, | |
| "learning_rate": 4.6246112838738336e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -217.0, | |
| "logps/rejected": -217.0, | |
| "loss": 0.5173, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.7109375, | |
| "rewards/margins": 0.625, | |
| "rewards/rejected": -2.328125, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.5155875299760192, | |
| "grad_norm": 22.796145877614705, | |
| "learning_rate": 4.602398933807197e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -221.0, | |
| "logps/rejected": -254.0, | |
| "loss": 0.5249, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.1015625, | |
| "rewards/margins": 0.419921875, | |
| "rewards/rejected": -1.5234375, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.5275779376498801, | |
| "grad_norm": 21.830523584389546, | |
| "learning_rate": 4.5801865837405597e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -231.0, | |
| "logps/rejected": -234.0, | |
| "loss": 0.4868, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.97265625, | |
| "rewards/margins": 0.75, | |
| "rewards/rejected": -1.7265625, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.539568345323741, | |
| "grad_norm": 22.863177511462844, | |
| "learning_rate": 4.5579742336739224e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -217.0, | |
| "logps/rejected": -206.0, | |
| "loss": 0.5349, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.51953125, | |
| "rewards/margins": 0.79296875, | |
| "rewards/rejected": -1.3125, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.5515587529976019, | |
| "grad_norm": 22.106098891434247, | |
| "learning_rate": 4.535761883607285e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.4969, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.0078125, | |
| "rewards/margins": 0.5234375, | |
| "rewards/rejected": -1.53125, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.5635491606714629, | |
| "grad_norm": 21.83764765940895, | |
| "learning_rate": 4.513549533540648e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -221.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.4531, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.9609375, | |
| "rewards/margins": 0.75390625, | |
| "rewards/rejected": -1.7109375, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5755395683453237, | |
| "grad_norm": 27.75012920313157, | |
| "learning_rate": 4.491337183474012e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -227.0, | |
| "logps/rejected": -221.0, | |
| "loss": 0.5074, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.234375, | |
| "rewards/margins": 0.625, | |
| "rewards/rejected": -1.8515625, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.5875299760191847, | |
| "grad_norm": 17.36030484793641, | |
| "learning_rate": 4.4691248334073746e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -225.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.48, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.375, | |
| "rewards/margins": 0.73828125, | |
| "rewards/rejected": -2.109375, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.5995203836930456, | |
| "grad_norm": 22.808492833261823, | |
| "learning_rate": 4.4469124833407373e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -214.0, | |
| "loss": 0.5236, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.3984375, | |
| "rewards/margins": 0.443359375, | |
| "rewards/rejected": -1.84375, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.6115107913669064, | |
| "grad_norm": 23.84926176790092, | |
| "learning_rate": 4.4247001332741e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -215.0, | |
| "logps/rejected": -235.0, | |
| "loss": 0.4798, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.96484375, | |
| "rewards/margins": 0.76953125, | |
| "rewards/rejected": -1.734375, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.6235011990407674, | |
| "grad_norm": 21.734949611768823, | |
| "learning_rate": 4.402487783207463e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -215.0, | |
| "logps/rejected": -238.0, | |
| "loss": 0.437, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.0546875, | |
| "rewards/margins": 0.9921875, | |
| "rewards/rejected": -2.046875, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.6354916067146283, | |
| "grad_norm": 17.985205119836653, | |
| "learning_rate": 4.380275433140826e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.453125, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.4438, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.0703125, | |
| "rewards/margins": 1.0625, | |
| "rewards/rejected": -2.125, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.6474820143884892, | |
| "grad_norm": 19.26308613682953, | |
| "learning_rate": 4.358063083074189e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -244.0, | |
| "loss": 0.4505, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -1.5, | |
| "rewards/margins": 1.234375, | |
| "rewards/rejected": -2.734375, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.6594724220623501, | |
| "grad_norm": 24.596146010502707, | |
| "learning_rate": 4.335850733007552e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -221.0, | |
| "logps/rejected": -233.0, | |
| "loss": 0.4943, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.6796875, | |
| "rewards/margins": 0.7421875, | |
| "rewards/rejected": -2.421875, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.6714628297362111, | |
| "grad_norm": 15.013286046936686, | |
| "learning_rate": 4.313638382940915e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -217.0, | |
| "logps/rejected": -236.0, | |
| "loss": 0.4661, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.9921875, | |
| "rewards/margins": 0.984375, | |
| "rewards/rejected": -1.9765625, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.6834532374100719, | |
| "grad_norm": 19.571343970012546, | |
| "learning_rate": 4.291426032874278e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -218.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.43, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.2265625, | |
| "rewards/margins": 1.03125, | |
| "rewards/rejected": -2.25, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.6954436450839329, | |
| "grad_norm": 20.377379214041834, | |
| "learning_rate": 4.269213682807641e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -227.0, | |
| "logps/rejected": -221.0, | |
| "loss": 0.4403, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.83984375, | |
| "rewards/margins": 0.9375, | |
| "rewards/rejected": -1.7734375, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.7074340527577938, | |
| "grad_norm": 25.88308379473588, | |
| "learning_rate": 4.247001332741004e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.4357, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.78515625, | |
| "rewards/margins": 1.0703125, | |
| "rewards/rejected": -1.8515625, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.7194244604316546, | |
| "grad_norm": 21.223402057931814, | |
| "learning_rate": 4.2247889826743666e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -227.0, | |
| "logps/rejected": -232.0, | |
| "loss": 0.4865, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.8515625, | |
| "rewards/margins": 0.8125, | |
| "rewards/rejected": -1.6640625, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.7314148681055156, | |
| "grad_norm": 26.413684307486818, | |
| "learning_rate": 4.2025766326077294e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -222.0, | |
| "logps/rejected": -231.0, | |
| "loss": 0.4473, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.453125, | |
| "rewards/margins": 0.984375, | |
| "rewards/rejected": -2.4375, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.7434052757793765, | |
| "grad_norm": 19.274016165683612, | |
| "learning_rate": 4.1803642825410926e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -218.0, | |
| "logps/rejected": -241.0, | |
| "loss": 0.4574, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.2734375, | |
| "rewards/margins": 0.94140625, | |
| "rewards/rejected": -2.21875, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.7553956834532374, | |
| "grad_norm": 21.962129244766075, | |
| "learning_rate": 4.158151932474456e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -234.0, | |
| "loss": 0.4202, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.546875, | |
| "rewards/margins": 1.09375, | |
| "rewards/rejected": -2.640625, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.7673860911270983, | |
| "grad_norm": 19.250420656857816, | |
| "learning_rate": 4.1359395824078187e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -231.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.423, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -2.109375, | |
| "rewards/margins": 1.1640625, | |
| "rewards/rejected": -3.28125, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.7793764988009593, | |
| "grad_norm": 17.26038839326138, | |
| "learning_rate": 4.1137272323411815e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -223.0, | |
| "logps/rejected": -245.0, | |
| "loss": 0.4144, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.140625, | |
| "rewards/margins": 1.140625, | |
| "rewards/rejected": -3.28125, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.7913669064748201, | |
| "grad_norm": 22.725812386482403, | |
| "learning_rate": 4.091514882274544e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -224.0, | |
| "logps/rejected": -223.0, | |
| "loss": 0.4619, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.6328125, | |
| "rewards/margins": 0.75, | |
| "rewards/rejected": -2.390625, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.8033573141486811, | |
| "grad_norm": 26.346880341026353, | |
| "learning_rate": 4.069302532207907e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -234.0, | |
| "logps/rejected": -233.0, | |
| "loss": 0.4198, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.09375, | |
| "rewards/margins": 1.2109375, | |
| "rewards/rejected": -2.296875, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.815347721822542, | |
| "grad_norm": 16.341325440490397, | |
| "learning_rate": 4.047090182141271e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -237.0, | |
| "loss": 0.4215, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.828125, | |
| "rewards/margins": 0.8359375, | |
| "rewards/rejected": -2.65625, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.8273381294964028, | |
| "grad_norm": 19.597918931735652, | |
| "learning_rate": 4.0248778320746336e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.421875, | |
| "logps/chosen": -225.0, | |
| "logps/rejected": -246.0, | |
| "loss": 0.4179, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.6640625, | |
| "rewards/margins": 1.046875, | |
| "rewards/rejected": -2.71875, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.8393285371702638, | |
| "grad_norm": 16.19010210226658, | |
| "learning_rate": 4.0026654820079964e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -253.0, | |
| "loss": 0.4064, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -1.6875, | |
| "rewards/margins": 1.1875, | |
| "rewards/rejected": -2.875, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.8513189448441247, | |
| "grad_norm": 19.378154801231126, | |
| "learning_rate": 3.980453131941359e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -243.0, | |
| "loss": 0.382, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.703125, | |
| "rewards/margins": 1.234375, | |
| "rewards/rejected": -2.9375, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.8633093525179856, | |
| "grad_norm": 17.875697953531226, | |
| "learning_rate": 3.958240781874722e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -231.0, | |
| "loss": 0.3839, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -1.6171875, | |
| "rewards/margins": 1.28125, | |
| "rewards/rejected": -2.90625, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.8752997601918465, | |
| "grad_norm": 21.20250450740525, | |
| "learning_rate": 3.936028431808085e-07, | |
| "logits/chosen": -2.390625, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -231.0, | |
| "logps/rejected": -231.0, | |
| "loss": 0.3574, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.609375, | |
| "rewards/margins": 1.421875, | |
| "rewards/rejected": -3.03125, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.8872901678657075, | |
| "grad_norm": 17.46000299554747, | |
| "learning_rate": 3.913816081741448e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -237.0, | |
| "loss": 0.3473, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -1.703125, | |
| "rewards/margins": 1.625, | |
| "rewards/rejected": -3.328125, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.8992805755395683, | |
| "grad_norm": 17.53738029583566, | |
| "learning_rate": 3.8916037316748113e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -222.0, | |
| "logps/rejected": -232.0, | |
| "loss": 0.371, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -2.09375, | |
| "rewards/margins": 1.2421875, | |
| "rewards/rejected": -3.328125, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.9112709832134293, | |
| "grad_norm": 13.845287693810857, | |
| "learning_rate": 3.869391381608174e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -227.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.3346, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.359375, | |
| "rewards/margins": 1.28125, | |
| "rewards/rejected": -2.640625, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.9232613908872902, | |
| "grad_norm": 18.53800658699909, | |
| "learning_rate": 3.847179031541537e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -225.0, | |
| "logps/rejected": -227.0, | |
| "loss": 0.3642, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -1.671875, | |
| "rewards/margins": 1.3125, | |
| "rewards/rejected": -2.984375, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.935251798561151, | |
| "grad_norm": 21.872485752073064, | |
| "learning_rate": 3.8249666814749e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -240.0, | |
| "logps/rejected": -237.0, | |
| "loss": 0.3391, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -2.015625, | |
| "rewards/margins": 1.5078125, | |
| "rewards/rejected": -3.53125, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.947242206235012, | |
| "grad_norm": 21.111212041972365, | |
| "learning_rate": 3.802754331408263e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -232.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.3679, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.296875, | |
| "rewards/margins": 1.3203125, | |
| "rewards/rejected": -3.625, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.9592326139088729, | |
| "grad_norm": 21.61033072214478, | |
| "learning_rate": 3.7805419813416256e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -226.0, | |
| "logps/rejected": -241.0, | |
| "loss": 0.3427, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -2.5625, | |
| "rewards/margins": 1.4140625, | |
| "rewards/rejected": -3.96875, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.9712230215827338, | |
| "grad_norm": 15.32449671679388, | |
| "learning_rate": 3.7583296312749884e-07, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -234.0, | |
| "logps/rejected": -252.0, | |
| "loss": 0.3787, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -2.71875, | |
| "rewards/margins": 1.53125, | |
| "rewards/rejected": -4.25, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.9832134292565947, | |
| "grad_norm": 19.67405128244237, | |
| "learning_rate": 3.7361172812083517e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -232.0, | |
| "loss": 0.3254, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -1.9453125, | |
| "rewards/margins": 1.46875, | |
| "rewards/rejected": -3.421875, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.9952038369304557, | |
| "grad_norm": 21.005105323056593, | |
| "learning_rate": 3.713904931141715e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -237.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.2971, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -1.5234375, | |
| "rewards/margins": 1.6328125, | |
| "rewards/rejected": -3.15625, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -2.296875, | |
| "eval_logits/rejected": -2.328125, | |
| "eval_logps/chosen": -228.0, | |
| "eval_logps/rejected": -236.0, | |
| "eval_loss": 0.6048940420150757, | |
| "eval_rewards/accuracies": 0.6617646813392639, | |
| "eval_rewards/chosen": -1.9921875, | |
| "eval_rewards/margins": 0.93359375, | |
| "eval_rewards/rejected": -2.921875, | |
| "eval_runtime": 19.9965, | |
| "eval_samples_per_second": 20.053, | |
| "eval_steps_per_second": 0.85, | |
| "step": 834 | |
| }, | |
| { | |
| "epoch": 1.0071942446043165, | |
| "grad_norm": 18.731496491285114, | |
| "learning_rate": 3.691692581075078e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -237.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.2766, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -1.9453125, | |
| "rewards/margins": 1.984375, | |
| "rewards/rejected": -3.921875, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 1.0191846522781776, | |
| "grad_norm": 18.062862619034647, | |
| "learning_rate": 3.6694802310084405e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -226.0, | |
| "logps/rejected": -246.0, | |
| "loss": 0.2457, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.625, | |
| "rewards/margins": 1.8046875, | |
| "rewards/rejected": -3.4375, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 1.0311750599520384, | |
| "grad_norm": 18.43054843437252, | |
| "learning_rate": 3.6472678809418033e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -234.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.2445, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -1.7734375, | |
| "rewards/margins": 1.984375, | |
| "rewards/rejected": -3.765625, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 1.0431654676258992, | |
| "grad_norm": 14.792270635313406, | |
| "learning_rate": 3.625055530875166e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -239.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.2187, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.046875, | |
| "rewards/margins": 2.15625, | |
| "rewards/rejected": -4.1875, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 1.0551558752997603, | |
| "grad_norm": 14.450058470084105, | |
| "learning_rate": 3.6028431808085294e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -238.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.2253, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -1.984375, | |
| "rewards/margins": 1.953125, | |
| "rewards/rejected": -3.953125, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 1.0671462829736211, | |
| "grad_norm": 14.392869335197297, | |
| "learning_rate": 3.5806308307418926e-07, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.1993, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.15625, | |
| "rewards/margins": 2.25, | |
| "rewards/rejected": -4.40625, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 1.079136690647482, | |
| "grad_norm": 11.360043739301632, | |
| "learning_rate": 3.5584184806752554e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -224.0, | |
| "logps/rejected": -252.0, | |
| "loss": 0.2037, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.8984375, | |
| "rewards/margins": 2.390625, | |
| "rewards/rejected": -4.28125, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.091127098321343, | |
| "grad_norm": 17.249700366406504, | |
| "learning_rate": 3.536206130608618e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -224.0, | |
| "logps/rejected": -244.0, | |
| "loss": 0.1878, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.921875, | |
| "rewards/margins": 2.34375, | |
| "rewards/rejected": -4.28125, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 1.1031175059952039, | |
| "grad_norm": 17.500202516480442, | |
| "learning_rate": 3.513993780541981e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.375, | |
| "logps/chosen": -203.0, | |
| "logps/rejected": -246.0, | |
| "loss": 0.2128, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.875, | |
| "rewards/margins": 2.375, | |
| "rewards/rejected": -4.25, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 1.1151079136690647, | |
| "grad_norm": 15.16644938505708, | |
| "learning_rate": 3.491781430475344e-07, | |
| "logits/chosen": -2.359375, | |
| "logits/rejected": -2.40625, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.1957, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.4375, | |
| "rewards/margins": 2.40625, | |
| "rewards/rejected": -3.84375, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 1.1270983213429258, | |
| "grad_norm": 12.741517880568063, | |
| "learning_rate": 3.469569080408707e-07, | |
| "logits/chosen": -2.21875, | |
| "logits/rejected": -2.234375, | |
| "logps/chosen": -225.0, | |
| "logps/rejected": -246.0, | |
| "loss": 0.2034, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.296875, | |
| "rewards/margins": 2.359375, | |
| "rewards/rejected": -4.65625, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 1.1390887290167866, | |
| "grad_norm": 21.470206899673656, | |
| "learning_rate": 3.4473567303420703e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.1846, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.34375, | |
| "rewards/margins": 2.515625, | |
| "rewards/rejected": -4.875, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.1510791366906474, | |
| "grad_norm": 17.20708415573527, | |
| "learning_rate": 3.425144380275433e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.359375, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.2031, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.3125, | |
| "rewards/margins": 2.5625, | |
| "rewards/rejected": -4.875, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 1.1630695443645085, | |
| "grad_norm": 14.942195535135212, | |
| "learning_rate": 3.402932030208796e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.2032, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.234375, | |
| "rewards/margins": 2.734375, | |
| "rewards/rejected": -4.96875, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 1.1750599520383693, | |
| "grad_norm": 11.998202232926158, | |
| "learning_rate": 3.380719680142159e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -226.0, | |
| "logps/rejected": -248.0, | |
| "loss": 0.1639, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.734375, | |
| "rewards/margins": 2.6875, | |
| "rewards/rejected": -4.40625, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 1.1870503597122302, | |
| "grad_norm": 12.340216646822865, | |
| "learning_rate": 3.358507330075522e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -238.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.1902, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -2.75, | |
| "rewards/margins": 2.296875, | |
| "rewards/rejected": -5.0625, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 1.1990407673860912, | |
| "grad_norm": 10.240701799756478, | |
| "learning_rate": 3.3362949800088847e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -237.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.1614, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.859375, | |
| "rewards/margins": 2.65625, | |
| "rewards/rejected": -4.5, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.211031175059952, | |
| "grad_norm": 16.753004063840088, | |
| "learning_rate": 3.3140826299422474e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.390625, | |
| "logps/chosen": -234.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.2139, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.15625, | |
| "rewards/margins": 2.5625, | |
| "rewards/rejected": -4.71875, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 1.223021582733813, | |
| "grad_norm": 22.84859803236856, | |
| "learning_rate": 3.291870279875611e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.25, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.1757, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.359375, | |
| "rewards/margins": 2.578125, | |
| "rewards/rejected": -4.9375, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 1.235011990407674, | |
| "grad_norm": 14.789129276628799, | |
| "learning_rate": 3.269657929808974e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -234.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.1719, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.078125, | |
| "rewards/margins": 2.515625, | |
| "rewards/rejected": -4.59375, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 1.2470023980815348, | |
| "grad_norm": 21.148440474411384, | |
| "learning_rate": 3.247445579742337e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -228.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.1689, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.6875, | |
| "rewards/margins": 2.546875, | |
| "rewards/rejected": -5.25, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 1.2589928057553956, | |
| "grad_norm": 19.272034515617282, | |
| "learning_rate": 3.2252332296756996e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -227.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.1591, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.65625, | |
| "rewards/margins": 2.96875, | |
| "rewards/rejected": -5.625, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.2709832134292567, | |
| "grad_norm": 16.85198782317795, | |
| "learning_rate": 3.2030208796090623e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -229.0, | |
| "logps/rejected": -251.0, | |
| "loss": 0.2236, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -1.765625, | |
| "rewards/margins": 2.890625, | |
| "rewards/rejected": -4.65625, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 1.2829736211031175, | |
| "grad_norm": 15.836692596712956, | |
| "learning_rate": 3.180808529542425e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.168, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.671875, | |
| "rewards/margins": 2.703125, | |
| "rewards/rejected": -5.375, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 1.2949640287769784, | |
| "grad_norm": 8.91604867948452, | |
| "learning_rate": 3.1585961794757884e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -240.0, | |
| "logps/rejected": -253.0, | |
| "loss": 0.1711, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.59375, | |
| "rewards/margins": 2.75, | |
| "rewards/rejected": -4.34375, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 1.3069544364508392, | |
| "grad_norm": 14.305189630485472, | |
| "learning_rate": 3.1363838294091517e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -251.0, | |
| "loss": 0.1415, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.640625, | |
| "rewards/margins": 2.9375, | |
| "rewards/rejected": -4.59375, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 1.3189448441247003, | |
| "grad_norm": 12.352735099035222, | |
| "learning_rate": 3.1141714793425145e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -249.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.1572, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.1875, | |
| "rewards/margins": 2.953125, | |
| "rewards/rejected": -5.125, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.330935251798561, | |
| "grad_norm": 14.924171575427302, | |
| "learning_rate": 3.091959129275877e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -215.0, | |
| "logps/rejected": -241.0, | |
| "loss": 0.1714, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -1.9375, | |
| "rewards/margins": 2.84375, | |
| "rewards/rejected": -4.78125, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 1.3429256594724222, | |
| "grad_norm": 17.499923046869892, | |
| "learning_rate": 3.06974677920924e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -254.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.1393, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.953125, | |
| "rewards/margins": 2.640625, | |
| "rewards/rejected": -5.59375, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 1.354916067146283, | |
| "grad_norm": 16.158253954777205, | |
| "learning_rate": 3.0475344291426033e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -229.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.1656, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.1875, | |
| "rewards/margins": 3.375, | |
| "rewards/rejected": -5.5625, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 1.3669064748201438, | |
| "grad_norm": 12.087846600129158, | |
| "learning_rate": 3.025322079075966e-07, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -225.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.1966, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.96875, | |
| "rewards/margins": 3.171875, | |
| "rewards/rejected": -5.15625, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 1.3788968824940047, | |
| "grad_norm": 17.576347022242885, | |
| "learning_rate": 3.003109729009329e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.1847, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.8046875, | |
| "rewards/margins": 2.703125, | |
| "rewards/rejected": -4.5, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.3908872901678657, | |
| "grad_norm": 12.233329087327826, | |
| "learning_rate": 2.980897378942692e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.234375, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.1373, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.34375, | |
| "rewards/margins": 2.890625, | |
| "rewards/rejected": -5.21875, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 1.4028776978417266, | |
| "grad_norm": 14.731501359525842, | |
| "learning_rate": 2.958685028876055e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.234375, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.1843, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.1875, | |
| "rewards/margins": 2.625, | |
| "rewards/rejected": -4.8125, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 1.4148681055155876, | |
| "grad_norm": 10.33569347198834, | |
| "learning_rate": 2.936472678809418e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.1912, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.40625, | |
| "rewards/margins": 2.640625, | |
| "rewards/rejected": -5.0625, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 1.4268585131894485, | |
| "grad_norm": 16.986111059013577, | |
| "learning_rate": 2.914260328742781e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.25, | |
| "logps/chosen": -231.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.1611, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.3125, | |
| "rewards/margins": 2.703125, | |
| "rewards/rejected": -5.03125, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 1.4388489208633093, | |
| "grad_norm": 25.435636014827384, | |
| "learning_rate": 2.8920479786761437e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.1512, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.28125, | |
| "rewards/margins": 2.859375, | |
| "rewards/rejected": -5.15625, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.4508393285371701, | |
| "grad_norm": 13.56680412334536, | |
| "learning_rate": 2.8698356286095065e-07, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -260.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.1255, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.171875, | |
| "rewards/margins": 3.40625, | |
| "rewards/rejected": -6.59375, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 1.4628297362110312, | |
| "grad_norm": 8.141641110335339, | |
| "learning_rate": 2.847623278542869e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.1428, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.296875, | |
| "rewards/margins": 3.03125, | |
| "rewards/rejected": -5.34375, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 1.474820143884892, | |
| "grad_norm": 21.115162166818017, | |
| "learning_rate": 2.825410928476233e-07, | |
| "logits/chosen": -2.1875, | |
| "logits/rejected": -2.25, | |
| "logps/chosen": -221.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.136, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.796875, | |
| "rewards/margins": 3.09375, | |
| "rewards/rejected": -5.90625, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 1.486810551558753, | |
| "grad_norm": 18.47262881897785, | |
| "learning_rate": 2.803198578409596e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.25, | |
| "logps/chosen": -228.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.1688, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -1.9375, | |
| "rewards/margins": 3.171875, | |
| "rewards/rejected": -5.125, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 1.498800959232614, | |
| "grad_norm": 7.863109211524176, | |
| "learning_rate": 2.7809862283429586e-07, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -222.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.1298, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.046875, | |
| "rewards/margins": 3.53125, | |
| "rewards/rejected": -5.5625, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.5107913669064748, | |
| "grad_norm": 12.978026445303877, | |
| "learning_rate": 2.7587738782763214e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.1403, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.40625, | |
| "rewards/margins": 3.421875, | |
| "rewards/rejected": -5.8125, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 1.5227817745803356, | |
| "grad_norm": 15.921496272567383, | |
| "learning_rate": 2.736561528209684e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -234.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.1493, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.28125, | |
| "rewards/margins": 3.140625, | |
| "rewards/rejected": -5.40625, | |
| "step": 1270 | |
| }, | |
| { | |
| "epoch": 1.5347721822541966, | |
| "grad_norm": 14.590593841694169, | |
| "learning_rate": 2.7143491781430474e-07, | |
| "logits/chosen": -2.21875, | |
| "logits/rejected": -2.234375, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.1147, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.8125, | |
| "rewards/margins": 3.078125, | |
| "rewards/rejected": -5.875, | |
| "step": 1280 | |
| }, | |
| { | |
| "epoch": 1.5467625899280577, | |
| "grad_norm": 16.15992031571578, | |
| "learning_rate": 2.692136828076411e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.1222, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.0625, | |
| "rewards/margins": 3.453125, | |
| "rewards/rejected": -6.53125, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 1.5587529976019185, | |
| "grad_norm": 10.104750544983435, | |
| "learning_rate": 2.6699244780097735e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.1225, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.453125, | |
| "rewards/margins": 3.34375, | |
| "rewards/rejected": -5.8125, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.5707434052757794, | |
| "grad_norm": 16.414889130607445, | |
| "learning_rate": 2.6477121279431363e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -227.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.1865, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.59375, | |
| "rewards/margins": 3.421875, | |
| "rewards/rejected": -6.03125, | |
| "step": 1310 | |
| }, | |
| { | |
| "epoch": 1.5827338129496402, | |
| "grad_norm": 19.612192416903497, | |
| "learning_rate": 2.625499777876499e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.1157, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.484375, | |
| "rewards/margins": 3.265625, | |
| "rewards/rejected": -6.75, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 1.594724220623501, | |
| "grad_norm": 14.340619052654011, | |
| "learning_rate": 2.6032874278098623e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.1074, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.421875, | |
| "rewards/margins": 3.90625, | |
| "rewards/rejected": -6.34375, | |
| "step": 1330 | |
| }, | |
| { | |
| "epoch": 1.6067146282973621, | |
| "grad_norm": 10.947155370834658, | |
| "learning_rate": 2.581075077743225e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.1142, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.53125, | |
| "rewards/margins": 3.640625, | |
| "rewards/rejected": -6.1875, | |
| "step": 1340 | |
| }, | |
| { | |
| "epoch": 1.6187050359712232, | |
| "grad_norm": 6.379407877957059, | |
| "learning_rate": 2.558862727676588e-07, | |
| "logits/chosen": -2.3125, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.1175, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.546875, | |
| "rewards/margins": 3.6875, | |
| "rewards/rejected": -6.21875, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.630695443645084, | |
| "grad_norm": 14.584427754669349, | |
| "learning_rate": 2.536650377609951e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.088, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.65625, | |
| "rewards/margins": 3.515625, | |
| "rewards/rejected": -6.1875, | |
| "step": 1360 | |
| }, | |
| { | |
| "epoch": 1.6426858513189448, | |
| "grad_norm": 18.386633865745754, | |
| "learning_rate": 2.514438027543314e-07, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -2.203125, | |
| "logps/chosen": -240.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0843, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.375, | |
| "rewards/margins": 4.1875, | |
| "rewards/rejected": -6.5625, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 1.6546762589928057, | |
| "grad_norm": 13.249684798008296, | |
| "learning_rate": 2.492225677476677e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.328125, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -282.0, | |
| "loss": 0.11, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.515625, | |
| "rewards/margins": 3.46875, | |
| "rewards/rejected": -5.96875, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 1.6666666666666665, | |
| "grad_norm": 24.989358834998974, | |
| "learning_rate": 2.47001332741004e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -240.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.134, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.59375, | |
| "rewards/margins": 3.3125, | |
| "rewards/rejected": -5.90625, | |
| "step": 1390 | |
| }, | |
| { | |
| "epoch": 1.6786570743405276, | |
| "grad_norm": 21.225948587869443, | |
| "learning_rate": 2.447800977343403e-07, | |
| "logits/chosen": -2.203125, | |
| "logits/rejected": -2.21875, | |
| "logps/chosen": -227.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.1539, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -2.671875, | |
| "rewards/margins": 3.546875, | |
| "rewards/rejected": -6.21875, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.6906474820143886, | |
| "grad_norm": 22.09693803294425, | |
| "learning_rate": 2.425588627276766e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -232.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.1301, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.8125, | |
| "rewards/margins": 3.59375, | |
| "rewards/rejected": -6.40625, | |
| "step": 1410 | |
| }, | |
| { | |
| "epoch": 1.7026378896882495, | |
| "grad_norm": 13.568164124099978, | |
| "learning_rate": 2.403376277210129e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.0997, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.734375, | |
| "rewards/margins": 3.609375, | |
| "rewards/rejected": -7.34375, | |
| "step": 1420 | |
| }, | |
| { | |
| "epoch": 1.7146282973621103, | |
| "grad_norm": 12.536257234153933, | |
| "learning_rate": 2.3811639271434916e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.25, | |
| "logps/chosen": -243.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.1365, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.765625, | |
| "rewards/margins": 3.515625, | |
| "rewards/rejected": -6.28125, | |
| "step": 1430 | |
| }, | |
| { | |
| "epoch": 1.7266187050359711, | |
| "grad_norm": 17.74910297920457, | |
| "learning_rate": 2.3589515770768546e-07, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -256.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.1436, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -3.3125, | |
| "rewards/margins": 3.53125, | |
| "rewards/rejected": -6.84375, | |
| "step": 1440 | |
| }, | |
| { | |
| "epoch": 1.738609112709832, | |
| "grad_norm": 27.091464814087082, | |
| "learning_rate": 2.3367392270102177e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.1031, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.640625, | |
| "rewards/margins": 4.1875, | |
| "rewards/rejected": -6.84375, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.750599520383693, | |
| "grad_norm": 14.870002835228663, | |
| "learning_rate": 2.3145268769435804e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.203125, | |
| "logps/chosen": -247.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.1465, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.046875, | |
| "rewards/margins": 3.609375, | |
| "rewards/rejected": -6.65625, | |
| "step": 1460 | |
| }, | |
| { | |
| "epoch": 1.762589928057554, | |
| "grad_norm": 17.480506529993356, | |
| "learning_rate": 2.2923145268769435e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -223.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.0896, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.484375, | |
| "rewards/margins": 4.3125, | |
| "rewards/rejected": -6.78125, | |
| "step": 1470 | |
| }, | |
| { | |
| "epoch": 1.774580335731415, | |
| "grad_norm": 6.540878635676034, | |
| "learning_rate": 2.2701021768103065e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.25, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.1107, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.703125, | |
| "rewards/margins": 3.984375, | |
| "rewards/rejected": -6.6875, | |
| "step": 1480 | |
| }, | |
| { | |
| "epoch": 1.7865707434052758, | |
| "grad_norm": 16.51964209338177, | |
| "learning_rate": 2.2478898267436695e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -241.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.0963, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.671875, | |
| "rewards/margins": 3.71875, | |
| "rewards/rejected": -7.375, | |
| "step": 1490 | |
| }, | |
| { | |
| "epoch": 1.7985611510791366, | |
| "grad_norm": 17.145067233621255, | |
| "learning_rate": 2.2256774766770323e-07, | |
| "logits/chosen": -2.1875, | |
| "logits/rejected": -2.171875, | |
| "logps/chosen": -234.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.1016, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.71875, | |
| "rewards/margins": 3.953125, | |
| "rewards/rejected": -6.6875, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.8105515587529974, | |
| "grad_norm": 9.418365551008751, | |
| "learning_rate": 2.2034651266103953e-07, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -2.234375, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0902, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.65625, | |
| "rewards/margins": 3.875, | |
| "rewards/rejected": -6.53125, | |
| "step": 1510 | |
| }, | |
| { | |
| "epoch": 1.8225419664268585, | |
| "grad_norm": 8.454143934854207, | |
| "learning_rate": 2.1812527765437583e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.25, | |
| "logps/chosen": -247.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.1036, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.875, | |
| "rewards/margins": 3.75, | |
| "rewards/rejected": -6.625, | |
| "step": 1520 | |
| }, | |
| { | |
| "epoch": 1.8345323741007196, | |
| "grad_norm": 23.877830813468428, | |
| "learning_rate": 2.159040426477121e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.3125, | |
| "logps/chosen": -249.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.0974, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.625, | |
| "rewards/margins": 3.953125, | |
| "rewards/rejected": -7.5625, | |
| "step": 1530 | |
| }, | |
| { | |
| "epoch": 1.8465227817745804, | |
| "grad_norm": 17.59044264644789, | |
| "learning_rate": 2.1368280764104841e-07, | |
| "logits/chosen": -2.171875, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.0921, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.734375, | |
| "rewards/margins": 3.984375, | |
| "rewards/rejected": -7.71875, | |
| "step": 1540 | |
| }, | |
| { | |
| "epoch": 1.8585131894484412, | |
| "grad_norm": 7.53671069585816, | |
| "learning_rate": 2.1146157263438472e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -237.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.0677, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.734375, | |
| "rewards/margins": 4.1875, | |
| "rewards/rejected": -6.9375, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.870503597122302, | |
| "grad_norm": 15.114334538388968, | |
| "learning_rate": 2.09240337627721e-07, | |
| "logits/chosen": -2.28125, | |
| "logits/rejected": -2.25, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.1064, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -3.0, | |
| "rewards/margins": 3.78125, | |
| "rewards/rejected": -6.78125, | |
| "step": 1560 | |
| }, | |
| { | |
| "epoch": 1.882494004796163, | |
| "grad_norm": 24.921093039405555, | |
| "learning_rate": 2.070191026210573e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -246.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.1109, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.6875, | |
| "rewards/margins": 3.46875, | |
| "rewards/rejected": -6.15625, | |
| "step": 1570 | |
| }, | |
| { | |
| "epoch": 1.894484412470024, | |
| "grad_norm": 17.229671646251834, | |
| "learning_rate": 2.047978676143936e-07, | |
| "logits/chosen": -2.203125, | |
| "logits/rejected": -2.21875, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0787, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.609375, | |
| "rewards/margins": 3.875, | |
| "rewards/rejected": -6.5, | |
| "step": 1580 | |
| }, | |
| { | |
| "epoch": 1.906474820143885, | |
| "grad_norm": 10.939803601324265, | |
| "learning_rate": 2.025766326077299e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.0871, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.515625, | |
| "rewards/margins": 4.125, | |
| "rewards/rejected": -6.625, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 1.9184652278177459, | |
| "grad_norm": 16.05247518882454, | |
| "learning_rate": 2.0035539760106618e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -241.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.0611, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.5625, | |
| "rewards/margins": 4.78125, | |
| "rewards/rejected": -7.34375, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.9304556354916067, | |
| "grad_norm": 9.855796335839392, | |
| "learning_rate": 1.9813416259440246e-07, | |
| "logits/chosen": -2.234375, | |
| "logits/rejected": -2.28125, | |
| "logps/chosen": -245.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.0889, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.625, | |
| "rewards/margins": 3.921875, | |
| "rewards/rejected": -7.53125, | |
| "step": 1610 | |
| }, | |
| { | |
| "epoch": 1.9424460431654675, | |
| "grad_norm": 7.909249841156511, | |
| "learning_rate": 1.9591292758773879e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -235.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.1069, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.59375, | |
| "rewards/margins": 4.40625, | |
| "rewards/rejected": -8.0, | |
| "step": 1620 | |
| }, | |
| { | |
| "epoch": 1.9544364508393284, | |
| "grad_norm": 10.195924724786721, | |
| "learning_rate": 1.9369169258107506e-07, | |
| "logits/chosen": -2.328125, | |
| "logits/rejected": -2.234375, | |
| "logps/chosen": -251.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0671, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.296875, | |
| "rewards/margins": 4.0625, | |
| "rewards/rejected": -7.375, | |
| "step": 1630 | |
| }, | |
| { | |
| "epoch": 1.9664268585131894, | |
| "grad_norm": 4.981347556467333, | |
| "learning_rate": 1.9147045757441137e-07, | |
| "logits/chosen": -2.265625, | |
| "logits/rejected": -2.296875, | |
| "logps/chosen": -232.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.0626, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.40625, | |
| "rewards/margins": 4.03125, | |
| "rewards/rejected": -7.4375, | |
| "step": 1640 | |
| }, | |
| { | |
| "epoch": 1.9784172661870505, | |
| "grad_norm": 8.780872639436975, | |
| "learning_rate": 1.8924922256774767e-07, | |
| "logits/chosen": -2.25, | |
| "logits/rejected": -2.1875, | |
| "logps/chosen": -248.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0836, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.84375, | |
| "rewards/margins": 3.953125, | |
| "rewards/rejected": -7.8125, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.9904076738609113, | |
| "grad_norm": 13.586444001383855, | |
| "learning_rate": 1.8702798756108395e-07, | |
| "logits/chosen": -2.34375, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -253.0, | |
| "logps/rejected": -282.0, | |
| "loss": 0.0676, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.046875, | |
| "rewards/margins": 4.21875, | |
| "rewards/rejected": -7.25, | |
| "step": 1660 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -2.265625, | |
| "eval_logits/rejected": -2.296875, | |
| "eval_logps/chosen": -255.0, | |
| "eval_logps/rejected": -270.0, | |
| "eval_loss": 0.8723406195640564, | |
| "eval_rewards/accuracies": 0.6397058963775635, | |
| "eval_rewards/chosen": -4.71875, | |
| "eval_rewards/margins": 1.59375, | |
| "eval_rewards/rejected": -6.3125, | |
| "eval_runtime": 20.0713, | |
| "eval_samples_per_second": 19.979, | |
| "eval_steps_per_second": 0.847, | |
| "step": 1668 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 2502, | |
| "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": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |