| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 2502, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.001199040767386091, | |
| "grad_norm": 20.168596571210266, | |
| "learning_rate": 1.99203187250996e-09, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -226.0, | |
| "logps/rejected": -218.0, | |
| "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": 18.74753353986141, | |
| "learning_rate": 1.99203187250996e-08, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.5390625, | |
| "logps/chosen": -185.0, | |
| "logps/rejected": -176.0, | |
| "loss": 0.6943, | |
| "rewards/accuracies": 0.125, | |
| "rewards/chosen": 0.006256103515625, | |
| "rewards/margins": -0.0093994140625, | |
| "rewards/rejected": 0.015625, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.023980815347721823, | |
| "grad_norm": 20.00098472032445, | |
| "learning_rate": 3.98406374501992e-08, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -182.0, | |
| "logps/rejected": -175.0, | |
| "loss": 0.6993, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": 0.00750732421875, | |
| "rewards/margins": -0.01190185546875, | |
| "rewards/rejected": 0.0194091796875, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03597122302158273, | |
| "grad_norm": 19.082683285927587, | |
| "learning_rate": 5.97609561752988e-08, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -157.0, | |
| "logps/rejected": -167.0, | |
| "loss": 0.6934, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.01531982421875, | |
| "rewards/margins": 0.000308990478515625, | |
| "rewards/rejected": 0.0150146484375, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.047961630695443645, | |
| "grad_norm": 18.804321980067716, | |
| "learning_rate": 7.96812749003984e-08, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -172.0, | |
| "logps/rejected": -160.0, | |
| "loss": 0.6978, | |
| "rewards/accuracies": 0.17499999701976776, | |
| "rewards/chosen": 0.004058837890625, | |
| "rewards/margins": -0.01251220703125, | |
| "rewards/rejected": 0.0166015625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.05995203836930456, | |
| "grad_norm": 17.83886367720322, | |
| "learning_rate": 9.9601593625498e-08, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -185.0, | |
| "logps/rejected": -179.0, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.23749999701976776, | |
| "rewards/chosen": 0.0106201171875, | |
| "rewards/margins": -0.000926971435546875, | |
| "rewards/rejected": 0.0115966796875, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.07194244604316546, | |
| "grad_norm": 22.450580380910267, | |
| "learning_rate": 1.195219123505976e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -170.0, | |
| "logps/rejected": -180.0, | |
| "loss": 0.6885, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": 0.01190185546875, | |
| "rewards/margins": 0.01031494140625, | |
| "rewards/rejected": 0.0015716552734375, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.08393285371702638, | |
| "grad_norm": 19.042392981886824, | |
| "learning_rate": 1.394422310756972e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -180.0, | |
| "logps/rejected": -172.0, | |
| "loss": 0.6887, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.01531982421875, | |
| "rewards/margins": -0.00061798095703125, | |
| "rewards/rejected": 0.0159912109375, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.09592326139088729, | |
| "grad_norm": 19.035173662979318, | |
| "learning_rate": 1.593625498007968e-07, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -178.0, | |
| "logps/rejected": -191.0, | |
| "loss": 0.6917, | |
| "rewards/accuracies": 0.22499999403953552, | |
| "rewards/chosen": 0.01190185546875, | |
| "rewards/margins": -0.00250244140625, | |
| "rewards/rejected": 0.014404296875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.1079136690647482, | |
| "grad_norm": 19.033868457427005, | |
| "learning_rate": 1.7928286852589642e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -190.0, | |
| "loss": 0.6935, | |
| "rewards/accuracies": 0.20000000298023224, | |
| "rewards/chosen": 0.0174560546875, | |
| "rewards/margins": 0.0031280517578125, | |
| "rewards/rejected": 0.014404296875, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.11990407673860912, | |
| "grad_norm": 19.178986080183737, | |
| "learning_rate": 1.99203187250996e-07, | |
| "logits/chosen": -1.4296875, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -162.0, | |
| "logps/rejected": -175.0, | |
| "loss": 0.6941, | |
| "rewards/accuracies": 0.21250000596046448, | |
| "rewards/chosen": 0.00909423828125, | |
| "rewards/margins": -0.00750732421875, | |
| "rewards/rejected": 0.0166015625, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.13189448441247004, | |
| "grad_norm": 18.51014085479843, | |
| "learning_rate": 2.191235059760956e-07, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -170.0, | |
| "logps/rejected": -186.0, | |
| "loss": 0.69, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": 0.0322265625, | |
| "rewards/margins": 0.015625, | |
| "rewards/rejected": 0.0166015625, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.14388489208633093, | |
| "grad_norm": 18.632268534190302, | |
| "learning_rate": 2.390438247011952e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -176.0, | |
| "logps/rejected": -158.0, | |
| "loss": 0.6893, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": 0.031982421875, | |
| "rewards/margins": 0.0234375, | |
| "rewards/rejected": 0.0084228515625, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.15587529976019185, | |
| "grad_norm": 20.464332592180558, | |
| "learning_rate": 2.589641434262948e-07, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -181.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.6818, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": 0.024658203125, | |
| "rewards/margins": 0.017822265625, | |
| "rewards/rejected": 0.00689697265625, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.16786570743405277, | |
| "grad_norm": 18.785864772733888, | |
| "learning_rate": 2.788844621513944e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -184.0, | |
| "logps/rejected": -178.0, | |
| "loss": 0.6853, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": 0.032470703125, | |
| "rewards/margins": 0.01470947265625, | |
| "rewards/rejected": 0.017822265625, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.17985611510791366, | |
| "grad_norm": 18.715152675718606, | |
| "learning_rate": 2.98804780876494e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -181.0, | |
| "logps/rejected": -183.0, | |
| "loss": 0.6824, | |
| "rewards/accuracies": 0.4000000059604645, | |
| "rewards/chosen": 0.0294189453125, | |
| "rewards/margins": 0.027587890625, | |
| "rewards/rejected": 0.0018768310546875, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.19184652278177458, | |
| "grad_norm": 19.40478993623526, | |
| "learning_rate": 3.187250996015936e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -174.0, | |
| "logps/rejected": -182.0, | |
| "loss": 0.6854, | |
| "rewards/accuracies": 0.38749998807907104, | |
| "rewards/chosen": 0.0419921875, | |
| "rewards/margins": 0.026611328125, | |
| "rewards/rejected": 0.01531982421875, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.2038369304556355, | |
| "grad_norm": 18.06542256703035, | |
| "learning_rate": 3.386454183266932e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -188.0, | |
| "loss": 0.6775, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.043212890625, | |
| "rewards/margins": 0.045654296875, | |
| "rewards/rejected": -0.00250244140625, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.2158273381294964, | |
| "grad_norm": 18.751232483418, | |
| "learning_rate": 3.5856573705179284e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -180.0, | |
| "logps/rejected": -183.0, | |
| "loss": 0.673, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.0218505859375, | |
| "rewards/margins": 0.04638671875, | |
| "rewards/rejected": -0.0244140625, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.2278177458033573, | |
| "grad_norm": 17.469498378813427, | |
| "learning_rate": 3.784860557768924e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -169.0, | |
| "logps/rejected": -158.0, | |
| "loss": 0.6819, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.01312255859375, | |
| "rewards/margins": 0.031005859375, | |
| "rewards/rejected": -0.017822265625, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.23980815347721823, | |
| "grad_norm": 19.476666022603656, | |
| "learning_rate": 3.98406374501992e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -181.0, | |
| "logps/rejected": -177.0, | |
| "loss": 0.6788, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 0.01190185546875, | |
| "rewards/margins": 0.0281982421875, | |
| "rewards/rejected": -0.0162353515625, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2517985611510791, | |
| "grad_norm": 15.783370845580528, | |
| "learning_rate": 4.1832669322709163e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -174.0, | |
| "logps/rejected": -182.0, | |
| "loss": 0.6673, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.029052734375, | |
| "rewards/margins": 0.05859375, | |
| "rewards/rejected": -0.0294189453125, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.2637889688249401, | |
| "grad_norm": 17.69748999194774, | |
| "learning_rate": 4.382470119521912e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -185.0, | |
| "logps/rejected": -179.0, | |
| "loss": 0.6637, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.0244140625, | |
| "rewards/margins": 0.058837890625, | |
| "rewards/rejected": -0.08349609375, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.27577937649880097, | |
| "grad_norm": 18.608453625360337, | |
| "learning_rate": 4.581673306772908e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -174.0, | |
| "logps/rejected": -178.0, | |
| "loss": 0.6603, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.045654296875, | |
| "rewards/margins": 0.07470703125, | |
| "rewards/rejected": -0.1201171875, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.28776978417266186, | |
| "grad_norm": 20.06258278752274, | |
| "learning_rate": 4.780876494023904e-07, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -186.0, | |
| "logps/rejected": -172.0, | |
| "loss": 0.652, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.022216796875, | |
| "rewards/margins": 0.11474609375, | |
| "rewards/rejected": -0.13671875, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.2997601918465228, | |
| "grad_norm": 18.760059177733364, | |
| "learning_rate": 4.9800796812749e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -161.0, | |
| "logps/rejected": -173.0, | |
| "loss": 0.649, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.130859375, | |
| "rewards/margins": 0.07568359375, | |
| "rewards/rejected": -0.2060546875, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.3117505995203837, | |
| "grad_norm": 18.23235067079184, | |
| "learning_rate": 4.980008884940026e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -174.0, | |
| "logps/rejected": -174.0, | |
| "loss": 0.649, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.1357421875, | |
| "rewards/margins": 0.1435546875, | |
| "rewards/rejected": -0.279296875, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.3237410071942446, | |
| "grad_norm": 16.049599214987126, | |
| "learning_rate": 4.957796534873389e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -175.0, | |
| "logps/rejected": -171.0, | |
| "loss": 0.6365, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.2109375, | |
| "rewards/margins": 0.150390625, | |
| "rewards/rejected": -0.36328125, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.33573141486810554, | |
| "grad_norm": 16.211195894914958, | |
| "learning_rate": 4.935584184806753e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -176.0, | |
| "logps/rejected": -177.0, | |
| "loss": 0.624, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.330078125, | |
| "rewards/margins": 0.1513671875, | |
| "rewards/rejected": -0.48046875, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.34772182254196643, | |
| "grad_norm": 20.12569666248882, | |
| "learning_rate": 4.913371834740116e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -176.0, | |
| "logps/rejected": -180.0, | |
| "loss": 0.6387, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": -0.498046875, | |
| "rewards/margins": 0.11279296875, | |
| "rewards/rejected": -0.609375, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.3597122302158273, | |
| "grad_norm": 17.476880445567698, | |
| "learning_rate": 4.891159484673478e-07, | |
| "logits/chosen": -1.4296875, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -187.0, | |
| "logps/rejected": -189.0, | |
| "loss": 0.624, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.58984375, | |
| "rewards/margins": 0.1591796875, | |
| "rewards/rejected": -0.74609375, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.37170263788968827, | |
| "grad_norm": 15.507480637830666, | |
| "learning_rate": 4.868947134606841e-07, | |
| "logits/chosen": -1.4140625, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -186.0, | |
| "logps/rejected": -196.0, | |
| "loss": 0.5885, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.734375, | |
| "rewards/margins": 0.3203125, | |
| "rewards/rejected": -1.0546875, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.38369304556354916, | |
| "grad_norm": 18.563307603684667, | |
| "learning_rate": 4.846734784540204e-07, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -203.0, | |
| "loss": 0.587, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.92578125, | |
| "rewards/margins": 0.2734375, | |
| "rewards/rejected": -1.1953125, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.39568345323741005, | |
| "grad_norm": 23.31061683348206, | |
| "learning_rate": 4.824522434473567e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -203.0, | |
| "logps/rejected": -202.0, | |
| "loss": 0.5723, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -1.3125, | |
| "rewards/margins": 0.29296875, | |
| "rewards/rejected": -1.609375, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.407673860911271, | |
| "grad_norm": 16.604718922811248, | |
| "learning_rate": 4.80231008440693e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -195.0, | |
| "loss": 0.55, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.09375, | |
| "rewards/margins": 0.466796875, | |
| "rewards/rejected": -1.5625, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.4196642685851319, | |
| "grad_norm": 21.812258860080746, | |
| "learning_rate": 4.780097734340293e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -199.0, | |
| "loss": 0.5831, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -1.2734375, | |
| "rewards/margins": 0.349609375, | |
| "rewards/rejected": -1.6171875, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.4316546762589928, | |
| "grad_norm": 21.975218415102187, | |
| "learning_rate": 4.757885384273656e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -192.0, | |
| "logps/rejected": -204.0, | |
| "loss": 0.5577, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -1.65625, | |
| "rewards/margins": 0.4375, | |
| "rewards/rejected": -2.09375, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.44364508393285373, | |
| "grad_norm": 18.21517367189722, | |
| "learning_rate": 4.7356730342070187e-07, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -198.0, | |
| "logps/rejected": -202.0, | |
| "loss": 0.5576, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.6171875, | |
| "rewards/margins": 0.5390625, | |
| "rewards/rejected": -2.15625, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.4556354916067146, | |
| "grad_norm": 17.44710723350182, | |
| "learning_rate": 4.713460684140382e-07, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -195.0, | |
| "logps/rejected": -195.0, | |
| "loss": 0.5889, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -1.65625, | |
| "rewards/margins": 0.310546875, | |
| "rewards/rejected": -1.96875, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.4676258992805755, | |
| "grad_norm": 17.175546716735866, | |
| "learning_rate": 4.691248334073745e-07, | |
| "logits/chosen": -1.359375, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -187.0, | |
| "logps/rejected": -199.0, | |
| "loss": 0.5365, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.390625, | |
| "rewards/margins": 0.4375, | |
| "rewards/rejected": -1.828125, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.47961630695443647, | |
| "grad_norm": 21.590406692659766, | |
| "learning_rate": 4.6690359840071075e-07, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.3515625, | |
| "logps/chosen": -171.0, | |
| "logps/rejected": -185.0, | |
| "loss": 0.5199, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.6640625, | |
| "rewards/margins": 0.55078125, | |
| "rewards/rejected": -1.21875, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.49160671462829736, | |
| "grad_norm": 17.161105944130238, | |
| "learning_rate": 4.646823633940471e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -185.0, | |
| "logps/rejected": -194.0, | |
| "loss": 0.5379, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.5625, | |
| "rewards/margins": 0.5078125, | |
| "rewards/rejected": -1.0703125, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.5035971223021583, | |
| "grad_norm": 20.897701687493683, | |
| "learning_rate": 4.6246112838738336e-07, | |
| "logits/chosen": -1.4140625, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -187.0, | |
| "logps/rejected": -187.0, | |
| "loss": 0.5262, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.8125, | |
| "rewards/margins": 0.62890625, | |
| "rewards/rejected": -2.4375, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.5155875299760192, | |
| "grad_norm": 21.505932353833877, | |
| "learning_rate": 4.602398933807197e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -192.0, | |
| "logps/rejected": -225.0, | |
| "loss": 0.5387, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.5078125, | |
| "rewards/margins": 0.435546875, | |
| "rewards/rejected": -1.9375, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.5275779376498801, | |
| "grad_norm": 20.010972596147028, | |
| "learning_rate": 4.5801865837405597e-07, | |
| "logits/chosen": -1.4296875, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -200.0, | |
| "logps/rejected": -203.0, | |
| "loss": 0.5316, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.1796875, | |
| "rewards/margins": 0.53125, | |
| "rewards/rejected": -1.703125, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.539568345323741, | |
| "grad_norm": 19.522837250813602, | |
| "learning_rate": 4.5579742336739224e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -178.0, | |
| "loss": 0.5376, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.8046875, | |
| "rewards/margins": 0.7578125, | |
| "rewards/rejected": -1.5625, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.5515587529976019, | |
| "grad_norm": 19.08163221164428, | |
| "learning_rate": 4.535761883607285e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -209.0, | |
| "logps/rejected": -197.0, | |
| "loss": 0.5262, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.0546875, | |
| "rewards/margins": 0.42578125, | |
| "rewards/rejected": -1.4765625, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.5635491606714629, | |
| "grad_norm": 19.252670782360564, | |
| "learning_rate": 4.513549533540648e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -195.0, | |
| "loss": 0.4992, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.91796875, | |
| "rewards/margins": 0.5859375, | |
| "rewards/rejected": -1.5, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5755395683453237, | |
| "grad_norm": 25.774716921623085, | |
| "learning_rate": 4.491337183474012e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -194.0, | |
| "loss": 0.5192, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.71875, | |
| "rewards/margins": 0.55078125, | |
| "rewards/rejected": -2.28125, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.5875299760191847, | |
| "grad_norm": 14.548009215483642, | |
| "learning_rate": 4.4691248334073746e-07, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -211.0, | |
| "loss": 0.4748, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.6484375, | |
| "rewards/margins": 0.77734375, | |
| "rewards/rejected": -2.421875, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.5995203836930456, | |
| "grad_norm": 25.516181574963138, | |
| "learning_rate": 4.4469124833407373e-07, | |
| "logits/chosen": -1.3828125, | |
| "logits/rejected": -1.375, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -190.0, | |
| "loss": 0.5534, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -2.03125, | |
| "rewards/margins": 0.3984375, | |
| "rewards/rejected": -2.421875, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.6115107913669064, | |
| "grad_norm": 22.58747620927269, | |
| "learning_rate": 4.4247001332741e-07, | |
| "logits/chosen": -1.40625, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -210.0, | |
| "loss": 0.4916, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -1.5078125, | |
| "rewards/margins": 0.8046875, | |
| "rewards/rejected": -2.3125, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.6235011990407674, | |
| "grad_norm": 17.897088219787737, | |
| "learning_rate": 4.402487783207463e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -188.0, | |
| "logps/rejected": -209.0, | |
| "loss": 0.4577, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.4765625, | |
| "rewards/margins": 0.9296875, | |
| "rewards/rejected": -2.40625, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.6354916067146283, | |
| "grad_norm": 16.86391324984424, | |
| "learning_rate": 4.380275433140826e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -201.0, | |
| "loss": 0.4514, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -1.4609375, | |
| "rewards/margins": 0.93359375, | |
| "rewards/rejected": -2.390625, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.6474820143884892, | |
| "grad_norm": 21.115226866999087, | |
| "learning_rate": 4.358063083074189e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -203.0, | |
| "logps/rejected": -216.0, | |
| "loss": 0.4654, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.03125, | |
| "rewards/margins": 1.0234375, | |
| "rewards/rejected": -3.0625, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.6594724220623501, | |
| "grad_norm": 20.504857524123096, | |
| "learning_rate": 4.335850733007552e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -207.0, | |
| "loss": 0.4951, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.4375, | |
| "rewards/margins": 0.67578125, | |
| "rewards/rejected": -3.125, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.6714628297362111, | |
| "grad_norm": 12.606241771319064, | |
| "learning_rate": 4.313638382940915e-07, | |
| "logits/chosen": -1.4296875, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -193.0, | |
| "logps/rejected": -211.0, | |
| "loss": 0.5071, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.8125, | |
| "rewards/margins": 0.8671875, | |
| "rewards/rejected": -2.671875, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.6834532374100719, | |
| "grad_norm": 21.06544638125072, | |
| "learning_rate": 4.291426032874278e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -228.0, | |
| "loss": 0.466, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.703125, | |
| "rewards/margins": 0.8984375, | |
| "rewards/rejected": -2.59375, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.6954436450839329, | |
| "grad_norm": 20.105608002851536, | |
| "learning_rate": 4.269213682807641e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -193.0, | |
| "loss": 0.461, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.2578125, | |
| "rewards/margins": 0.87109375, | |
| "rewards/rejected": -2.125, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.7074340527577938, | |
| "grad_norm": 19.41006530799305, | |
| "learning_rate": 4.247001332741004e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -190.0, | |
| "logps/rejected": -208.0, | |
| "loss": 0.4635, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -1.1015625, | |
| "rewards/margins": 0.8828125, | |
| "rewards/rejected": -1.984375, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.7194244604316546, | |
| "grad_norm": 19.775747633899154, | |
| "learning_rate": 4.2247889826743666e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -202.0, | |
| "loss": 0.532, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.125, | |
| "rewards/margins": 0.74609375, | |
| "rewards/rejected": -1.8671875, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.7314148681055156, | |
| "grad_norm": 20.975128500826628, | |
| "learning_rate": 4.2025766326077294e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -200.0, | |
| "loss": 0.4693, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.6171875, | |
| "rewards/margins": 0.82421875, | |
| "rewards/rejected": -2.4375, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.7434052757793765, | |
| "grad_norm": 19.86967099333255, | |
| "learning_rate": 4.1803642825410926e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -187.0, | |
| "logps/rejected": -208.0, | |
| "loss": 0.4997, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -1.34375, | |
| "rewards/margins": 0.77734375, | |
| "rewards/rejected": -2.125, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.7553956834532374, | |
| "grad_norm": 18.483183442280176, | |
| "learning_rate": 4.158151932474456e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -200.0, | |
| "logps/rejected": -199.0, | |
| "loss": 0.4335, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -1.453125, | |
| "rewards/margins": 1.0546875, | |
| "rewards/rejected": -2.5, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.7673860911270983, | |
| "grad_norm": 27.16526184426748, | |
| "learning_rate": 4.1359395824078187e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -222.0, | |
| "loss": 0.4636, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.15625, | |
| "rewards/margins": 0.92578125, | |
| "rewards/rejected": -3.078125, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.7793764988009593, | |
| "grad_norm": 15.242340480747277, | |
| "learning_rate": 4.1137272323411815e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -216.0, | |
| "loss": 0.4529, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.578125, | |
| "rewards/margins": 0.9296875, | |
| "rewards/rejected": -3.515625, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.7913669064748201, | |
| "grad_norm": 19.772842905766947, | |
| "learning_rate": 4.091514882274544e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -195.0, | |
| "logps/rejected": -194.0, | |
| "loss": 0.5111, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.8828125, | |
| "rewards/margins": 0.59375, | |
| "rewards/rejected": -2.46875, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.8033573141486811, | |
| "grad_norm": 20.04801706897612, | |
| "learning_rate": 4.069302532207907e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -205.0, | |
| "logps/rejected": -205.0, | |
| "loss": 0.4599, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.4765625, | |
| "rewards/margins": 1.1484375, | |
| "rewards/rejected": -2.625, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.815347721822542, | |
| "grad_norm": 16.63035578458501, | |
| "learning_rate": 4.047090182141271e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.3984375, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -208.0, | |
| "loss": 0.4646, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -2.34375, | |
| "rewards/margins": 0.65234375, | |
| "rewards/rejected": -3.0, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.8273381294964028, | |
| "grad_norm": 16.867616740093794, | |
| "learning_rate": 4.0248778320746336e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -213.0, | |
| "loss": 0.4477, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.8515625, | |
| "rewards/margins": 1.0078125, | |
| "rewards/rejected": -2.859375, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.8393285371702638, | |
| "grad_norm": 15.441471505065069, | |
| "learning_rate": 4.0026654820079964e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -207.0, | |
| "logps/rejected": -219.0, | |
| "loss": 0.4527, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -1.6875, | |
| "rewards/margins": 1.0234375, | |
| "rewards/rejected": -2.71875, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.8513189448441247, | |
| "grad_norm": 17.143760908308213, | |
| "learning_rate": 3.980453131941359e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -209.0, | |
| "loss": 0.4297, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.546875, | |
| "rewards/margins": 1.03125, | |
| "rewards/rejected": -2.578125, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.8633093525179856, | |
| "grad_norm": 14.198886301598709, | |
| "learning_rate": 3.958240781874722e-07, | |
| "logits/chosen": -1.4765625, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -198.0, | |
| "logps/rejected": -198.0, | |
| "loss": 0.4112, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.640625, | |
| "rewards/margins": 1.0546875, | |
| "rewards/rejected": -2.6875, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.8752997601918465, | |
| "grad_norm": 19.10768891251934, | |
| "learning_rate": 3.936028431808085e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -200.0, | |
| "logps/rejected": -198.0, | |
| "loss": 0.3902, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.6640625, | |
| "rewards/margins": 1.109375, | |
| "rewards/rejected": -2.765625, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.8872901678657075, | |
| "grad_norm": 15.879596623648883, | |
| "learning_rate": 3.913816081741448e-07, | |
| "logits/chosen": -1.4140625, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -186.0, | |
| "logps/rejected": -204.0, | |
| "loss": 0.3893, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.796875, | |
| "rewards/margins": 1.375, | |
| "rewards/rejected": -3.171875, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.8992805755395683, | |
| "grad_norm": 17.436078043677007, | |
| "learning_rate": 3.8916037316748113e-07, | |
| "logits/chosen": -1.421875, | |
| "logits/rejected": -1.421875, | |
| "logps/chosen": -193.0, | |
| "logps/rejected": -200.0, | |
| "loss": 0.4263, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.3125, | |
| "rewards/margins": 0.91796875, | |
| "rewards/rejected": -3.234375, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.9112709832134293, | |
| "grad_norm": 15.704331891226987, | |
| "learning_rate": 3.869391381608174e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -201.0, | |
| "logps/rejected": -211.0, | |
| "loss": 0.3875, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.9609375, | |
| "rewards/margins": 1.015625, | |
| "rewards/rejected": -2.984375, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.9232613908872902, | |
| "grad_norm": 19.28806504511243, | |
| "learning_rate": 3.847179031541537e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -190.0, | |
| "loss": 0.421, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -1.453125, | |
| "rewards/margins": 1.046875, | |
| "rewards/rejected": -2.5, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.935251798561151, | |
| "grad_norm": 17.706917404985195, | |
| "learning_rate": 3.8249666814749e-07, | |
| "logits/chosen": -1.4375, | |
| "logits/rejected": -1.40625, | |
| "logps/chosen": -205.0, | |
| "logps/rejected": -201.0, | |
| "loss": 0.3674, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -1.78125, | |
| "rewards/margins": 1.25, | |
| "rewards/rejected": -3.03125, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.947242206235012, | |
| "grad_norm": 14.788710817068432, | |
| "learning_rate": 3.802754331408263e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4453125, | |
| "logps/chosen": -198.0, | |
| "logps/rejected": -222.0, | |
| "loss": 0.4113, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.171875, | |
| "rewards/margins": 1.078125, | |
| "rewards/rejected": -3.25, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.9592326139088729, | |
| "grad_norm": 20.22248928411642, | |
| "learning_rate": 3.7805419813416256e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -192.0, | |
| "logps/rejected": -204.0, | |
| "loss": 0.3646, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -2.3125, | |
| "rewards/margins": 1.15625, | |
| "rewards/rejected": -3.484375, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.9712230215827338, | |
| "grad_norm": 13.171309874145647, | |
| "learning_rate": 3.7583296312749884e-07, | |
| "logits/chosen": -1.3984375, | |
| "logits/rejected": -1.4296875, | |
| "logps/chosen": -198.0, | |
| "logps/rejected": -215.0, | |
| "loss": 0.4278, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -2.234375, | |
| "rewards/margins": 1.34375, | |
| "rewards/rejected": -3.578125, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.9832134292565947, | |
| "grad_norm": 19.50680457515295, | |
| "learning_rate": 3.7361172812083517e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.4140625, | |
| "logps/chosen": -195.0, | |
| "logps/rejected": -196.0, | |
| "loss": 0.3729, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -1.640625, | |
| "rewards/margins": 1.1953125, | |
| "rewards/rejected": -2.828125, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.9952038369304557, | |
| "grad_norm": 19.43880275481757, | |
| "learning_rate": 3.713904931141715e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -202.0, | |
| "logps/rejected": -202.0, | |
| "loss": 0.3305, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -1.3203125, | |
| "rewards/margins": 1.2890625, | |
| "rewards/rejected": -2.609375, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_logits/chosen": -1.4765625, | |
| "eval_logits/rejected": -1.484375, | |
| "eval_logps/chosen": -190.0, | |
| "eval_logps/rejected": -195.0, | |
| "eval_loss": 0.6123461127281189, | |
| "eval_rewards/accuracies": 0.6029411554336548, | |
| "eval_rewards/chosen": -1.2265625, | |
| "eval_rewards/margins": 0.63671875, | |
| "eval_rewards/rejected": -1.8671875, | |
| "eval_runtime": 23.6977, | |
| "eval_samples_per_second": 16.921, | |
| "eval_steps_per_second": 0.717, | |
| "step": 834 | |
| }, | |
| { | |
| "epoch": 1.0071942446043165, | |
| "grad_norm": 17.925966726812263, | |
| "learning_rate": 3.691692581075078e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -198.0, | |
| "logps/rejected": -228.0, | |
| "loss": 0.3292, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.28125, | |
| "rewards/margins": 1.6328125, | |
| "rewards/rejected": -2.921875, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 1.0191846522781776, | |
| "grad_norm": 15.13245137359342, | |
| "learning_rate": 3.6694802310084405e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -209.0, | |
| "loss": 0.2941, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.390625, | |
| "rewards/margins": 1.390625, | |
| "rewards/rejected": -2.78125, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 1.0311750599520384, | |
| "grad_norm": 15.85613126792236, | |
| "learning_rate": 3.6472678809418033e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -200.0, | |
| "loss": 0.3029, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.28125, | |
| "rewards/margins": 1.65625, | |
| "rewards/rejected": -2.9375, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 1.0431654676258992, | |
| "grad_norm": 17.934043297524138, | |
| "learning_rate": 3.625055530875166e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.4765625, | |
| "logps/chosen": -201.0, | |
| "logps/rejected": -217.0, | |
| "loss": 0.2741, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -1.59375, | |
| "rewards/margins": 1.828125, | |
| "rewards/rejected": -3.421875, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 1.0551558752997603, | |
| "grad_norm": 12.6431049483951, | |
| "learning_rate": 3.6028431808085294e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.46875, | |
| "logps/chosen": -203.0, | |
| "logps/rejected": -221.0, | |
| "loss": 0.2871, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.7890625, | |
| "rewards/margins": 1.5859375, | |
| "rewards/rejected": -3.375, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 1.0671462829736211, | |
| "grad_norm": 14.781119124887237, | |
| "learning_rate": 3.5806308307418926e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -200.0, | |
| "logps/rejected": -222.0, | |
| "loss": 0.2536, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.90625, | |
| "rewards/margins": 1.828125, | |
| "rewards/rejected": -3.71875, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 1.079136690647482, | |
| "grad_norm": 13.675786724224842, | |
| "learning_rate": 3.5584184806752554e-07, | |
| "logits/chosen": -1.46875, | |
| "logits/rejected": -1.453125, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -216.0, | |
| "loss": 0.2305, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.84375, | |
| "rewards/margins": 2.03125, | |
| "rewards/rejected": -3.875, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.091127098321343, | |
| "grad_norm": 15.133495835048825, | |
| "learning_rate": 3.536206130608618e-07, | |
| "logits/chosen": -1.4609375, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -208.0, | |
| "loss": 0.2446, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -1.7734375, | |
| "rewards/margins": 2.015625, | |
| "rewards/rejected": -3.796875, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 1.1031175059952039, | |
| "grad_norm": 17.223627578304455, | |
| "learning_rate": 3.513993780541981e-07, | |
| "logits/chosen": -1.453125, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -172.0, | |
| "logps/rejected": -210.0, | |
| "loss": 0.2433, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.796875, | |
| "rewards/margins": 1.9765625, | |
| "rewards/rejected": -3.765625, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 1.1151079136690647, | |
| "grad_norm": 12.606788377121838, | |
| "learning_rate": 3.491781430475344e-07, | |
| "logits/chosen": -1.5390625, | |
| "logits/rejected": -1.5546875, | |
| "logps/chosen": -192.0, | |
| "logps/rejected": -231.0, | |
| "loss": 0.2498, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -1.78125, | |
| "rewards/margins": 1.796875, | |
| "rewards/rejected": -3.578125, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 1.1270983213429258, | |
| "grad_norm": 20.113412627019386, | |
| "learning_rate": 3.469569080408707e-07, | |
| "logits/chosen": -1.4453125, | |
| "logits/rejected": -1.4375, | |
| "logps/chosen": -185.0, | |
| "logps/rejected": -204.0, | |
| "loss": 0.258, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -1.5234375, | |
| "rewards/margins": 2.125, | |
| "rewards/rejected": -3.640625, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 1.1390887290167866, | |
| "grad_norm": 22.896453855913, | |
| "learning_rate": 3.4473567303420703e-07, | |
| "logits/chosen": -1.53125, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -200.0, | |
| "logps/rejected": -217.0, | |
| "loss": 0.2356, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.0, | |
| "rewards/margins": 2.09375, | |
| "rewards/rejected": -4.09375, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 1.1510791366906474, | |
| "grad_norm": 17.597413512159072, | |
| "learning_rate": 3.425144380275433e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -213.0, | |
| "logps/rejected": -235.0, | |
| "loss": 0.2713, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.359375, | |
| "rewards/margins": 1.8984375, | |
| "rewards/rejected": -4.25, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 1.1630695443645085, | |
| "grad_norm": 14.323268951448473, | |
| "learning_rate": 3.402932030208796e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.5, | |
| "logps/chosen": -210.0, | |
| "logps/rejected": -235.0, | |
| "loss": 0.2388, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.40625, | |
| "rewards/margins": 2.171875, | |
| "rewards/rejected": -4.5625, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 1.1750599520383693, | |
| "grad_norm": 13.502480458522491, | |
| "learning_rate": 3.380719680142159e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.546875, | |
| "logps/chosen": -192.0, | |
| "logps/rejected": -209.0, | |
| "loss": 0.2241, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -1.6953125, | |
| "rewards/margins": 2.078125, | |
| "rewards/rejected": -3.765625, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 1.1870503597122302, | |
| "grad_norm": 12.261596643016196, | |
| "learning_rate": 3.358507330075522e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.5390625, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -219.0, | |
| "loss": 0.2554, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.140625, | |
| "rewards/margins": 1.953125, | |
| "rewards/rejected": -4.09375, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 1.1990407673860912, | |
| "grad_norm": 11.618624189476918, | |
| "learning_rate": 3.3362949800088847e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -204.0, | |
| "logps/rejected": -222.0, | |
| "loss": 0.2051, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.71875, | |
| "rewards/margins": 2.171875, | |
| "rewards/rejected": -3.890625, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.211031175059952, | |
| "grad_norm": 22.814769403029445, | |
| "learning_rate": 3.3140826299422474e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -201.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.2417, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.015625, | |
| "rewards/margins": 2.25, | |
| "rewards/rejected": -4.28125, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 1.223021582733813, | |
| "grad_norm": 23.105005233123116, | |
| "learning_rate": 3.291870279875611e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.578125, | |
| "logps/chosen": -209.0, | |
| "logps/rejected": -235.0, | |
| "loss": 0.2123, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.4375, | |
| "rewards/margins": 2.21875, | |
| "rewards/rejected": -4.65625, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 1.235011990407674, | |
| "grad_norm": 16.59163456560542, | |
| "learning_rate": 3.269657929808974e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.2316, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.3125, | |
| "rewards/margins": 1.953125, | |
| "rewards/rejected": -4.28125, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 1.2470023980815348, | |
| "grad_norm": 16.97136930640841, | |
| "learning_rate": 3.247445579742337e-07, | |
| "logits/chosen": -1.4921875, | |
| "logits/rejected": -1.4609375, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -228.0, | |
| "loss": 0.233, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.640625, | |
| "rewards/margins": 2.078125, | |
| "rewards/rejected": -4.71875, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 1.2589928057553956, | |
| "grad_norm": 11.658846274024977, | |
| "learning_rate": 3.2252332296756996e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.5546875, | |
| "logps/chosen": -191.0, | |
| "logps/rejected": -217.0, | |
| "loss": 0.2061, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.078125, | |
| "rewards/margins": 2.265625, | |
| "rewards/rejected": -4.34375, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 1.2709832134292567, | |
| "grad_norm": 17.907683758231656, | |
| "learning_rate": 3.2030208796090623e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -218.0, | |
| "loss": 0.2564, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.125, | |
| "rewards/margins": 2.46875, | |
| "rewards/rejected": -4.59375, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 1.2829736211031175, | |
| "grad_norm": 13.891330543214426, | |
| "learning_rate": 3.180808529542425e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -236.0, | |
| "loss": 0.2253, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.5625, | |
| "rewards/margins": 2.078125, | |
| "rewards/rejected": -4.625, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 1.2949640287769784, | |
| "grad_norm": 12.255143244628048, | |
| "learning_rate": 3.1585961794757884e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.5546875, | |
| "logps/chosen": -213.0, | |
| "logps/rejected": -220.0, | |
| "loss": 0.1986, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.15625, | |
| "rewards/margins": 2.25, | |
| "rewards/rejected": -4.40625, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 1.3069544364508392, | |
| "grad_norm": 16.986661579673758, | |
| "learning_rate": 3.1363838294091517e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -201.0, | |
| "logps/rejected": -215.0, | |
| "loss": 0.1726, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -1.5078125, | |
| "rewards/margins": 2.515625, | |
| "rewards/rejected": -4.0, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 1.3189448441247003, | |
| "grad_norm": 11.362437294228451, | |
| "learning_rate": 3.1141714793425145e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.5546875, | |
| "logps/chosen": -217.0, | |
| "logps/rejected": -240.0, | |
| "loss": 0.1936, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.265625, | |
| "rewards/margins": 2.34375, | |
| "rewards/rejected": -4.625, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.330935251798561, | |
| "grad_norm": 18.06174034251652, | |
| "learning_rate": 3.091959129275877e-07, | |
| "logits/chosen": -1.5, | |
| "logits/rejected": -1.4921875, | |
| "logps/chosen": -184.0, | |
| "logps/rejected": -205.0, | |
| "loss": 0.2132, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -1.78125, | |
| "rewards/margins": 2.40625, | |
| "rewards/rejected": -4.1875, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 1.3429256594724222, | |
| "grad_norm": 21.343977198752015, | |
| "learning_rate": 3.06974677920924e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.5078125, | |
| "logps/chosen": -221.0, | |
| "logps/rejected": -240.0, | |
| "loss": 0.2055, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -2.90625, | |
| "rewards/margins": 2.171875, | |
| "rewards/rejected": -5.09375, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 1.354916067146283, | |
| "grad_norm": 19.364261897999658, | |
| "learning_rate": 3.0475344291426033e-07, | |
| "logits/chosen": -1.546875, | |
| "logits/rejected": -1.515625, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -230.0, | |
| "loss": 0.1925, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.140625, | |
| "rewards/margins": 2.90625, | |
| "rewards/rejected": -5.03125, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 1.3669064748201438, | |
| "grad_norm": 12.49182324344103, | |
| "learning_rate": 3.025322079075966e-07, | |
| "logits/chosen": -1.484375, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.2206, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.453125, | |
| "rewards/margins": 2.65625, | |
| "rewards/rejected": -5.09375, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 1.3788968824940047, | |
| "grad_norm": 12.604128787864076, | |
| "learning_rate": 3.003109729009329e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.5234375, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -235.0, | |
| "loss": 0.2102, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.671875, | |
| "rewards/margins": 2.1875, | |
| "rewards/rejected": -4.875, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.3908872901678657, | |
| "grad_norm": 16.0572765821895, | |
| "learning_rate": 2.980897378942692e-07, | |
| "logits/chosen": -1.5546875, | |
| "logits/rejected": -1.546875, | |
| "logps/chosen": -207.0, | |
| "logps/rejected": -226.0, | |
| "loss": 0.1899, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.6875, | |
| "rewards/margins": 2.171875, | |
| "rewards/rejected": -4.84375, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 1.4028776978417266, | |
| "grad_norm": 14.62570261847443, | |
| "learning_rate": 2.958685028876055e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.546875, | |
| "logps/chosen": -204.0, | |
| "logps/rejected": -223.0, | |
| "loss": 0.2256, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -2.1875, | |
| "rewards/margins": 2.28125, | |
| "rewards/rejected": -4.46875, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 1.4148681055155876, | |
| "grad_norm": 16.69089195431221, | |
| "learning_rate": 2.936472678809418e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.5703125, | |
| "logps/chosen": -215.0, | |
| "logps/rejected": -234.0, | |
| "loss": 0.2164, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.625, | |
| "rewards/margins": 2.265625, | |
| "rewards/rejected": -4.90625, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 1.4268585131894485, | |
| "grad_norm": 14.208012618540776, | |
| "learning_rate": 2.914260328742781e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.546875, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -226.0, | |
| "loss": 0.1728, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.90625, | |
| "rewards/margins": 2.21875, | |
| "rewards/rejected": -5.125, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 1.4388489208633093, | |
| "grad_norm": 19.814051144030756, | |
| "learning_rate": 2.8920479786761437e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.59375, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -232.0, | |
| "loss": 0.2096, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.484375, | |
| "rewards/margins": 2.40625, | |
| "rewards/rejected": -4.875, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.4508393285371701, | |
| "grad_norm": 16.967462896165113, | |
| "learning_rate": 2.8698356286095065e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.5703125, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.1481, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.5625, | |
| "rewards/margins": 2.796875, | |
| "rewards/rejected": -5.34375, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 1.4628297362110312, | |
| "grad_norm": 15.206385611452236, | |
| "learning_rate": 2.847623278542869e-07, | |
| "logits/chosen": -1.609375, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -208.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.1846, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.5625, | |
| "rewards/margins": 2.375, | |
| "rewards/rejected": -4.9375, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 1.474820143884892, | |
| "grad_norm": 17.869462519073895, | |
| "learning_rate": 2.825410928476233e-07, | |
| "logits/chosen": -1.515625, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -187.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.1821, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.703125, | |
| "rewards/margins": 2.578125, | |
| "rewards/rejected": -5.28125, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 1.486810551558753, | |
| "grad_norm": 11.533712996176597, | |
| "learning_rate": 2.803198578409596e-07, | |
| "logits/chosen": -1.609375, | |
| "logits/rejected": -1.5625, | |
| "logps/chosen": -204.0, | |
| "logps/rejected": -228.0, | |
| "loss": 0.2133, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.703125, | |
| "rewards/margins": 2.515625, | |
| "rewards/rejected": -5.21875, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 1.498800959232614, | |
| "grad_norm": 11.144943270392623, | |
| "learning_rate": 2.7809862283429586e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.5703125, | |
| "logps/chosen": -193.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.1494, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.3125, | |
| "rewards/margins": 3.078125, | |
| "rewards/rejected": -5.375, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.5107913669064748, | |
| "grad_norm": 15.93725977415015, | |
| "learning_rate": 2.7587738782763214e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.59375, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -236.0, | |
| "loss": 0.204, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -3.53125, | |
| "rewards/margins": 2.546875, | |
| "rewards/rejected": -6.0625, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 1.5227817745803356, | |
| "grad_norm": 23.59190890231536, | |
| "learning_rate": 2.736561528209684e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -210.0, | |
| "logps/rejected": -244.0, | |
| "loss": 0.1853, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -3.09375, | |
| "rewards/margins": 2.625, | |
| "rewards/rejected": -5.71875, | |
| "step": 1270 | |
| }, | |
| { | |
| "epoch": 1.5347721822541966, | |
| "grad_norm": 22.364117581141223, | |
| "learning_rate": 2.7143491781430474e-07, | |
| "logits/chosen": -1.5078125, | |
| "logits/rejected": -1.5390625, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.1952, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.78125, | |
| "rewards/margins": 2.296875, | |
| "rewards/rejected": -5.09375, | |
| "step": 1280 | |
| }, | |
| { | |
| "epoch": 1.5467625899280577, | |
| "grad_norm": 11.715030229741336, | |
| "learning_rate": 2.692136828076411e-07, | |
| "logits/chosen": -1.578125, | |
| "logits/rejected": -1.5546875, | |
| "logps/chosen": -210.0, | |
| "logps/rejected": -234.0, | |
| "loss": 0.1717, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.96875, | |
| "rewards/margins": 2.609375, | |
| "rewards/rejected": -5.5625, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 1.5587529976019185, | |
| "grad_norm": 16.74676425409118, | |
| "learning_rate": 2.6699244780097735e-07, | |
| "logits/chosen": -1.609375, | |
| "logits/rejected": -1.59375, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -231.0, | |
| "loss": 0.1883, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.875, | |
| "rewards/margins": 2.515625, | |
| "rewards/rejected": -5.375, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.5707434052757794, | |
| "grad_norm": 16.31256275621072, | |
| "learning_rate": 2.6477121279431363e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.5859375, | |
| "logps/chosen": -194.0, | |
| "logps/rejected": -226.0, | |
| "loss": 0.2258, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -2.5, | |
| "rewards/margins": 2.4375, | |
| "rewards/rejected": -4.9375, | |
| "step": 1310 | |
| }, | |
| { | |
| "epoch": 1.5827338129496402, | |
| "grad_norm": 15.174270462866287, | |
| "learning_rate": 2.625499777876499e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.5703125, | |
| "logps/chosen": -204.0, | |
| "logps/rejected": -229.0, | |
| "loss": 0.1577, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -3.375, | |
| "rewards/margins": 2.421875, | |
| "rewards/rejected": -5.78125, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 1.594724220623501, | |
| "grad_norm": 20.949271053556263, | |
| "learning_rate": 2.6032874278098623e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -218.0, | |
| "logps/rejected": -234.0, | |
| "loss": 0.1542, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.46875, | |
| "rewards/margins": 3.078125, | |
| "rewards/rejected": -5.5625, | |
| "step": 1330 | |
| }, | |
| { | |
| "epoch": 1.6067146282973621, | |
| "grad_norm": 21.582166006491406, | |
| "learning_rate": 2.581075077743225e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -215.0, | |
| "logps/rejected": -231.0, | |
| "loss": 0.1642, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.0625, | |
| "rewards/margins": 2.703125, | |
| "rewards/rejected": -5.75, | |
| "step": 1340 | |
| }, | |
| { | |
| "epoch": 1.6187050359712232, | |
| "grad_norm": 12.490052192808845, | |
| "learning_rate": 2.558862727676588e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.546875, | |
| "logps/chosen": -211.0, | |
| "logps/rejected": -227.0, | |
| "loss": 0.1549, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.71875, | |
| "rewards/margins": 2.921875, | |
| "rewards/rejected": -5.625, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.630695443645084, | |
| "grad_norm": 14.27659020234575, | |
| "learning_rate": 2.536650377609951e-07, | |
| "logits/chosen": -1.6484375, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -208.0, | |
| "logps/rejected": -230.0, | |
| "loss": 0.1123, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.515625, | |
| "rewards/margins": 3.0625, | |
| "rewards/rejected": -5.5625, | |
| "step": 1360 | |
| }, | |
| { | |
| "epoch": 1.6426858513189448, | |
| "grad_norm": 14.433767508436462, | |
| "learning_rate": 2.514438027543314e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.5625, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -235.0, | |
| "loss": 0.1312, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.359375, | |
| "rewards/margins": 3.359375, | |
| "rewards/rejected": -5.71875, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 1.6546762589928057, | |
| "grad_norm": 13.424589730956683, | |
| "learning_rate": 2.492225677476677e-07, | |
| "logits/chosen": -1.609375, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -198.0, | |
| "logps/rejected": -242.0, | |
| "loss": 0.1744, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -2.609375, | |
| "rewards/margins": 2.71875, | |
| "rewards/rejected": -5.34375, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 1.6666666666666665, | |
| "grad_norm": 23.17900601362576, | |
| "learning_rate": 2.47001332741004e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.5703125, | |
| "logps/chosen": -203.0, | |
| "logps/rejected": -217.0, | |
| "loss": 0.1801, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.109375, | |
| "rewards/margins": 2.8125, | |
| "rewards/rejected": -4.90625, | |
| "step": 1390 | |
| }, | |
| { | |
| "epoch": 1.6786570743405276, | |
| "grad_norm": 15.035036612326222, | |
| "learning_rate": 2.447800977343403e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.5625, | |
| "logps/chosen": -195.0, | |
| "logps/rejected": -226.0, | |
| "loss": 0.1659, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -2.78125, | |
| "rewards/margins": 2.796875, | |
| "rewards/rejected": -5.5625, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.6906474820143886, | |
| "grad_norm": 23.256101786444912, | |
| "learning_rate": 2.425588627276766e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.5859375, | |
| "logps/chosen": -199.0, | |
| "logps/rejected": -238.0, | |
| "loss": 0.15, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.734375, | |
| "rewards/margins": 3.140625, | |
| "rewards/rejected": -5.875, | |
| "step": 1410 | |
| }, | |
| { | |
| "epoch": 1.7026378896882495, | |
| "grad_norm": 16.44114210272047, | |
| "learning_rate": 2.403376277210129e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.578125, | |
| "logps/chosen": -207.0, | |
| "logps/rejected": -242.0, | |
| "loss": 0.1413, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -3.265625, | |
| "rewards/margins": 2.890625, | |
| "rewards/rejected": -6.125, | |
| "step": 1420 | |
| }, | |
| { | |
| "epoch": 1.7146282973621103, | |
| "grad_norm": 19.104299358854444, | |
| "learning_rate": 2.3811639271434916e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.5390625, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -240.0, | |
| "loss": 0.1638, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -3.265625, | |
| "rewards/margins": 2.875, | |
| "rewards/rejected": -6.15625, | |
| "step": 1430 | |
| }, | |
| { | |
| "epoch": 1.7266187050359711, | |
| "grad_norm": 8.57838400028329, | |
| "learning_rate": 2.3589515770768546e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.5546875, | |
| "logps/chosen": -221.0, | |
| "logps/rejected": -236.0, | |
| "loss": 0.1491, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -3.09375, | |
| "rewards/margins": 2.875, | |
| "rewards/rejected": -5.96875, | |
| "step": 1440 | |
| }, | |
| { | |
| "epoch": 1.738609112709832, | |
| "grad_norm": 12.545882511992975, | |
| "learning_rate": 2.3367392270102177e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -200.0, | |
| "logps/rejected": -252.0, | |
| "loss": 0.1236, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.328125, | |
| "rewards/margins": 3.359375, | |
| "rewards/rejected": -5.6875, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.750599520383693, | |
| "grad_norm": 16.11932503670354, | |
| "learning_rate": 2.3145268769435804e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.578125, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -242.0, | |
| "loss": 0.1705, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.078125, | |
| "rewards/margins": 3.125, | |
| "rewards/rejected": -6.21875, | |
| "step": 1460 | |
| }, | |
| { | |
| "epoch": 1.762589928057554, | |
| "grad_norm": 19.767212861451107, | |
| "learning_rate": 2.2923145268769435e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -197.0, | |
| "logps/rejected": -232.0, | |
| "loss": 0.1243, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.96875, | |
| "rewards/margins": 3.59375, | |
| "rewards/rejected": -6.5625, | |
| "step": 1470 | |
| }, | |
| { | |
| "epoch": 1.774580335731415, | |
| "grad_norm": 17.770676013426908, | |
| "learning_rate": 2.2701021768103065e-07, | |
| "logits/chosen": -1.609375, | |
| "logits/rejected": -1.5859375, | |
| "logps/chosen": -212.0, | |
| "logps/rejected": -244.0, | |
| "loss": 0.162, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -3.5, | |
| "rewards/margins": 3.0625, | |
| "rewards/rejected": -6.5625, | |
| "step": 1480 | |
| }, | |
| { | |
| "epoch": 1.7865707434052758, | |
| "grad_norm": 9.354271003661482, | |
| "learning_rate": 2.2478898267436695e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -207.0, | |
| "logps/rejected": -245.0, | |
| "loss": 0.1463, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.421875, | |
| "rewards/margins": 3.109375, | |
| "rewards/rejected": -6.53125, | |
| "step": 1490 | |
| }, | |
| { | |
| "epoch": 1.7985611510791366, | |
| "grad_norm": 16.23301041511246, | |
| "learning_rate": 2.2256774766770323e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.53125, | |
| "logps/chosen": -205.0, | |
| "logps/rejected": -218.0, | |
| "loss": 0.147, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.046875, | |
| "rewards/margins": 3.0625, | |
| "rewards/rejected": -6.125, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.8105515587529974, | |
| "grad_norm": 10.88674570833714, | |
| "learning_rate": 2.2034651266103953e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.5546875, | |
| "logps/chosen": -186.0, | |
| "logps/rejected": -227.0, | |
| "loss": 0.1292, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.734375, | |
| "rewards/margins": 3.109375, | |
| "rewards/rejected": -5.84375, | |
| "step": 1510 | |
| }, | |
| { | |
| "epoch": 1.8225419664268585, | |
| "grad_norm": 11.649861361880111, | |
| "learning_rate": 2.1812527765437583e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.578125, | |
| "logps/chosen": -211.0, | |
| "logps/rejected": -235.0, | |
| "loss": 0.1467, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -2.5, | |
| "rewards/margins": 3.046875, | |
| "rewards/rejected": -5.5625, | |
| "step": 1520 | |
| }, | |
| { | |
| "epoch": 1.8345323741007196, | |
| "grad_norm": 17.611539910897193, | |
| "learning_rate": 2.159040426477121e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.1337, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.546875, | |
| "rewards/margins": 3.0625, | |
| "rewards/rejected": -6.59375, | |
| "step": 1530 | |
| }, | |
| { | |
| "epoch": 1.8465227817745804, | |
| "grad_norm": 20.32024320505495, | |
| "learning_rate": 2.1368280764104841e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.6015625, | |
| "logps/chosen": -210.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.1444, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.59375, | |
| "rewards/margins": 3.421875, | |
| "rewards/rejected": -7.0, | |
| "step": 1540 | |
| }, | |
| { | |
| "epoch": 1.8585131894484412, | |
| "grad_norm": 10.61654419093803, | |
| "learning_rate": 2.1146157263438472e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -208.0, | |
| "logps/rejected": -241.0, | |
| "loss": 0.1117, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.140625, | |
| "rewards/margins": 3.390625, | |
| "rewards/rejected": -6.53125, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.870503597122302, | |
| "grad_norm": 15.023137073497749, | |
| "learning_rate": 2.09240337627721e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.59375, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -225.0, | |
| "loss": 0.1284, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": -3.40625, | |
| "rewards/margins": 2.96875, | |
| "rewards/rejected": -6.375, | |
| "step": 1560 | |
| }, | |
| { | |
| "epoch": 1.882494004796163, | |
| "grad_norm": 23.48973810623808, | |
| "learning_rate": 2.070191026210573e-07, | |
| "logits/chosen": -1.5703125, | |
| "logits/rejected": -1.6015625, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.1468, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -2.953125, | |
| "rewards/margins": 2.96875, | |
| "rewards/rejected": -5.9375, | |
| "step": 1570 | |
| }, | |
| { | |
| "epoch": 1.894484412470024, | |
| "grad_norm": 9.16497112579326, | |
| "learning_rate": 2.047978676143936e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.578125, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -242.0, | |
| "loss": 0.1041, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.84375, | |
| "rewards/margins": 3.453125, | |
| "rewards/rejected": -6.28125, | |
| "step": 1580 | |
| }, | |
| { | |
| "epoch": 1.906474820143885, | |
| "grad_norm": 11.983080775973534, | |
| "learning_rate": 2.025766326077299e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -207.0, | |
| "logps/rejected": -248.0, | |
| "loss": 0.1184, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.984375, | |
| "rewards/margins": 3.28125, | |
| "rewards/rejected": -6.25, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 1.9184652278177459, | |
| "grad_norm": 20.978396501783582, | |
| "learning_rate": 2.0035539760106618e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -212.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.1042, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.859375, | |
| "rewards/margins": 3.53125, | |
| "rewards/rejected": -6.40625, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.9304556354916067, | |
| "grad_norm": 9.307796901849636, | |
| "learning_rate": 1.9813416259440246e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.6015625, | |
| "logps/chosen": -209.0, | |
| "logps/rejected": -233.0, | |
| "loss": 0.1347, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.171875, | |
| "rewards/margins": 2.984375, | |
| "rewards/rejected": -6.15625, | |
| "step": 1610 | |
| }, | |
| { | |
| "epoch": 1.9424460431654675, | |
| "grad_norm": 22.24759097187433, | |
| "learning_rate": 1.9591292758773879e-07, | |
| "logits/chosen": -1.625, | |
| "logits/rejected": -1.6015625, | |
| "logps/chosen": -196.0, | |
| "logps/rejected": -251.0, | |
| "loss": 0.1543, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -2.765625, | |
| "rewards/margins": 3.640625, | |
| "rewards/rejected": -6.40625, | |
| "step": 1620 | |
| }, | |
| { | |
| "epoch": 1.9544364508393284, | |
| "grad_norm": 22.879947315900267, | |
| "learning_rate": 1.9369169258107506e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.5703125, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -239.0, | |
| "loss": 0.1174, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.84375, | |
| "rewards/margins": 3.359375, | |
| "rewards/rejected": -6.21875, | |
| "step": 1630 | |
| }, | |
| { | |
| "epoch": 1.9664268585131894, | |
| "grad_norm": 10.630808616318225, | |
| "learning_rate": 1.9147045757441137e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -205.0, | |
| "logps/rejected": -249.0, | |
| "loss": 0.113, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.6875, | |
| "rewards/margins": 3.140625, | |
| "rewards/rejected": -6.8125, | |
| "step": 1640 | |
| }, | |
| { | |
| "epoch": 1.9784172661870505, | |
| "grad_norm": 20.455023304923063, | |
| "learning_rate": 1.8924922256774767e-07, | |
| "logits/chosen": -1.5625, | |
| "logits/rejected": -1.484375, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -230.0, | |
| "loss": 0.133, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.46875, | |
| "rewards/margins": 2.84375, | |
| "rewards/rejected": -6.3125, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.9904076738609113, | |
| "grad_norm": 12.942728073820737, | |
| "learning_rate": 1.8702798756108395e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.578125, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -234.0, | |
| "loss": 0.1466, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -2.71875, | |
| "rewards/margins": 3.078125, | |
| "rewards/rejected": -5.78125, | |
| "step": 1660 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_logits/chosen": -1.6484375, | |
| "eval_logits/rejected": -1.6640625, | |
| "eval_logps/chosen": -219.0, | |
| "eval_logps/rejected": -228.0, | |
| "eval_loss": 0.7706417441368103, | |
| "eval_rewards/accuracies": 0.6397058963775635, | |
| "eval_rewards/chosen": -4.15625, | |
| "eval_rewards/margins": 1.03125, | |
| "eval_rewards/rejected": -5.1875, | |
| "eval_runtime": 23.2393, | |
| "eval_samples_per_second": 17.255, | |
| "eval_steps_per_second": 0.732, | |
| "step": 1668 | |
| }, | |
| { | |
| "epoch": 2.002398081534772, | |
| "grad_norm": 9.573444683512719, | |
| "learning_rate": 1.8480675255442025e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.1096, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -3.6875, | |
| "rewards/margins": 3.40625, | |
| "rewards/rejected": -7.09375, | |
| "step": 1670 | |
| }, | |
| { | |
| "epoch": 2.014388489208633, | |
| "grad_norm": 13.805204899223623, | |
| "learning_rate": 1.8258551754775655e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -236.0, | |
| "loss": 0.0794, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.015625, | |
| "rewards/margins": 3.375, | |
| "rewards/rejected": -6.40625, | |
| "step": 1680 | |
| }, | |
| { | |
| "epoch": 2.026378896882494, | |
| "grad_norm": 5.467275936133308, | |
| "learning_rate": 1.8036428254109286e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.5625, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0638, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.71875, | |
| "rewards/margins": 3.8125, | |
| "rewards/rejected": -7.53125, | |
| "step": 1690 | |
| }, | |
| { | |
| "epoch": 2.038369304556355, | |
| "grad_norm": 10.769684058873244, | |
| "learning_rate": 1.7814304753442913e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -215.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0615, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.1875, | |
| "rewards/margins": 3.875, | |
| "rewards/rejected": -7.09375, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 2.050359712230216, | |
| "grad_norm": 5.977598972558003, | |
| "learning_rate": 1.759218125277654e-07, | |
| "logits/chosen": -1.59375, | |
| "logits/rejected": -1.578125, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -251.0, | |
| "loss": 0.0588, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.703125, | |
| "rewards/margins": 3.5625, | |
| "rewards/rejected": -7.25, | |
| "step": 1710 | |
| }, | |
| { | |
| "epoch": 2.062350119904077, | |
| "grad_norm": 6.11324671011277, | |
| "learning_rate": 1.7370057752110174e-07, | |
| "logits/chosen": -1.5859375, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -221.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0604, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.546875, | |
| "rewards/margins": 4.125, | |
| "rewards/rejected": -7.65625, | |
| "step": 1720 | |
| }, | |
| { | |
| "epoch": 2.0743405275779376, | |
| "grad_norm": 7.0419948282884555, | |
| "learning_rate": 1.7147934251443802e-07, | |
| "logits/chosen": -1.703125, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -218.0, | |
| "logps/rejected": -250.0, | |
| "loss": 0.0604, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.53125, | |
| "rewards/margins": 3.90625, | |
| "rewards/rejected": -7.4375, | |
| "step": 1730 | |
| }, | |
| { | |
| "epoch": 2.0863309352517985, | |
| "grad_norm": 13.536498874343664, | |
| "learning_rate": 1.6925810750777432e-07, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -234.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0559, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.890625, | |
| "rewards/margins": 4.09375, | |
| "rewards/rejected": -8.0, | |
| "step": 1740 | |
| }, | |
| { | |
| "epoch": 2.0983213429256593, | |
| "grad_norm": 7.857771486014022, | |
| "learning_rate": 1.6703687250111062e-07, | |
| "logits/chosen": -1.6015625, | |
| "logits/rejected": -1.5703125, | |
| "logps/chosen": -228.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0637, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 3.84375, | |
| "rewards/rejected": -8.375, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 2.1103117505995206, | |
| "grad_norm": 24.029270438668547, | |
| "learning_rate": 1.648156374944469e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -227.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.0691, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 4.28125, | |
| "rewards/rejected": -8.4375, | |
| "step": 1760 | |
| }, | |
| { | |
| "epoch": 2.1223021582733814, | |
| "grad_norm": 9.575862100985665, | |
| "learning_rate": 1.625944024877832e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.5703125, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -242.0, | |
| "loss": 0.0761, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.734375, | |
| "rewards/margins": 3.609375, | |
| "rewards/rejected": -7.34375, | |
| "step": 1770 | |
| }, | |
| { | |
| "epoch": 2.1342925659472423, | |
| "grad_norm": 8.445962784666682, | |
| "learning_rate": 1.603731674811195e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -229.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0666, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.921875, | |
| "rewards/margins": 3.953125, | |
| "rewards/rejected": -7.875, | |
| "step": 1780 | |
| }, | |
| { | |
| "epoch": 2.146282973621103, | |
| "grad_norm": 12.5096726751056, | |
| "learning_rate": 1.581519324744558e-07, | |
| "logits/chosen": -1.703125, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -229.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.0554, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.515625, | |
| "rewards/margins": 3.875, | |
| "rewards/rejected": -7.375, | |
| "step": 1790 | |
| }, | |
| { | |
| "epoch": 2.158273381294964, | |
| "grad_norm": 7.424474262544135, | |
| "learning_rate": 1.5593069746779208e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -250.0, | |
| "loss": 0.0665, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.15625, | |
| "rewards/margins": 3.796875, | |
| "rewards/rejected": -7.9375, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.1702637889688248, | |
| "grad_norm": 7.886336061164498, | |
| "learning_rate": 1.5370946246112836e-07, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -212.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.0565, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.25, | |
| "rewards/margins": 4.34375, | |
| "rewards/rejected": -8.5625, | |
| "step": 1810 | |
| }, | |
| { | |
| "epoch": 2.182254196642686, | |
| "grad_norm": 5.804381015966981, | |
| "learning_rate": 1.514882274544647e-07, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0475, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.03125, | |
| "rewards/margins": 4.5625, | |
| "rewards/rejected": -8.625, | |
| "step": 1820 | |
| }, | |
| { | |
| "epoch": 2.194244604316547, | |
| "grad_norm": 8.523559366958416, | |
| "learning_rate": 1.4926699244780097e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -215.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0732, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.828125, | |
| "rewards/margins": 4.21875, | |
| "rewards/rejected": -8.0625, | |
| "step": 1830 | |
| }, | |
| { | |
| "epoch": 2.2062350119904077, | |
| "grad_norm": 15.092361293159154, | |
| "learning_rate": 1.4704575744113727e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -227.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0544, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.25, | |
| "rewards/margins": 4.125, | |
| "rewards/rejected": -8.375, | |
| "step": 1840 | |
| }, | |
| { | |
| "epoch": 2.2182254196642686, | |
| "grad_norm": 7.809268443424338, | |
| "learning_rate": 1.4482452243447357e-07, | |
| "logits/chosen": -1.6171875, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0463, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.0, | |
| "rewards/margins": 4.625, | |
| "rewards/rejected": -8.625, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 2.2302158273381294, | |
| "grad_norm": 6.175474139623565, | |
| "learning_rate": 1.4260328742780985e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.59375, | |
| "logps/chosen": -210.0, | |
| "logps/rejected": -251.0, | |
| "loss": 0.0524, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.8125, | |
| "rewards/margins": 4.40625, | |
| "rewards/rejected": -8.1875, | |
| "step": 1860 | |
| }, | |
| { | |
| "epoch": 2.2422062350119902, | |
| "grad_norm": 8.016024927468784, | |
| "learning_rate": 1.4038205242114615e-07, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.5703125, | |
| "logps/chosen": -202.0, | |
| "logps/rejected": -250.0, | |
| "loss": 0.0476, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.671875, | |
| "rewards/margins": 4.21875, | |
| "rewards/rejected": -7.875, | |
| "step": 1870 | |
| }, | |
| { | |
| "epoch": 2.2541966426858515, | |
| "grad_norm": 46.585404251146365, | |
| "learning_rate": 1.3816081741448243e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -243.0, | |
| "loss": 0.0878, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": -3.703125, | |
| "rewards/margins": 4.125, | |
| "rewards/rejected": -7.84375, | |
| "step": 1880 | |
| }, | |
| { | |
| "epoch": 2.2661870503597124, | |
| "grad_norm": 14.496947844310064, | |
| "learning_rate": 1.3593958240781876e-07, | |
| "logits/chosen": -1.75, | |
| "logits/rejected": -1.71875, | |
| "logps/chosen": -244.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.0442, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.953125, | |
| "rewards/margins": 4.59375, | |
| "rewards/rejected": -8.5625, | |
| "step": 1890 | |
| }, | |
| { | |
| "epoch": 2.278177458033573, | |
| "grad_norm": 10.300232011520219, | |
| "learning_rate": 1.3371834740115504e-07, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -286.0, | |
| "loss": 0.0516, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.34375, | |
| "rewards/margins": 4.3125, | |
| "rewards/rejected": -8.625, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 2.290167865707434, | |
| "grad_norm": 8.877528274051476, | |
| "learning_rate": 1.3149711239449131e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -225.0, | |
| "logps/rejected": -276.0, | |
| "loss": 0.0473, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.71875, | |
| "rewards/margins": 4.375, | |
| "rewards/rejected": -9.125, | |
| "step": 1910 | |
| }, | |
| { | |
| "epoch": 2.302158273381295, | |
| "grad_norm": 8.315723064599236, | |
| "learning_rate": 1.2927587738782764e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -218.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0593, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.09375, | |
| "rewards/margins": 4.28125, | |
| "rewards/rejected": -8.375, | |
| "step": 1920 | |
| }, | |
| { | |
| "epoch": 2.3141486810551557, | |
| "grad_norm": 13.642713487818822, | |
| "learning_rate": 1.2705464238116392e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -209.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0621, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.09375, | |
| "rewards/margins": 4.59375, | |
| "rewards/rejected": -8.6875, | |
| "step": 1930 | |
| }, | |
| { | |
| "epoch": 2.326139088729017, | |
| "grad_norm": 13.131558338647796, | |
| "learning_rate": 1.2483340737450022e-07, | |
| "logits/chosen": -1.65625, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0602, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -3.84375, | |
| "rewards/margins": 4.28125, | |
| "rewards/rejected": -8.125, | |
| "step": 1940 | |
| }, | |
| { | |
| "epoch": 2.338129496402878, | |
| "grad_norm": 10.187009587459894, | |
| "learning_rate": 1.226121723678365e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -206.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.05, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.671875, | |
| "rewards/margins": 4.625, | |
| "rewards/rejected": -8.3125, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 2.3501199040767387, | |
| "grad_norm": 7.106372818158467, | |
| "learning_rate": 1.203909373611728e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.6015625, | |
| "logps/chosen": -213.0, | |
| "logps/rejected": -249.0, | |
| "loss": 0.0516, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.03125, | |
| "rewards/margins": 4.21875, | |
| "rewards/rejected": -8.25, | |
| "step": 1960 | |
| }, | |
| { | |
| "epoch": 2.3621103117505995, | |
| "grad_norm": 6.0024755856013305, | |
| "learning_rate": 1.1816970235450909e-07, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.609375, | |
| "logps/chosen": -224.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.045, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.1875, | |
| "rewards/margins": 4.75, | |
| "rewards/rejected": -8.9375, | |
| "step": 1970 | |
| }, | |
| { | |
| "epoch": 2.3741007194244603, | |
| "grad_norm": 7.689477505076036, | |
| "learning_rate": 1.159484673478454e-07, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -226.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.043, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.0625, | |
| "rewards/margins": 4.40625, | |
| "rewards/rejected": -8.4375, | |
| "step": 1980 | |
| }, | |
| { | |
| "epoch": 2.386091127098321, | |
| "grad_norm": 7.812529725908209, | |
| "learning_rate": 1.1372723234118169e-07, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0501, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.25, | |
| "rewards/margins": 4.125, | |
| "rewards/rejected": -8.375, | |
| "step": 1990 | |
| }, | |
| { | |
| "epoch": 2.3980815347721824, | |
| "grad_norm": 49.4364201724221, | |
| "learning_rate": 1.1150599733451799e-07, | |
| "logits/chosen": -1.7265625, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -232.0, | |
| "logps/rejected": -298.0, | |
| "loss": 0.0593, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.1875, | |
| "rewards/margins": 4.625, | |
| "rewards/rejected": -8.8125, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.4100719424460433, | |
| "grad_norm": 4.317275529656694, | |
| "learning_rate": 1.0928476232785429e-07, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0362, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 4.78125, | |
| "rewards/rejected": -8.875, | |
| "step": 2010 | |
| }, | |
| { | |
| "epoch": 2.422062350119904, | |
| "grad_norm": 7.089907087105454, | |
| "learning_rate": 1.0706352732119057e-07, | |
| "logits/chosen": -1.71875, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.0413, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -3.671875, | |
| "rewards/margins": 4.4375, | |
| "rewards/rejected": -8.125, | |
| "step": 2020 | |
| }, | |
| { | |
| "epoch": 2.434052757793765, | |
| "grad_norm": 10.12450913020343, | |
| "learning_rate": 1.0484229231452687e-07, | |
| "logits/chosen": -1.7265625, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.0598, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 4.375, | |
| "rewards/rejected": -9.0625, | |
| "step": 2030 | |
| }, | |
| { | |
| "epoch": 2.446043165467626, | |
| "grad_norm": 10.848851517796744, | |
| "learning_rate": 1.0262105730786316e-07, | |
| "logits/chosen": -1.703125, | |
| "logits/rejected": -1.7265625, | |
| "logps/chosen": -213.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0475, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5, | |
| "rewards/margins": 4.1875, | |
| "rewards/rejected": -8.6875, | |
| "step": 2040 | |
| }, | |
| { | |
| "epoch": 2.4580335731414866, | |
| "grad_norm": 3.5169019202687584, | |
| "learning_rate": 1.0039982230119947e-07, | |
| "logits/chosen": -1.6953125, | |
| "logits/rejected": -1.703125, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0752, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": -4.875, | |
| "rewards/margins": 4.34375, | |
| "rewards/rejected": -9.1875, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 2.470023980815348, | |
| "grad_norm": 11.726493677946326, | |
| "learning_rate": 9.817858729453577e-08, | |
| "logits/chosen": -1.7109375, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.054, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.21875, | |
| "rewards/margins": 4.625, | |
| "rewards/rejected": -8.875, | |
| "step": 2060 | |
| }, | |
| { | |
| "epoch": 2.4820143884892087, | |
| "grad_norm": 10.3815257923276, | |
| "learning_rate": 9.595735228787204e-08, | |
| "logits/chosen": -1.6953125, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -221.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0545, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 4.21875, | |
| "rewards/rejected": -8.875, | |
| "step": 2070 | |
| }, | |
| { | |
| "epoch": 2.4940047961630696, | |
| "grad_norm": 5.520164218604047, | |
| "learning_rate": 9.373611728120835e-08, | |
| "logits/chosen": -1.7109375, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.0396, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.28125, | |
| "rewards/margins": 4.09375, | |
| "rewards/rejected": -8.375, | |
| "step": 2080 | |
| }, | |
| { | |
| "epoch": 2.5059952038369304, | |
| "grad_norm": 8.376756986846386, | |
| "learning_rate": 9.151488227454464e-08, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -208.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0427, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.0, | |
| "rewards/margins": 4.75, | |
| "rewards/rejected": -8.75, | |
| "step": 2090 | |
| }, | |
| { | |
| "epoch": 2.5179856115107913, | |
| "grad_norm": 20.20395693776781, | |
| "learning_rate": 8.929364726788094e-08, | |
| "logits/chosen": -1.734375, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -226.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.0661, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 4.5, | |
| "rewards/rejected": -9.1875, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.5299760191846525, | |
| "grad_norm": 30.373447856641558, | |
| "learning_rate": 8.707241226121723e-08, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.703125, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -284.0, | |
| "loss": 0.0453, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 4.59375, | |
| "rewards/rejected": -9.3125, | |
| "step": 2110 | |
| }, | |
| { | |
| "epoch": 2.5419664268585134, | |
| "grad_norm": 3.057799168034458, | |
| "learning_rate": 8.485117725455352e-08, | |
| "logits/chosen": -1.7109375, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -241.0, | |
| "logps/rejected": -288.0, | |
| "loss": 0.0326, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.84375, | |
| "rewards/margins": 4.6875, | |
| "rewards/rejected": -9.5, | |
| "step": 2120 | |
| }, | |
| { | |
| "epoch": 2.553956834532374, | |
| "grad_norm": 8.47051827415506, | |
| "learning_rate": 8.262994224788982e-08, | |
| "logits/chosen": -1.734375, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -282.0, | |
| "loss": 0.0523, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -5.09375, | |
| "rewards/margins": 4.25, | |
| "rewards/rejected": -9.3125, | |
| "step": 2130 | |
| }, | |
| { | |
| "epoch": 2.565947242206235, | |
| "grad_norm": 3.8022915967693662, | |
| "learning_rate": 8.040870724122611e-08, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.6640625, | |
| "logps/chosen": -202.0, | |
| "logps/rejected": -256.0, | |
| "loss": 0.0353, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -3.953125, | |
| "rewards/margins": 4.625, | |
| "rewards/rejected": -8.5625, | |
| "step": 2140 | |
| }, | |
| { | |
| "epoch": 2.577937649880096, | |
| "grad_norm": 22.652691529528166, | |
| "learning_rate": 7.818747223456242e-08, | |
| "logits/chosen": -1.703125, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -231.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.0421, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 4.84375, | |
| "rewards/rejected": -9.4375, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 2.5899280575539567, | |
| "grad_norm": 5.621319686196611, | |
| "learning_rate": 7.596623722789871e-08, | |
| "logits/chosen": -1.6328125, | |
| "logits/rejected": -1.625, | |
| "logps/chosen": -189.0, | |
| "logps/rejected": -243.0, | |
| "loss": 0.0383, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 4.625, | |
| "rewards/rejected": -8.75, | |
| "step": 2160 | |
| }, | |
| { | |
| "epoch": 2.6019184652278176, | |
| "grad_norm": 9.164491714974064, | |
| "learning_rate": 7.3745002221235e-08, | |
| "logits/chosen": -1.75, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -242.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0414, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.53125, | |
| "rewards/margins": 4.53125, | |
| "rewards/rejected": -9.0625, | |
| "step": 2170 | |
| }, | |
| { | |
| "epoch": 2.6139088729016784, | |
| "grad_norm": 5.372990484400878, | |
| "learning_rate": 7.15237672145713e-08, | |
| "logits/chosen": -1.71875, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -212.0, | |
| "logps/rejected": -258.0, | |
| "loss": 0.0366, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.09375, | |
| "rewards/margins": 5.3125, | |
| "rewards/rejected": -9.375, | |
| "step": 2180 | |
| }, | |
| { | |
| "epoch": 2.6258992805755397, | |
| "grad_norm": 10.746900341332498, | |
| "learning_rate": 6.930253220790759e-08, | |
| "logits/chosen": -1.75, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.0452, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 4.4375, | |
| "rewards/rejected": -9.125, | |
| "step": 2190 | |
| }, | |
| { | |
| "epoch": 2.6378896882494005, | |
| "grad_norm": 11.993659380624129, | |
| "learning_rate": 6.70812972012439e-08, | |
| "logits/chosen": -1.7421875, | |
| "logits/rejected": -1.703125, | |
| "logps/chosen": -212.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0386, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.40625, | |
| "rewards/margins": 4.9375, | |
| "rewards/rejected": -9.3125, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.6498800959232613, | |
| "grad_norm": 10.522371499574833, | |
| "learning_rate": 6.486006219458018e-08, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -216.0, | |
| "logps/rejected": -255.0, | |
| "loss": 0.0462, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.4375, | |
| "rewards/margins": 4.625, | |
| "rewards/rejected": -9.0625, | |
| "step": 2210 | |
| }, | |
| { | |
| "epoch": 2.661870503597122, | |
| "grad_norm": 12.225253766846958, | |
| "learning_rate": 6.263882718791647e-08, | |
| "logits/chosen": -1.6484375, | |
| "logits/rejected": -1.6171875, | |
| "logps/chosen": -228.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0522, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.125, | |
| "rewards/margins": 4.84375, | |
| "rewards/rejected": -8.9375, | |
| "step": 2220 | |
| }, | |
| { | |
| "epoch": 2.6738609112709835, | |
| "grad_norm": 12.629562732845173, | |
| "learning_rate": 6.041759218125278e-08, | |
| "logits/chosen": -1.671875, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -221.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0446, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.46875, | |
| "rewards/margins": 4.40625, | |
| "rewards/rejected": -8.875, | |
| "step": 2230 | |
| }, | |
| { | |
| "epoch": 2.6858513189448443, | |
| "grad_norm": 8.807729780911718, | |
| "learning_rate": 5.8196357174589066e-08, | |
| "logits/chosen": -1.734375, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -241.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0384, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.59375, | |
| "rewards/margins": 4.625, | |
| "rewards/rejected": -9.25, | |
| "step": 2240 | |
| }, | |
| { | |
| "epoch": 2.697841726618705, | |
| "grad_norm": 2.884098452133717, | |
| "learning_rate": 5.597512216792536e-08, | |
| "logits/chosen": -1.734375, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -224.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.036, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.09375, | |
| "rewards/margins": 4.40625, | |
| "rewards/rejected": -8.5, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 2.709832134292566, | |
| "grad_norm": 7.527525000207272, | |
| "learning_rate": 5.375388716126166e-08, | |
| "logits/chosen": -1.703125, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -212.0, | |
| "logps/rejected": -249.0, | |
| "loss": 0.0436, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.46875, | |
| "rewards/margins": 4.65625, | |
| "rewards/rejected": -9.125, | |
| "step": 2260 | |
| }, | |
| { | |
| "epoch": 2.721822541966427, | |
| "grad_norm": 22.77210894040614, | |
| "learning_rate": 5.1532652154597956e-08, | |
| "logits/chosen": -1.703125, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -236.0, | |
| "logps/rejected": -294.0, | |
| "loss": 0.0441, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.3125, | |
| "rewards/margins": 5.21875, | |
| "rewards/rejected": -10.5625, | |
| "step": 2270 | |
| }, | |
| { | |
| "epoch": 2.7338129496402876, | |
| "grad_norm": 7.128969603906596, | |
| "learning_rate": 4.9311417147934246e-08, | |
| "logits/chosen": -1.703125, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -222.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0474, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.625, | |
| "rewards/margins": 4.46875, | |
| "rewards/rejected": -9.0625, | |
| "step": 2280 | |
| }, | |
| { | |
| "epoch": 2.7458033573141485, | |
| "grad_norm": 6.150942157607421, | |
| "learning_rate": 4.709018214127054e-08, | |
| "logits/chosen": -1.78125, | |
| "logits/rejected": -1.71875, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0352, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.5625, | |
| "rewards/margins": 4.78125, | |
| "rewards/rejected": -9.375, | |
| "step": 2290 | |
| }, | |
| { | |
| "epoch": 2.7577937649880093, | |
| "grad_norm": 2.8164036629567843, | |
| "learning_rate": 4.486894713460684e-08, | |
| "logits/chosen": -1.7421875, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -222.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.0347, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 4.65625, | |
| "rewards/rejected": -9.375, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.7697841726618706, | |
| "grad_norm": 7.133912331859571, | |
| "learning_rate": 4.2647712127943136e-08, | |
| "logits/chosen": -1.765625, | |
| "logits/rejected": -1.7578125, | |
| "logps/chosen": -239.0, | |
| "logps/rejected": -304.0, | |
| "loss": 0.0356, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.75, | |
| "rewards/margins": 4.90625, | |
| "rewards/rejected": -9.6875, | |
| "step": 2310 | |
| }, | |
| { | |
| "epoch": 2.7817745803357314, | |
| "grad_norm": 5.416214434119401, | |
| "learning_rate": 4.042647712127943e-08, | |
| "logits/chosen": -1.71875, | |
| "logits/rejected": -1.6953125, | |
| "logps/chosen": -230.0, | |
| "logps/rejected": -278.0, | |
| "loss": 0.0358, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -5.625, | |
| "rewards/margins": 4.5625, | |
| "rewards/rejected": -10.1875, | |
| "step": 2320 | |
| }, | |
| { | |
| "epoch": 2.7937649880095923, | |
| "grad_norm": 7.05206840398353, | |
| "learning_rate": 3.820524211461572e-08, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -214.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.0468, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -5.1875, | |
| "rewards/margins": 4.40625, | |
| "rewards/rejected": -9.5625, | |
| "step": 2330 | |
| }, | |
| { | |
| "epoch": 2.805755395683453, | |
| "grad_norm": 6.8237218283267245, | |
| "learning_rate": 3.598400710795202e-08, | |
| "logits/chosen": -1.7109375, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -239.0, | |
| "logps/rejected": -296.0, | |
| "loss": 0.0383, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.9375, | |
| "rewards/margins": 5.25, | |
| "rewards/rejected": -10.1875, | |
| "step": 2340 | |
| }, | |
| { | |
| "epoch": 2.8177458033573144, | |
| "grad_norm": 17.473074202115587, | |
| "learning_rate": 3.3762772101288315e-08, | |
| "logits/chosen": -1.640625, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -218.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0388, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.84375, | |
| "rewards/margins": 4.4375, | |
| "rewards/rejected": -9.25, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 2.8297362110311752, | |
| "grad_norm": 6.423408193266802, | |
| "learning_rate": 3.1541537094624605e-08, | |
| "logits/chosen": -1.734375, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -233.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.0436, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.78125, | |
| "rewards/margins": 4.84375, | |
| "rewards/rejected": -9.625, | |
| "step": 2360 | |
| }, | |
| { | |
| "epoch": 2.841726618705036, | |
| "grad_norm": 9.316314943177023, | |
| "learning_rate": 2.9320302087960905e-08, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -223.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0494, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -5.25, | |
| "rewards/margins": 4.875, | |
| "rewards/rejected": -10.125, | |
| "step": 2370 | |
| }, | |
| { | |
| "epoch": 2.853717026378897, | |
| "grad_norm": 4.563487715111162, | |
| "learning_rate": 2.70990670812972e-08, | |
| "logits/chosen": -1.6796875, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -215.0, | |
| "logps/rejected": -264.0, | |
| "loss": 0.0411, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -5.5625, | |
| "rewards/margins": 4.71875, | |
| "rewards/rejected": -10.25, | |
| "step": 2380 | |
| }, | |
| { | |
| "epoch": 2.8657074340527577, | |
| "grad_norm": 7.265458467508151, | |
| "learning_rate": 2.4877832074633495e-08, | |
| "logits/chosen": -1.734375, | |
| "logits/rejected": -1.6328125, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -266.0, | |
| "loss": 0.035, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.96875, | |
| "rewards/margins": 4.875, | |
| "rewards/rejected": -9.8125, | |
| "step": 2390 | |
| }, | |
| { | |
| "epoch": 2.8776978417266186, | |
| "grad_norm": 3.9055052792545744, | |
| "learning_rate": 2.2656597067969788e-08, | |
| "logits/chosen": -1.6875, | |
| "logits/rejected": -1.7109375, | |
| "logps/chosen": -217.0, | |
| "logps/rejected": -272.0, | |
| "loss": 0.0387, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.46875, | |
| "rewards/margins": 5.0, | |
| "rewards/rejected": -9.5, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.8896882494004794, | |
| "grad_norm": 3.850000762426418, | |
| "learning_rate": 2.0435362061306088e-08, | |
| "logits/chosen": -1.703125, | |
| "logits/rejected": -1.6875, | |
| "logps/chosen": -222.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0553, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.1875, | |
| "rewards/margins": 4.8125, | |
| "rewards/rejected": -10.0, | |
| "step": 2410 | |
| }, | |
| { | |
| "epoch": 2.9016786570743403, | |
| "grad_norm": 3.8530749601946486, | |
| "learning_rate": 1.821412705464238e-08, | |
| "logits/chosen": -1.765625, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -225.0, | |
| "logps/rejected": -300.0, | |
| "loss": 0.033, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.4375, | |
| "rewards/margins": 5.25, | |
| "rewards/rejected": -10.6875, | |
| "step": 2420 | |
| }, | |
| { | |
| "epoch": 2.9136690647482015, | |
| "grad_norm": 5.640559221099878, | |
| "learning_rate": 1.5992892047978674e-08, | |
| "logits/chosen": -1.6953125, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -226.0, | |
| "logps/rejected": -280.0, | |
| "loss": 0.0343, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.0, | |
| "rewards/margins": 4.71875, | |
| "rewards/rejected": -9.75, | |
| "step": 2430 | |
| }, | |
| { | |
| "epoch": 2.9256594724220624, | |
| "grad_norm": 4.969640872073211, | |
| "learning_rate": 1.377165704131497e-08, | |
| "logits/chosen": -1.6640625, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -218.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.038, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.65625, | |
| "rewards/margins": 5.0625, | |
| "rewards/rejected": -9.6875, | |
| "step": 2440 | |
| }, | |
| { | |
| "epoch": 2.937649880095923, | |
| "grad_norm": 5.866247247572104, | |
| "learning_rate": 1.1550422034651266e-08, | |
| "logits/chosen": -1.75, | |
| "logits/rejected": -1.6484375, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -270.0, | |
| "loss": 0.039, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.78125, | |
| "rewards/margins": 4.78125, | |
| "rewards/rejected": -9.5625, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 2.949640287769784, | |
| "grad_norm": 4.3423386310633365, | |
| "learning_rate": 9.32918702798756e-09, | |
| "logits/chosen": -1.71875, | |
| "logits/rejected": -1.65625, | |
| "logps/chosen": -220.0, | |
| "logps/rejected": -262.0, | |
| "loss": 0.0506, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -5.1875, | |
| "rewards/margins": 4.21875, | |
| "rewards/rejected": -9.4375, | |
| "step": 2460 | |
| }, | |
| { | |
| "epoch": 2.9616306954436453, | |
| "grad_norm": 7.560339052869031, | |
| "learning_rate": 7.107952021323855e-09, | |
| "logits/chosen": -1.7265625, | |
| "logits/rejected": -1.6796875, | |
| "logps/chosen": -219.0, | |
| "logps/rejected": -268.0, | |
| "loss": 0.0335, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.96875, | |
| "rewards/margins": 5.09375, | |
| "rewards/rejected": -10.0625, | |
| "step": 2470 | |
| }, | |
| { | |
| "epoch": 2.973621103117506, | |
| "grad_norm": 10.814073549670898, | |
| "learning_rate": 4.886717014660151e-09, | |
| "logits/chosen": -1.75, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -232.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0393, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": -4.875, | |
| "rewards/margins": 4.78125, | |
| "rewards/rejected": -9.6875, | |
| "step": 2480 | |
| }, | |
| { | |
| "epoch": 2.985611510791367, | |
| "grad_norm": 4.9409390032818665, | |
| "learning_rate": 2.665482007996446e-09, | |
| "logits/chosen": -1.7265625, | |
| "logits/rejected": -1.671875, | |
| "logps/chosen": -228.0, | |
| "logps/rejected": -260.0, | |
| "loss": 0.0378, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": -4.96875, | |
| "rewards/margins": 4.28125, | |
| "rewards/rejected": -9.25, | |
| "step": 2490 | |
| }, | |
| { | |
| "epoch": 2.997601918465228, | |
| "grad_norm": 19.048106847839737, | |
| "learning_rate": 4.4424700133274096e-10, | |
| "logits/chosen": -1.71875, | |
| "logits/rejected": -1.640625, | |
| "logps/chosen": -238.0, | |
| "logps/rejected": -274.0, | |
| "loss": 0.0398, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -4.6875, | |
| "rewards/margins": 4.9375, | |
| "rewards/rejected": -9.625, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_logits/chosen": -1.71875, | |
| "eval_logits/rejected": -1.7421875, | |
| "eval_logps/chosen": -240.0, | |
| "eval_logps/rejected": -250.0, | |
| "eval_loss": 0.9926745891571045, | |
| "eval_rewards/accuracies": 0.6029411554336548, | |
| "eval_rewards/chosen": -6.21875, | |
| "eval_rewards/margins": 1.1953125, | |
| "eval_rewards/rejected": -7.4375, | |
| "eval_runtime": 28.9311, | |
| "eval_samples_per_second": 13.861, | |
| "eval_steps_per_second": 0.588, | |
| "step": 2502 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 2502, | |
| "total_flos": 0.0, | |
| "train_loss": 0.2617608135004791, | |
| "train_runtime": 8206.5437, | |
| "train_samples_per_second": 7.311, | |
| "train_steps_per_second": 0.305 | |
| } | |
| ], | |
| "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": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |