| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 100, | |
| "global_step": 478, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0020920502092050207, | |
| "grad_norm": 6.6646714922569, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": -0.69921875, | |
| "logits/rejected": -0.828125, | |
| "logps/chosen": -326.0, | |
| "logps/rejected": -376.0, | |
| "loss": 1.3828, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02092050209205021, | |
| "grad_norm": 6.489408950256911, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -0.6328125, | |
| "logits/rejected": -0.56640625, | |
| "logps/chosen": -272.0, | |
| "logps/rejected": -260.0, | |
| "loss": 1.3839, | |
| "rewards/accuracies": 0.1736111044883728, | |
| "rewards/chosen": -0.000797271728515625, | |
| "rewards/margins": -0.000865936279296875, | |
| "rewards/rejected": 7.43865966796875e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04184100418410042, | |
| "grad_norm": 5.9529855573917985, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -0.5390625, | |
| "logits/rejected": -0.5625, | |
| "logps/chosen": -296.0, | |
| "logps/rejected": -304.0, | |
| "loss": 1.384, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": 0.00148773193359375, | |
| "rewards/margins": 0.00162506103515625, | |
| "rewards/rejected": -0.0001373291015625, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06276150627615062, | |
| "grad_norm": 6.276450797610096, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -0.59375, | |
| "logits/rejected": -0.625, | |
| "logps/chosen": -314.0, | |
| "logps/rejected": -276.0, | |
| "loss": 1.3838, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": -0.00025177001953125, | |
| "rewards/margins": 0.001220703125, | |
| "rewards/rejected": -0.00147247314453125, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08368200836820083, | |
| "grad_norm": 6.188778447881269, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -0.5078125, | |
| "logits/rejected": -0.51171875, | |
| "logps/chosen": -270.0, | |
| "logps/rejected": -266.0, | |
| "loss": 1.3807, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": -0.003875732421875, | |
| "rewards/margins": 0.004547119140625, | |
| "rewards/rejected": -0.0084228515625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10460251046025104, | |
| "grad_norm": 6.297837269064405, | |
| "learning_rate": 4.999733114418725e-07, | |
| "logits/chosen": -0.5703125, | |
| "logits/rejected": -0.53515625, | |
| "logps/chosen": -318.0, | |
| "logps/rejected": -312.0, | |
| "loss": 1.3732, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.0089111328125, | |
| "rewards/margins": 0.010009765625, | |
| "rewards/rejected": -0.0189208984375, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12552301255230125, | |
| "grad_norm": 6.372919655819386, | |
| "learning_rate": 4.990398100856366e-07, | |
| "logits/chosen": -0.6328125, | |
| "logits/rejected": -0.6484375, | |
| "logps/chosen": -284.0, | |
| "logps/rejected": -314.0, | |
| "loss": 1.3635, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.0140380859375, | |
| "rewards/margins": 0.0238037109375, | |
| "rewards/rejected": -0.037841796875, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14644351464435146, | |
| "grad_norm": 5.9641189836535125, | |
| "learning_rate": 4.967775735898179e-07, | |
| "logits/chosen": -0.515625, | |
| "logits/rejected": -0.55078125, | |
| "logps/chosen": -284.0, | |
| "logps/rejected": -282.0, | |
| "loss": 1.3517, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.023681640625, | |
| "rewards/margins": 0.037353515625, | |
| "rewards/rejected": -0.06103515625, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16736401673640167, | |
| "grad_norm": 6.861806432649946, | |
| "learning_rate": 4.931986719649298e-07, | |
| "logits/chosen": -0.67578125, | |
| "logits/rejected": -0.6796875, | |
| "logps/chosen": -346.0, | |
| "logps/rejected": -286.0, | |
| "loss": 1.3391, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.0419921875, | |
| "rewards/margins": 0.0537109375, | |
| "rewards/rejected": -0.095703125, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18828451882845187, | |
| "grad_norm": 6.96525100509858, | |
| "learning_rate": 4.883222001996351e-07, | |
| "logits/chosen": -0.64453125, | |
| "logits/rejected": -0.69140625, | |
| "logps/chosen": -282.0, | |
| "logps/rejected": -286.0, | |
| "loss": 1.3164, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.054443359375, | |
| "rewards/margins": 0.08642578125, | |
| "rewards/rejected": -0.140625, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20920502092050208, | |
| "grad_norm": 6.5113717274375915, | |
| "learning_rate": 4.821741763807186e-07, | |
| "logits/chosen": -0.7421875, | |
| "logits/rejected": -0.78515625, | |
| "logps/chosen": -288.0, | |
| "logps/rejected": -294.0, | |
| "loss": 1.3058, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.06689453125, | |
| "rewards/margins": 0.0947265625, | |
| "rewards/rejected": -0.1611328125, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.20920502092050208, | |
| "eval_logits/chosen": -0.71875, | |
| "eval_logits/rejected": -0.6875, | |
| "eval_logps/chosen": -296.0, | |
| "eval_logps/rejected": -310.0, | |
| "eval_loss": 0.6498281359672546, | |
| "eval_rewards/accuracies": 0.72265625, | |
| "eval_rewards/chosen": -0.05078125, | |
| "eval_rewards/margins": 0.1201171875, | |
| "eval_rewards/rejected": -0.1708984375, | |
| "eval_runtime": 45.0509, | |
| "eval_samples_per_second": 44.394, | |
| "eval_steps_per_second": 0.71, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2301255230125523, | |
| "grad_norm": 7.903584697993889, | |
| "learning_rate": 4.747874028753375e-07, | |
| "logits/chosen": -0.74609375, | |
| "logits/rejected": -0.765625, | |
| "logps/chosen": -340.0, | |
| "logps/rejected": -296.0, | |
| "loss": 1.2953, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.07958984375, | |
| "rewards/margins": 0.126953125, | |
| "rewards/rejected": -0.20703125, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.2510460251046025, | |
| "grad_norm": 9.884652757231866, | |
| "learning_rate": 4.662012913161997e-07, | |
| "logits/chosen": -0.72265625, | |
| "logits/rejected": -0.79296875, | |
| "logps/chosen": -320.0, | |
| "logps/rejected": -306.0, | |
| "loss": 1.2558, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.0869140625, | |
| "rewards/margins": 0.169921875, | |
| "rewards/rejected": -0.255859375, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2719665271966527, | |
| "grad_norm": 8.396514429492079, | |
| "learning_rate": 4.5646165232345103e-07, | |
| "logits/chosen": -0.76953125, | |
| "logits/rejected": -0.74609375, | |
| "logps/chosen": -304.0, | |
| "logps/rejected": -320.0, | |
| "loss": 1.2234, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.130859375, | |
| "rewards/margins": 0.2138671875, | |
| "rewards/rejected": -0.34375, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.2928870292887029, | |
| "grad_norm": 17.35957280226942, | |
| "learning_rate": 4.456204510851956e-07, | |
| "logits/chosen": -0.8046875, | |
| "logits/rejected": -0.796875, | |
| "logps/chosen": -346.0, | |
| "logps/rejected": -360.0, | |
| "loss": 1.2176, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.181640625, | |
| "rewards/margins": 0.267578125, | |
| "rewards/rejected": -0.44921875, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.3138075313807531, | |
| "grad_norm": 13.326784597880309, | |
| "learning_rate": 4.337355301007335e-07, | |
| "logits/chosen": -0.828125, | |
| "logits/rejected": -0.78515625, | |
| "logps/chosen": -332.0, | |
| "logps/rejected": -348.0, | |
| "loss": 1.2148, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.298828125, | |
| "rewards/margins": 0.19921875, | |
| "rewards/rejected": -0.498046875, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33472803347280333, | |
| "grad_norm": 12.974101567844148, | |
| "learning_rate": 4.2087030056579986e-07, | |
| "logits/chosen": -0.83984375, | |
| "logits/rejected": -0.8828125, | |
| "logps/chosen": -322.0, | |
| "logps/rejected": -328.0, | |
| "loss": 1.1781, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.328125, | |
| "rewards/margins": 0.337890625, | |
| "rewards/rejected": -0.6640625, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.35564853556485354, | |
| "grad_norm": 20.042683883224868, | |
| "learning_rate": 4.070934040463998e-07, | |
| "logits/chosen": -0.73046875, | |
| "logits/rejected": -0.69921875, | |
| "logps/chosen": -300.0, | |
| "logps/rejected": -310.0, | |
| "loss": 1.1769, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.33203125, | |
| "rewards/margins": 0.314453125, | |
| "rewards/rejected": -0.6484375, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.37656903765690375, | |
| "grad_norm": 17.632457136702087, | |
| "learning_rate": 3.9247834624635404e-07, | |
| "logits/chosen": -0.7890625, | |
| "logits/rejected": -0.8359375, | |
| "logps/chosen": -324.0, | |
| "logps/rejected": -332.0, | |
| "loss": 1.1354, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.5234375, | |
| "rewards/margins": 0.33984375, | |
| "rewards/rejected": -0.86328125, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.39748953974895396, | |
| "grad_norm": 18.261499900174815, | |
| "learning_rate": 3.7710310482256523e-07, | |
| "logits/chosen": -0.81640625, | |
| "logits/rejected": -0.77734375, | |
| "logps/chosen": -338.0, | |
| "logps/rejected": -374.0, | |
| "loss": 1.1528, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.39453125, | |
| "rewards/margins": 0.365234375, | |
| "rewards/rejected": -0.7578125, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.41841004184100417, | |
| "grad_norm": 36.22103873399267, | |
| "learning_rate": 3.610497133404795e-07, | |
| "logits/chosen": -0.72265625, | |
| "logits/rejected": -0.84375, | |
| "logps/chosen": -332.0, | |
| "logps/rejected": -360.0, | |
| "loss": 1.1479, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.51953125, | |
| "rewards/margins": 0.412109375, | |
| "rewards/rejected": -0.9296875, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.41841004184100417, | |
| "eval_logits/chosen": -0.8203125, | |
| "eval_logits/rejected": -0.7734375, | |
| "eval_logps/chosen": -336.0, | |
| "eval_logps/rejected": -388.0, | |
| "eval_loss": 0.5718047022819519, | |
| "eval_rewards/accuracies": 0.765625, | |
| "eval_rewards/chosen": -0.45703125, | |
| "eval_rewards/margins": 0.486328125, | |
| "eval_rewards/rejected": -0.9453125, | |
| "eval_runtime": 45.2, | |
| "eval_samples_per_second": 44.248, | |
| "eval_steps_per_second": 0.708, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4393305439330544, | |
| "grad_norm": 23.76956215603603, | |
| "learning_rate": 3.4440382358952115e-07, | |
| "logits/chosen": -0.84765625, | |
| "logits/rejected": -0.82421875, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -366.0, | |
| "loss": 1.1625, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.52734375, | |
| "rewards/margins": 0.419921875, | |
| "rewards/rejected": -0.9453125, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.4602510460251046, | |
| "grad_norm": 20.334823182577452, | |
| "learning_rate": 3.272542485937368e-07, | |
| "logits/chosen": -0.80859375, | |
| "logits/rejected": -0.87109375, | |
| "logps/chosen": -326.0, | |
| "logps/rejected": -366.0, | |
| "loss": 1.1391, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.5078125, | |
| "rewards/margins": 0.5078125, | |
| "rewards/rejected": -1.015625, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.4811715481171548, | |
| "grad_norm": 17.582669378077018, | |
| "learning_rate": 3.096924887558854e-07, | |
| "logits/chosen": -0.71875, | |
| "logits/rejected": -0.765625, | |
| "logps/chosen": -320.0, | |
| "logps/rejected": -362.0, | |
| "loss": 1.1636, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.60546875, | |
| "rewards/margins": 0.447265625, | |
| "rewards/rejected": -1.0546875, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.502092050209205, | |
| "grad_norm": 22.105123968919926, | |
| "learning_rate": 2.9181224366319943e-07, | |
| "logits/chosen": -0.8125, | |
| "logits/rejected": -0.70703125, | |
| "logps/chosen": -332.0, | |
| "logps/rejected": -376.0, | |
| "loss": 1.0841, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.51953125, | |
| "rewards/margins": 0.5390625, | |
| "rewards/rejected": -1.0625, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5230125523012552, | |
| "grad_norm": 21.132221087126197, | |
| "learning_rate": 2.7370891215954565e-07, | |
| "logits/chosen": -0.90234375, | |
| "logits/rejected": -0.796875, | |
| "logps/chosen": -376.0, | |
| "logps/rejected": -404.0, | |
| "loss": 1.1108, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.6484375, | |
| "rewards/margins": 0.6171875, | |
| "rewards/rejected": -1.265625, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5439330543933054, | |
| "grad_norm": 25.53588211930074, | |
| "learning_rate": 2.55479083351317e-07, | |
| "logits/chosen": -0.921875, | |
| "logits/rejected": -0.90234375, | |
| "logps/chosen": -388.0, | |
| "logps/rejected": -396.0, | |
| "loss": 1.1147, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.796875, | |
| "rewards/margins": 0.5078125, | |
| "rewards/rejected": -1.3046875, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5648535564853556, | |
| "grad_norm": 18.721587490226884, | |
| "learning_rate": 2.3722002126275822e-07, | |
| "logits/chosen": -0.80859375, | |
| "logits/rejected": -0.76171875, | |
| "logps/chosen": -376.0, | |
| "logps/rejected": -398.0, | |
| "loss": 1.1106, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.7421875, | |
| "rewards/margins": 0.470703125, | |
| "rewards/rejected": -1.2109375, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5857740585774058, | |
| "grad_norm": 22.57831660933063, | |
| "learning_rate": 2.19029145890313e-07, | |
| "logits/chosen": -0.75, | |
| "logits/rejected": -0.6953125, | |
| "logps/chosen": -348.0, | |
| "logps/rejected": -374.0, | |
| "loss": 1.1478, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.625, | |
| "rewards/margins": 0.490234375, | |
| "rewards/rejected": -1.109375, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.606694560669456, | |
| "grad_norm": 29.03098369938414, | |
| "learning_rate": 2.0100351342479216e-07, | |
| "logits/chosen": -0.90625, | |
| "logits/rejected": -0.83203125, | |
| "logps/chosen": -326.0, | |
| "logps/rejected": -374.0, | |
| "loss": 1.1153, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.6328125, | |
| "rewards/margins": 0.54296875, | |
| "rewards/rejected": -1.1796875, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6276150627615062, | |
| "grad_norm": 20.85597506299168, | |
| "learning_rate": 1.8323929841460178e-07, | |
| "logits/chosen": -0.88671875, | |
| "logits/rejected": -0.89453125, | |
| "logps/chosen": -400.0, | |
| "logps/rejected": -406.0, | |
| "loss": 1.0991, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.78515625, | |
| "rewards/margins": 0.484375, | |
| "rewards/rejected": -1.2734375, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6276150627615062, | |
| "eval_logits/chosen": -0.87109375, | |
| "eval_logits/rejected": -0.8203125, | |
| "eval_logps/chosen": -356.0, | |
| "eval_logps/rejected": -420.0, | |
| "eval_loss": 0.5499843955039978, | |
| "eval_rewards/accuracies": 0.76171875, | |
| "eval_rewards/chosen": -0.6640625, | |
| "eval_rewards/margins": 0.60546875, | |
| "eval_rewards/rejected": -1.2734375, | |
| "eval_runtime": 44.9585, | |
| "eval_samples_per_second": 44.486, | |
| "eval_steps_per_second": 0.712, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6485355648535565, | |
| "grad_norm": 22.089992862975382, | |
| "learning_rate": 1.6583128063291573e-07, | |
| "logits/chosen": -0.7421875, | |
| "logits/rejected": -0.79296875, | |
| "logps/chosen": -386.0, | |
| "logps/rejected": -400.0, | |
| "loss": 1.0856, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.796875, | |
| "rewards/margins": 0.5, | |
| "rewards/rejected": -1.296875, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6694560669456067, | |
| "grad_norm": 20.274184855936905, | |
| "learning_rate": 1.488723393865766e-07, | |
| "logits/chosen": -0.890625, | |
| "logits/rejected": -0.828125, | |
| "logps/chosen": -388.0, | |
| "logps/rejected": -402.0, | |
| "loss": 1.0528, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.796875, | |
| "rewards/margins": 0.58203125, | |
| "rewards/rejected": -1.375, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6903765690376569, | |
| "grad_norm": 21.63062129092967, | |
| "learning_rate": 1.3245295796480788e-07, | |
| "logits/chosen": -0.90234375, | |
| "logits/rejected": -0.875, | |
| "logps/chosen": -356.0, | |
| "logps/rejected": -400.0, | |
| "loss": 1.1076, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.703125, | |
| "rewards/margins": 0.435546875, | |
| "rewards/rejected": -1.1328125, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7112970711297071, | |
| "grad_norm": 24.015672748732257, | |
| "learning_rate": 1.1666074087171627e-07, | |
| "logits/chosen": -0.953125, | |
| "logits/rejected": -0.90234375, | |
| "logps/chosen": -384.0, | |
| "logps/rejected": -440.0, | |
| "loss": 1.0859, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.75390625, | |
| "rewards/margins": 0.6484375, | |
| "rewards/rejected": -1.3984375, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7322175732217573, | |
| "grad_norm": 21.28258346430573, | |
| "learning_rate": 1.0157994641835734e-07, | |
| "logits/chosen": -0.8515625, | |
| "logits/rejected": -0.83203125, | |
| "logps/chosen": -366.0, | |
| "logps/rejected": -400.0, | |
| "loss": 1.055, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.80078125, | |
| "rewards/margins": 0.59765625, | |
| "rewards/rejected": -1.40625, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7531380753138075, | |
| "grad_norm": 22.98970715168108, | |
| "learning_rate": 8.729103716819111e-08, | |
| "logits/chosen": -0.9375, | |
| "logits/rejected": -0.80078125, | |
| "logps/chosen": -404.0, | |
| "logps/rejected": -426.0, | |
| "loss": 1.0977, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.82421875, | |
| "rewards/margins": 0.6171875, | |
| "rewards/rejected": -1.4453125, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7740585774058577, | |
| "grad_norm": 24.00942251834518, | |
| "learning_rate": 7.387025063449081e-08, | |
| "logits/chosen": -0.875, | |
| "logits/rejected": -0.88671875, | |
| "logps/chosen": -380.0, | |
| "logps/rejected": -388.0, | |
| "loss": 1.0832, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.86328125, | |
| "rewards/margins": 0.5, | |
| "rewards/rejected": -1.359375, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7949790794979079, | |
| "grad_norm": 23.422353760489106, | |
| "learning_rate": 6.138919252022435e-08, | |
| "logits/chosen": -0.77734375, | |
| "logits/rejected": -0.7734375, | |
| "logps/chosen": -336.0, | |
| "logps/rejected": -428.0, | |
| "loss": 1.0845, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.83984375, | |
| "rewards/margins": 0.67578125, | |
| "rewards/rejected": -1.515625, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8158995815899581, | |
| "grad_norm": 26.11219518409401, | |
| "learning_rate": 4.991445467064689e-08, | |
| "logits/chosen": -0.9609375, | |
| "logits/rejected": -0.84765625, | |
| "logps/chosen": -404.0, | |
| "logps/rejected": -450.0, | |
| "loss": 1.0617, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.81640625, | |
| "rewards/margins": 0.6328125, | |
| "rewards/rejected": -1.4453125, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8368200836820083, | |
| "grad_norm": 34.377683916741226, | |
| "learning_rate": 3.9507259776993954e-08, | |
| "logits/chosen": -0.9296875, | |
| "logits/rejected": -0.84375, | |
| "logps/chosen": -382.0, | |
| "logps/rejected": -438.0, | |
| "loss": 1.0834, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.921875, | |
| "rewards/margins": 0.5859375, | |
| "rewards/rejected": -1.5078125, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8368200836820083, | |
| "eval_logits/chosen": -0.89453125, | |
| "eval_logits/rejected": -0.83984375, | |
| "eval_logps/chosen": -376.0, | |
| "eval_logps/rejected": -444.0, | |
| "eval_loss": 0.5443593859672546, | |
| "eval_rewards/accuracies": 0.77734375, | |
| "eval_rewards/chosen": -0.85546875, | |
| "eval_rewards/margins": 0.65625, | |
| "eval_rewards/rejected": -1.515625, | |
| "eval_runtime": 45.0507, | |
| "eval_samples_per_second": 44.394, | |
| "eval_steps_per_second": 0.71, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8577405857740585, | |
| "grad_norm": 18.194751008767053, | |
| "learning_rate": 3.022313472693447e-08, | |
| "logits/chosen": -0.9375, | |
| "logits/rejected": -0.88671875, | |
| "logps/chosen": -396.0, | |
| "logps/rejected": -430.0, | |
| "loss": 1.0862, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.890625, | |
| "rewards/margins": 0.609375, | |
| "rewards/rejected": -1.5, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8786610878661087, | |
| "grad_norm": 19.92181443843633, | |
| "learning_rate": 2.2111614344599684e-08, | |
| "logits/chosen": -0.84765625, | |
| "logits/rejected": -0.88671875, | |
| "logps/chosen": -434.0, | |
| "logps/rejected": -462.0, | |
| "loss": 1.073, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.015625, | |
| "rewards/margins": 0.5390625, | |
| "rewards/rejected": -1.5546875, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.899581589958159, | |
| "grad_norm": 23.788606985497122, | |
| "learning_rate": 1.521597710086439e-08, | |
| "logits/chosen": -0.828125, | |
| "logits/rejected": -0.78515625, | |
| "logps/chosen": -394.0, | |
| "logps/rejected": -428.0, | |
| "loss": 1.0378, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.8671875, | |
| "rewards/margins": 0.62890625, | |
| "rewards/rejected": -1.5, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9205020920502092, | |
| "grad_norm": 31.039553332885696, | |
| "learning_rate": 9.57301420397924e-09, | |
| "logits/chosen": -0.8828125, | |
| "logits/rejected": -0.828125, | |
| "logps/chosen": -394.0, | |
| "logps/rejected": -418.0, | |
| "loss": 1.0753, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.83203125, | |
| "rewards/margins": 0.51953125, | |
| "rewards/rejected": -1.3515625, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9414225941422594, | |
| "grad_norm": 21.613511008065277, | |
| "learning_rate": 5.212833302556258e-09, | |
| "logits/chosen": -0.93359375, | |
| "logits/rejected": -0.953125, | |
| "logps/chosen": -396.0, | |
| "logps/rejected": -460.0, | |
| "loss": 1.0728, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.89453125, | |
| "rewards/margins": 0.484375, | |
| "rewards/rejected": -1.375, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9623430962343096, | |
| "grad_norm": 31.165177122367144, | |
| "learning_rate": 2.158697848236607e-09, | |
| "logits/chosen": -0.87109375, | |
| "logits/rejected": -0.90234375, | |
| "logps/chosen": -382.0, | |
| "logps/rejected": -396.0, | |
| "loss": 1.0866, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.87109375, | |
| "rewards/margins": 0.515625, | |
| "rewards/rejected": -1.390625, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.9832635983263598, | |
| "grad_norm": 21.45190789431367, | |
| "learning_rate": 4.269029751107489e-10, | |
| "logits/chosen": -0.85546875, | |
| "logits/rejected": -0.859375, | |
| "logps/chosen": -380.0, | |
| "logps/rejected": -436.0, | |
| "loss": 1.0369, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.8203125, | |
| "rewards/margins": 0.5234375, | |
| "rewards/rejected": -1.34375, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 478, | |
| "total_flos": 0.0, | |
| "train_loss": 1.1715184933969665, | |
| "train_runtime": 4575.3933, | |
| "train_samples_per_second": 13.361, | |
| "train_steps_per_second": 0.104 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 478, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |