| { | |
| "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": 18.792079300993226, | |
| "learning_rate": 1.0416666666666666e-08, | |
| "logits/chosen": -3.359375, | |
| "logits/rejected": -3.34375, | |
| "logps/chosen": -320.0, | |
| "logps/rejected": -380.0, | |
| "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": 18.23398541151873, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -3.171875, | |
| "logits/rejected": -3.15625, | |
| "logps/chosen": -284.0, | |
| "logps/rejected": -268.0, | |
| "loss": 1.3841, | |
| "rewards/accuracies": 0.2083333283662796, | |
| "rewards/chosen": 0.00051116943359375, | |
| "rewards/margins": 0.000637054443359375, | |
| "rewards/rejected": -0.00012111663818359375, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04184100418410042, | |
| "grad_norm": 17.016957918520955, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -3.203125, | |
| "logits/rejected": -3.203125, | |
| "logps/chosen": -308.0, | |
| "logps/rejected": -316.0, | |
| "loss": 1.3834, | |
| "rewards/accuracies": 0.26249998807907104, | |
| "rewards/chosen": -0.000522613525390625, | |
| "rewards/margins": 0.0001621246337890625, | |
| "rewards/rejected": -0.000690460205078125, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06276150627615062, | |
| "grad_norm": 16.880895386487456, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -3.25, | |
| "logits/rejected": -3.234375, | |
| "logps/chosen": -324.0, | |
| "logps/rejected": -282.0, | |
| "loss": 1.3779, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.0010528564453125, | |
| "rewards/margins": 0.008056640625, | |
| "rewards/rejected": -0.006988525390625, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.08368200836820083, | |
| "grad_norm": 17.084942061574367, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -3.078125, | |
| "logits/rejected": -3.140625, | |
| "logps/chosen": -288.0, | |
| "logps/rejected": -276.0, | |
| "loss": 1.3604, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.007110595703125, | |
| "rewards/margins": 0.038330078125, | |
| "rewards/rejected": -0.04541015625, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.10460251046025104, | |
| "grad_norm": 16.90050268313775, | |
| "learning_rate": 4.999733114418725e-07, | |
| "logits/chosen": -3.078125, | |
| "logits/rejected": -3.0, | |
| "logps/chosen": -326.0, | |
| "logps/rejected": -332.0, | |
| "loss": 1.3289, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.047607421875, | |
| "rewards/margins": 0.0478515625, | |
| "rewards/rejected": -0.09521484375, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.12552301255230125, | |
| "grad_norm": 20.78289142989757, | |
| "learning_rate": 4.990398100856366e-07, | |
| "logits/chosen": -3.15625, | |
| "logits/rejected": -3.09375, | |
| "logps/chosen": -300.0, | |
| "logps/rejected": -338.0, | |
| "loss": 1.2877, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.091796875, | |
| "rewards/margins": 0.11279296875, | |
| "rewards/rejected": -0.205078125, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.14644351464435146, | |
| "grad_norm": 23.133166135835843, | |
| "learning_rate": 4.967775735898179e-07, | |
| "logits/chosen": -2.96875, | |
| "logits/rejected": -3.03125, | |
| "logps/chosen": -298.0, | |
| "logps/rejected": -308.0, | |
| "loss": 1.2374, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.0888671875, | |
| "rewards/margins": 0.1982421875, | |
| "rewards/rejected": -0.287109375, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.16736401673640167, | |
| "grad_norm": 38.62320509226835, | |
| "learning_rate": 4.931986719649298e-07, | |
| "logits/chosen": -3.109375, | |
| "logits/rejected": -3.125, | |
| "logps/chosen": -376.0, | |
| "logps/rejected": -336.0, | |
| "loss": 1.2072, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.27734375, | |
| "rewards/margins": 0.26953125, | |
| "rewards/rejected": -0.546875, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.18828451882845187, | |
| "grad_norm": 40.482824838074976, | |
| "learning_rate": 4.883222001996351e-07, | |
| "logits/chosen": -3.0625, | |
| "logits/rejected": -3.0625, | |
| "logps/chosen": -350.0, | |
| "logps/rejected": -398.0, | |
| "loss": 1.1208, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.66015625, | |
| "rewards/margins": 0.5625, | |
| "rewards/rejected": -1.2265625, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.20920502092050208, | |
| "grad_norm": 35.10500983971475, | |
| "learning_rate": 4.821741763807186e-07, | |
| "logits/chosen": -2.921875, | |
| "logits/rejected": -3.015625, | |
| "logps/chosen": -354.0, | |
| "logps/rejected": -406.0, | |
| "loss": 1.1291, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.6328125, | |
| "rewards/margins": 0.6015625, | |
| "rewards/rejected": -1.234375, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.20920502092050208, | |
| "eval_logits/chosen": -2.96875, | |
| "eval_logits/rejected": -2.953125, | |
| "eval_logps/chosen": -368.0, | |
| "eval_logps/rejected": -422.0, | |
| "eval_loss": 0.5672109127044678, | |
| "eval_rewards/accuracies": 0.71484375, | |
| "eval_rewards/chosen": -0.73828125, | |
| "eval_rewards/margins": 0.546875, | |
| "eval_rewards/rejected": -1.28125, | |
| "eval_runtime": 40.1054, | |
| "eval_samples_per_second": 49.869, | |
| "eval_steps_per_second": 0.798, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2301255230125523, | |
| "grad_norm": 49.983830380259825, | |
| "learning_rate": 4.747874028753375e-07, | |
| "logits/chosen": -2.9375, | |
| "logits/rejected": -2.984375, | |
| "logps/chosen": -418.0, | |
| "logps/rejected": -422.0, | |
| "loss": 1.1272, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.84375, | |
| "rewards/margins": 0.56640625, | |
| "rewards/rejected": -1.4140625, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.2510460251046025, | |
| "grad_norm": 44.36112825504322, | |
| "learning_rate": 4.662012913161997e-07, | |
| "logits/chosen": -2.828125, | |
| "logits/rejected": -2.875, | |
| "logps/chosen": -380.0, | |
| "logps/rejected": -418.0, | |
| "loss": 1.0984, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.6640625, | |
| "rewards/margins": 0.6484375, | |
| "rewards/rejected": -1.3125, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.2719665271966527, | |
| "grad_norm": 46.661481140072915, | |
| "learning_rate": 4.5646165232345103e-07, | |
| "logits/chosen": -2.8125, | |
| "logits/rejected": -2.890625, | |
| "logps/chosen": -404.0, | |
| "logps/rejected": -448.0, | |
| "loss": 1.0745, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.9765625, | |
| "rewards/margins": 0.61328125, | |
| "rewards/rejected": -1.5859375, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.2928870292887029, | |
| "grad_norm": 42.823235443382885, | |
| "learning_rate": 4.456204510851956e-07, | |
| "logits/chosen": -2.84375, | |
| "logits/rejected": -2.875, | |
| "logps/chosen": -408.0, | |
| "logps/rejected": -452.0, | |
| "loss": 1.0635, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.6875, | |
| "rewards/margins": 0.6484375, | |
| "rewards/rejected": -1.3359375, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.3138075313807531, | |
| "grad_norm": 59.20252738056763, | |
| "learning_rate": 4.337355301007335e-07, | |
| "logits/chosen": -2.828125, | |
| "logits/rejected": -2.78125, | |
| "logps/chosen": -406.0, | |
| "logps/rejected": -452.0, | |
| "loss": 1.0733, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.93359375, | |
| "rewards/margins": 0.55078125, | |
| "rewards/rejected": -1.484375, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.33472803347280333, | |
| "grad_norm": 51.1296977483052, | |
| "learning_rate": 4.2087030056579986e-07, | |
| "logits/chosen": -2.8125, | |
| "logits/rejected": -2.796875, | |
| "logps/chosen": -376.0, | |
| "logps/rejected": -440.0, | |
| "loss": 1.066, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.796875, | |
| "rewards/margins": 0.83984375, | |
| "rewards/rejected": -1.6328125, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.35564853556485354, | |
| "grad_norm": 55.991618044626925, | |
| "learning_rate": 4.070934040463998e-07, | |
| "logits/chosen": -2.640625, | |
| "logits/rejected": -2.609375, | |
| "logps/chosen": -374.0, | |
| "logps/rejected": -412.0, | |
| "loss": 1.082, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.0078125, | |
| "rewards/margins": 0.609375, | |
| "rewards/rejected": -1.609375, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.37656903765690375, | |
| "grad_norm": 50.525931210242035, | |
| "learning_rate": 3.9247834624635404e-07, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.765625, | |
| "logps/chosen": -358.0, | |
| "logps/rejected": -398.0, | |
| "loss": 1.0064, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.7890625, | |
| "rewards/margins": 0.69921875, | |
| "rewards/rejected": -1.484375, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.39748953974895396, | |
| "grad_norm": 46.856348860317034, | |
| "learning_rate": 3.7710310482256523e-07, | |
| "logits/chosen": -2.671875, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -390.0, | |
| "logps/rejected": -448.0, | |
| "loss": 1.0301, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.8515625, | |
| "rewards/margins": 0.65234375, | |
| "rewards/rejected": -1.5078125, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.41841004184100417, | |
| "grad_norm": 106.94783278420675, | |
| "learning_rate": 3.610497133404795e-07, | |
| "logits/chosen": -2.546875, | |
| "logits/rejected": -2.640625, | |
| "logps/chosen": -368.0, | |
| "logps/rejected": -430.0, | |
| "loss": 1.0723, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.828125, | |
| "rewards/margins": 0.73828125, | |
| "rewards/rejected": -1.5703125, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.41841004184100417, | |
| "eval_logits/chosen": -2.6875, | |
| "eval_logits/rejected": -2.65625, | |
| "eval_logps/chosen": -376.0, | |
| "eval_logps/rejected": -460.0, | |
| "eval_loss": 0.5170312523841858, | |
| "eval_rewards/accuracies": 0.765625, | |
| "eval_rewards/chosen": -0.8203125, | |
| "eval_rewards/margins": 0.83203125, | |
| "eval_rewards/rejected": -1.65625, | |
| "eval_runtime": 39.9201, | |
| "eval_samples_per_second": 50.1, | |
| "eval_steps_per_second": 0.802, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.4393305439330544, | |
| "grad_norm": 56.92194462169853, | |
| "learning_rate": 3.4440382358952115e-07, | |
| "logits/chosen": -2.671875, | |
| "logits/rejected": -2.640625, | |
| "logps/chosen": -402.0, | |
| "logps/rejected": -438.0, | |
| "loss": 1.0592, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8828125, | |
| "rewards/margins": 0.734375, | |
| "rewards/rejected": -1.6171875, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.4602510460251046, | |
| "grad_norm": 91.8785262604961, | |
| "learning_rate": 3.272542485937368e-07, | |
| "logits/chosen": -2.515625, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -386.0, | |
| "logps/rejected": -454.0, | |
| "loss": 1.036, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.96484375, | |
| "rewards/margins": 0.81640625, | |
| "rewards/rejected": -1.78125, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.4811715481171548, | |
| "grad_norm": 51.218599770252155, | |
| "learning_rate": 3.096924887558854e-07, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -376.0, | |
| "logps/rejected": -448.0, | |
| "loss": 1.0678, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.046875, | |
| "rewards/margins": 0.79296875, | |
| "rewards/rejected": -1.8359375, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.502092050209205, | |
| "grad_norm": 52.33600440489297, | |
| "learning_rate": 2.9181224366319943e-07, | |
| "logits/chosen": -2.65625, | |
| "logits/rejected": -2.546875, | |
| "logps/chosen": -372.0, | |
| "logps/rejected": -428.0, | |
| "loss": 1.0064, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.80859375, | |
| "rewards/margins": 0.6953125, | |
| "rewards/rejected": -1.5, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.5230125523012552, | |
| "grad_norm": 51.27684803998907, | |
| "learning_rate": 2.7370891215954565e-07, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.5625, | |
| "logps/chosen": -424.0, | |
| "logps/rejected": -478.0, | |
| "loss": 1.0087, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.0390625, | |
| "rewards/margins": 0.8828125, | |
| "rewards/rejected": -1.921875, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.5439330543933054, | |
| "grad_norm": 63.264740099851814, | |
| "learning_rate": 2.55479083351317e-07, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.65625, | |
| "logps/chosen": -424.0, | |
| "logps/rejected": -472.0, | |
| "loss": 1.0101, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.09375, | |
| "rewards/margins": 0.875, | |
| "rewards/rejected": -1.96875, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.5648535564853556, | |
| "grad_norm": 43.895031684867064, | |
| "learning_rate": 2.3722002126275822e-07, | |
| "logits/chosen": -2.59375, | |
| "logits/rejected": -2.59375, | |
| "logps/chosen": -396.0, | |
| "logps/rejected": -438.0, | |
| "loss": 1.0423, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.8984375, | |
| "rewards/margins": 0.68359375, | |
| "rewards/rejected": -1.5859375, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.5857740585774058, | |
| "grad_norm": 53.26137168286982, | |
| "learning_rate": 2.19029145890313e-07, | |
| "logits/chosen": -2.546875, | |
| "logits/rejected": -2.5625, | |
| "logps/chosen": -374.0, | |
| "logps/rejected": -426.0, | |
| "loss": 1.0536, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.81640625, | |
| "rewards/margins": 0.80078125, | |
| "rewards/rejected": -1.6171875, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.606694560669456, | |
| "grad_norm": 57.71809127266035, | |
| "learning_rate": 2.0100351342479216e-07, | |
| "logits/chosen": -2.640625, | |
| "logits/rejected": -2.609375, | |
| "logps/chosen": -386.0, | |
| "logps/rejected": -450.0, | |
| "loss": 1.0334, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.1171875, | |
| "rewards/margins": 0.75, | |
| "rewards/rejected": -1.859375, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.6276150627615062, | |
| "grad_norm": 59.26080657165494, | |
| "learning_rate": 1.8323929841460178e-07, | |
| "logits/chosen": -2.640625, | |
| "logits/rejected": -2.609375, | |
| "logps/chosen": -422.0, | |
| "logps/rejected": -460.0, | |
| "loss": 0.9917, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.94140625, | |
| "rewards/margins": 0.81640625, | |
| "rewards/rejected": -1.7578125, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6276150627615062, | |
| "eval_logits/chosen": -2.625, | |
| "eval_logits/rejected": -2.59375, | |
| "eval_logps/chosen": -372.0, | |
| "eval_logps/rejected": -464.0, | |
| "eval_loss": 0.5082695484161377, | |
| "eval_rewards/accuracies": 0.75390625, | |
| "eval_rewards/chosen": -0.77734375, | |
| "eval_rewards/margins": 0.92578125, | |
| "eval_rewards/rejected": -1.703125, | |
| "eval_runtime": 39.9557, | |
| "eval_samples_per_second": 50.055, | |
| "eval_steps_per_second": 0.801, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.6485355648535565, | |
| "grad_norm": 63.68800654846809, | |
| "learning_rate": 1.6583128063291573e-07, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -416.0, | |
| "logps/rejected": -452.0, | |
| "loss": 0.9939, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.94140625, | |
| "rewards/margins": 0.7734375, | |
| "rewards/rejected": -1.71875, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.6694560669456067, | |
| "grad_norm": 65.12912368302251, | |
| "learning_rate": 1.488723393865766e-07, | |
| "logits/chosen": -2.59375, | |
| "logits/rejected": -2.546875, | |
| "logps/chosen": -412.0, | |
| "logps/rejected": -444.0, | |
| "loss": 0.97, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.95703125, | |
| "rewards/margins": 0.80078125, | |
| "rewards/rejected": -1.7578125, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.6903765690376569, | |
| "grad_norm": 64.88040587462643, | |
| "learning_rate": 1.3245295796480788e-07, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.578125, | |
| "logps/chosen": -388.0, | |
| "logps/rejected": -452.0, | |
| "loss": 0.9984, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.91796875, | |
| "rewards/margins": 0.67578125, | |
| "rewards/rejected": -1.59375, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.7112970711297071, | |
| "grad_norm": 97.36761933497387, | |
| "learning_rate": 1.1666074087171627e-07, | |
| "logits/chosen": -2.65625, | |
| "logits/rejected": -2.546875, | |
| "logps/chosen": -400.0, | |
| "logps/rejected": -490.0, | |
| "loss": 0.9862, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.80859375, | |
| "rewards/margins": 0.9921875, | |
| "rewards/rejected": -1.796875, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.7322175732217573, | |
| "grad_norm": 55.29816737345383, | |
| "learning_rate": 1.0157994641835734e-07, | |
| "logits/chosen": -2.453125, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -390.0, | |
| "logps/rejected": -452.0, | |
| "loss": 0.9503, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.0, | |
| "rewards/margins": 0.87109375, | |
| "rewards/rejected": -1.875, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.7531380753138075, | |
| "grad_norm": 50.20493217819224, | |
| "learning_rate": 8.729103716819111e-08, | |
| "logits/chosen": -2.625, | |
| "logits/rejected": -2.46875, | |
| "logps/chosen": -438.0, | |
| "logps/rejected": -486.0, | |
| "loss": 1.0227, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.09375, | |
| "rewards/margins": 0.921875, | |
| "rewards/rejected": -2.015625, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.7740585774058577, | |
| "grad_norm": 52.540692008100564, | |
| "learning_rate": 7.387025063449081e-08, | |
| "logits/chosen": -2.46875, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -420.0, | |
| "logps/rejected": -448.0, | |
| "loss": 1.0167, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.2109375, | |
| "rewards/margins": 0.6796875, | |
| "rewards/rejected": -1.890625, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.7949790794979079, | |
| "grad_norm": 52.1938302087835, | |
| "learning_rate": 6.138919252022435e-08, | |
| "logits/chosen": -2.296875, | |
| "logits/rejected": -2.34375, | |
| "logps/chosen": -390.0, | |
| "logps/rejected": -506.0, | |
| "loss": 1.0047, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.296875, | |
| "rewards/margins": 0.94921875, | |
| "rewards/rejected": -2.25, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.8158995815899581, | |
| "grad_norm": 58.62340593152117, | |
| "learning_rate": 4.991445467064689e-08, | |
| "logits/chosen": -2.515625, | |
| "logits/rejected": -2.46875, | |
| "logps/chosen": -442.0, | |
| "logps/rejected": -500.0, | |
| "loss": 0.9798, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.046875, | |
| "rewards/margins": 0.85546875, | |
| "rewards/rejected": -1.8984375, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.8368200836820083, | |
| "grad_norm": 59.42878062470151, | |
| "learning_rate": 3.9507259776993954e-08, | |
| "logits/chosen": -2.515625, | |
| "logits/rejected": -2.453125, | |
| "logps/chosen": -404.0, | |
| "logps/rejected": -490.0, | |
| "loss": 1.012, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.09375, | |
| "rewards/margins": 0.89453125, | |
| "rewards/rejected": -1.984375, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8368200836820083, | |
| "eval_logits/chosen": -2.515625, | |
| "eval_logits/rejected": -2.46875, | |
| "eval_logps/chosen": -386.0, | |
| "eval_logps/rejected": -484.0, | |
| "eval_loss": 0.4995781183242798, | |
| "eval_rewards/accuracies": 0.78125, | |
| "eval_rewards/chosen": -0.91015625, | |
| "eval_rewards/margins": 0.984375, | |
| "eval_rewards/rejected": -1.890625, | |
| "eval_runtime": 39.9442, | |
| "eval_samples_per_second": 50.07, | |
| "eval_steps_per_second": 0.801, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.8577405857740585, | |
| "grad_norm": 57.94077111648529, | |
| "learning_rate": 3.022313472693447e-08, | |
| "logits/chosen": -2.5625, | |
| "logits/rejected": -2.53125, | |
| "logps/chosen": -418.0, | |
| "logps/rejected": -468.0, | |
| "loss": 1.0227, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.0, | |
| "rewards/margins": 0.84375, | |
| "rewards/rejected": -1.84375, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.8786610878661087, | |
| "grad_norm": 54.220495002656904, | |
| "learning_rate": 2.2111614344599684e-08, | |
| "logits/chosen": -2.421875, | |
| "logits/rejected": -2.484375, | |
| "logps/chosen": -442.0, | |
| "logps/rejected": -502.0, | |
| "loss": 0.9755, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.0234375, | |
| "rewards/margins": 0.859375, | |
| "rewards/rejected": -1.8828125, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.899581589958159, | |
| "grad_norm": 57.28244651782191, | |
| "learning_rate": 1.521597710086439e-08, | |
| "logits/chosen": -2.375, | |
| "logits/rejected": -2.265625, | |
| "logps/chosen": -418.0, | |
| "logps/rejected": -480.0, | |
| "loss": 0.9483, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.0703125, | |
| "rewards/margins": 0.91015625, | |
| "rewards/rejected": -1.9765625, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.9205020920502092, | |
| "grad_norm": 63.185522298582946, | |
| "learning_rate": 9.57301420397924e-09, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -408.0, | |
| "logps/rejected": -460.0, | |
| "loss": 0.9976, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.9140625, | |
| "rewards/margins": 0.796875, | |
| "rewards/rejected": -1.7109375, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.9414225941422594, | |
| "grad_norm": 72.06238722374589, | |
| "learning_rate": 5.212833302556258e-09, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.515625, | |
| "logps/chosen": -420.0, | |
| "logps/rejected": -510.0, | |
| "loss": 0.9918, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.0, | |
| "rewards/margins": 0.8125, | |
| "rewards/rejected": -1.8125, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.9623430962343096, | |
| "grad_norm": 63.02015092631578, | |
| "learning_rate": 2.158697848236607e-09, | |
| "logits/chosen": -2.484375, | |
| "logits/rejected": -2.453125, | |
| "logps/chosen": -396.0, | |
| "logps/rejected": -440.0, | |
| "loss": 0.9981, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.95703125, | |
| "rewards/margins": 0.82421875, | |
| "rewards/rejected": -1.78125, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.9832635983263598, | |
| "grad_norm": 58.50751073465125, | |
| "learning_rate": 4.269029751107489e-10, | |
| "logits/chosen": -2.4375, | |
| "logits/rejected": -2.4375, | |
| "logps/chosen": -404.0, | |
| "logps/rejected": -498.0, | |
| "loss": 0.9687, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.93359375, | |
| "rewards/margins": 0.8984375, | |
| "rewards/rejected": -1.8359375, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 478, | |
| "total_flos": 0.0, | |
| "train_loss": 1.0759161147113625, | |
| "train_runtime": 3593.5482, | |
| "train_samples_per_second": 17.012, | |
| "train_steps_per_second": 0.133 | |
| } | |
| ], | |
| "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 | |
| } | |