| { |
| "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": 0.0, |
| "learning_rate": 1.0416666666666666e-08, |
| "logits/chosen": -3.948974132537842, |
| "logits/rejected": -3.9480834007263184, |
| "logps/chosen": -1228.763427734375, |
| "logps/rejected": -1094.858642578125, |
| "loss": 0.6931, |
| "rewards/accuracies": 0.0, |
| "rewards/chosen": 0.0, |
| "rewards/margins": 0.0, |
| "rewards/rejected": 0.0, |
| "step": 1 |
| }, |
| { |
| "epoch": 0.02, |
| "learning_rate": 1.0416666666666667e-07, |
| "logits/chosen": -4.02778434753418, |
| "logits/rejected": -4.042763710021973, |
| "logps/chosen": -989.794189453125, |
| "logps/rejected": -831.4175415039062, |
| "loss": 0.6932, |
| "rewards/accuracies": 0.5486111044883728, |
| "rewards/chosen": 0.0007523840176872909, |
| "rewards/margins": 0.0010657603852450848, |
| "rewards/rejected": -0.00031337636755779386, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.04, |
| "learning_rate": 2.0833333333333333e-07, |
| "logits/chosen": -4.056273460388184, |
| "logits/rejected": -4.030348777770996, |
| "logps/chosen": -1058.2525634765625, |
| "logps/rejected": -935.8595581054688, |
| "loss": 0.6926, |
| "rewards/accuracies": 0.48124998807907104, |
| "rewards/chosen": 0.001010889420285821, |
| "rewards/margins": 0.0009087520884349942, |
| "rewards/rejected": 0.00010213752102572471, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.06, |
| "learning_rate": 3.1249999999999997e-07, |
| "logits/chosen": -4.065732479095459, |
| "logits/rejected": -4.047877311706543, |
| "logps/chosen": -1169.9014892578125, |
| "logps/rejected": -1036.8551025390625, |
| "loss": 0.69, |
| "rewards/accuracies": 0.574999988079071, |
| "rewards/chosen": 0.003177918028086424, |
| "rewards/margins": 0.009787282906472683, |
| "rewards/rejected": -0.006609364412724972, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.08, |
| "learning_rate": 4.1666666666666667e-07, |
| "logits/chosen": -3.960789203643799, |
| "logits/rejected": -3.9753544330596924, |
| "logps/chosen": -1171.2132568359375, |
| "logps/rejected": -1020.228515625, |
| "loss": 0.6847, |
| "rewards/accuracies": 0.550000011920929, |
| "rewards/chosen": -0.02048645354807377, |
| "rewards/margins": 0.007100983057171106, |
| "rewards/rejected": -0.027587438002228737, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.1, |
| "learning_rate": 4.999733114418725e-07, |
| "logits/chosen": -3.9188435077667236, |
| "logits/rejected": -3.906773090362549, |
| "logps/chosen": -1267.8223876953125, |
| "logps/rejected": -1147.726806640625, |
| "loss": 0.6788, |
| "rewards/accuracies": 0.5874999761581421, |
| "rewards/chosen": -0.07415176928043365, |
| "rewards/margins": 0.033729877322912216, |
| "rewards/rejected": -0.10788165032863617, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.13, |
| "learning_rate": 4.990398100856366e-07, |
| "logits/chosen": -3.914989471435547, |
| "logits/rejected": -3.897732973098755, |
| "logps/chosen": -981.6224365234375, |
| "logps/rejected": -830.83154296875, |
| "loss": 0.6774, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": -0.15918272733688354, |
| "rewards/margins": 0.023405645042657852, |
| "rewards/rejected": -0.1825883686542511, |
| "step": 60 |
| }, |
| { |
| "epoch": 0.15, |
| "learning_rate": 4.967775735898179e-07, |
| "logits/chosen": -3.8780300617218018, |
| "logits/rejected": -3.8814609050750732, |
| "logps/chosen": -1141.1514892578125, |
| "logps/rejected": -967.8585205078125, |
| "loss": 0.6755, |
| "rewards/accuracies": 0.543749988079071, |
| "rewards/chosen": -0.17965973913669586, |
| "rewards/margins": 0.0425986722111702, |
| "rewards/rejected": -0.22225841879844666, |
| "step": 70 |
| }, |
| { |
| "epoch": 0.17, |
| "learning_rate": 4.931986719649298e-07, |
| "logits/chosen": -3.848781108856201, |
| "logits/rejected": -3.8655917644500732, |
| "logps/chosen": -1174.486083984375, |
| "logps/rejected": -1068.8419189453125, |
| "loss": 0.6682, |
| "rewards/accuracies": 0.6000000238418579, |
| "rewards/chosen": -0.2266133576631546, |
| "rewards/margins": 0.042132142931222916, |
| "rewards/rejected": -0.2687455415725708, |
| "step": 80 |
| }, |
| { |
| "epoch": 0.19, |
| "learning_rate": 4.883222001996351e-07, |
| "logits/chosen": -3.8744590282440186, |
| "logits/rejected": -3.883523464202881, |
| "logps/chosen": -1135.01611328125, |
| "logps/rejected": -965.5120239257812, |
| "loss": 0.6533, |
| "rewards/accuracies": 0.6499999761581421, |
| "rewards/chosen": -0.24657884240150452, |
| "rewards/margins": 0.1434774398803711, |
| "rewards/rejected": -0.3900562822818756, |
| "step": 90 |
| }, |
| { |
| "epoch": 0.21, |
| "learning_rate": 4.821741763807186e-07, |
| "logits/chosen": -3.895872116088867, |
| "logits/rejected": -3.920149326324463, |
| "logps/chosen": -1210.6611328125, |
| "logps/rejected": -1016.2737426757812, |
| "loss": 0.6596, |
| "rewards/accuracies": 0.5874999761581421, |
| "rewards/chosen": -0.3355700373649597, |
| "rewards/margins": 0.05883771926164627, |
| "rewards/rejected": -0.3944077491760254, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.21, |
| "eval_logits/chosen": -3.886859893798828, |
| "eval_logits/rejected": -3.893383502960205, |
| "eval_logps/chosen": -1158.9254150390625, |
| "eval_logps/rejected": -1061.103759765625, |
| "eval_loss": 0.6581785678863525, |
| "eval_rewards/accuracies": 0.6171875, |
| "eval_rewards/chosen": -0.3306271433830261, |
| "eval_rewards/margins": 0.1209341362118721, |
| "eval_rewards/rejected": -0.4515612721443176, |
| "eval_runtime": 139.8808, |
| "eval_samples_per_second": 14.298, |
| "eval_steps_per_second": 0.229, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.23, |
| "learning_rate": 4.747874028753375e-07, |
| "logits/chosen": -3.8644802570343018, |
| "logits/rejected": -3.8876547813415527, |
| "logps/chosen": -1274.4202880859375, |
| "logps/rejected": -1031.0589599609375, |
| "loss": 0.6637, |
| "rewards/accuracies": 0.637499988079071, |
| "rewards/chosen": -0.38339611887931824, |
| "rewards/margins": 0.12875327467918396, |
| "rewards/rejected": -0.5121493935585022, |
| "step": 110 |
| }, |
| { |
| "epoch": 0.25, |
| "learning_rate": 4.662012913161997e-07, |
| "logits/chosen": -3.820500135421753, |
| "logits/rejected": -3.8377246856689453, |
| "logps/chosen": -1214.8111572265625, |
| "logps/rejected": -960.5948486328125, |
| "loss": 0.6429, |
| "rewards/accuracies": 0.612500011920929, |
| "rewards/chosen": -0.5467379689216614, |
| "rewards/margins": 0.10745690017938614, |
| "rewards/rejected": -0.6541949510574341, |
| "step": 120 |
| }, |
| { |
| "epoch": 0.27, |
| "learning_rate": 4.5646165232345103e-07, |
| "logits/chosen": -3.8416264057159424, |
| "logits/rejected": -3.8344688415527344, |
| "logps/chosen": -1127.9312744140625, |
| "logps/rejected": -1137.460205078125, |
| "loss": 0.6629, |
| "rewards/accuracies": 0.6000000238418579, |
| "rewards/chosen": -0.805020809173584, |
| "rewards/margins": 0.14469251036643982, |
| "rewards/rejected": -0.9497132301330566, |
| "step": 130 |
| }, |
| { |
| "epoch": 0.29, |
| "learning_rate": 4.456204510851956e-07, |
| "logits/chosen": -3.8683509826660156, |
| "logits/rejected": -3.872532606124878, |
| "logps/chosen": -1258.8890380859375, |
| "logps/rejected": -1094.35888671875, |
| "loss": 0.6504, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": -0.6595749258995056, |
| "rewards/margins": 0.17474223673343658, |
| "rewards/rejected": -0.8343170881271362, |
| "step": 140 |
| }, |
| { |
| "epoch": 0.31, |
| "learning_rate": 4.337355301007335e-07, |
| "logits/chosen": -3.833204984664917, |
| "logits/rejected": -3.8267292976379395, |
| "logps/chosen": -1194.625732421875, |
| "logps/rejected": -976.2630615234375, |
| "loss": 0.6429, |
| "rewards/accuracies": 0.5874999761581421, |
| "rewards/chosen": -0.6854017972946167, |
| "rewards/margins": 0.17019221186637878, |
| "rewards/rejected": -0.8555939793586731, |
| "step": 150 |
| }, |
| { |
| "epoch": 0.33, |
| "learning_rate": 4.2087030056579986e-07, |
| "logits/chosen": -3.73237681388855, |
| "logits/rejected": -3.732792615890503, |
| "logps/chosen": -1198.712646484375, |
| "logps/rejected": -1073.5947265625, |
| "loss": 0.6366, |
| "rewards/accuracies": 0.5625, |
| "rewards/chosen": -0.7374709248542786, |
| "rewards/margins": 0.20430393517017365, |
| "rewards/rejected": -0.9417747259140015, |
| "step": 160 |
| }, |
| { |
| "epoch": 0.36, |
| "learning_rate": 4.070934040463998e-07, |
| "logits/chosen": -3.7786946296691895, |
| "logits/rejected": -3.758208751678467, |
| "logps/chosen": -1110.61376953125, |
| "logps/rejected": -1014.2169189453125, |
| "loss": 0.6622, |
| "rewards/accuracies": 0.5562499761581421, |
| "rewards/chosen": -0.7858539819717407, |
| "rewards/margins": 0.0793028324842453, |
| "rewards/rejected": -0.8651568293571472, |
| "step": 170 |
| }, |
| { |
| "epoch": 0.38, |
| "learning_rate": 3.9247834624635404e-07, |
| "logits/chosen": -3.7446670532226562, |
| "logits/rejected": -3.75555157661438, |
| "logps/chosen": -1262.5172119140625, |
| "logps/rejected": -1211.857177734375, |
| "loss": 0.6451, |
| "rewards/accuracies": 0.6187499761581421, |
| "rewards/chosen": -1.284397006034851, |
| "rewards/margins": 0.20995619893074036, |
| "rewards/rejected": -1.4943532943725586, |
| "step": 180 |
| }, |
| { |
| "epoch": 0.4, |
| "learning_rate": 3.7710310482256523e-07, |
| "logits/chosen": -3.844573974609375, |
| "logits/rejected": -3.835858106613159, |
| "logps/chosen": -1269.37548828125, |
| "logps/rejected": -1180.8189697265625, |
| "loss": 0.6382, |
| "rewards/accuracies": 0.6625000238418579, |
| "rewards/chosen": -1.0775713920593262, |
| "rewards/margins": 0.20712728798389435, |
| "rewards/rejected": -1.284698724746704, |
| "step": 190 |
| }, |
| { |
| "epoch": 0.42, |
| "learning_rate": 3.610497133404795e-07, |
| "logits/chosen": -3.800981044769287, |
| "logits/rejected": -3.7990424633026123, |
| "logps/chosen": -1206.808349609375, |
| "logps/rejected": -1050.677978515625, |
| "loss": 0.6329, |
| "rewards/accuracies": 0.6000000238418579, |
| "rewards/chosen": -0.9752827882766724, |
| "rewards/margins": 0.11567431688308716, |
| "rewards/rejected": -1.0909572839736938, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.42, |
| "eval_logits/chosen": -3.8271100521087646, |
| "eval_logits/rejected": -3.8301334381103516, |
| "eval_logps/chosen": -1215.5958251953125, |
| "eval_logps/rejected": -1124.1610107421875, |
| "eval_loss": 0.6346434354782104, |
| "eval_rewards/accuracies": 0.6171875, |
| "eval_rewards/chosen": -0.8973308205604553, |
| "eval_rewards/margins": 0.18480299413204193, |
| "eval_rewards/rejected": -1.0821337699890137, |
| "eval_runtime": 139.8865, |
| "eval_samples_per_second": 14.297, |
| "eval_steps_per_second": 0.229, |
| "step": 200 |
| }, |
| { |
| "epoch": 0.44, |
| "learning_rate": 3.4440382358952115e-07, |
| "logits/chosen": -3.7631287574768066, |
| "logits/rejected": -3.7575225830078125, |
| "logps/chosen": -1218.4136962890625, |
| "logps/rejected": -1040.703857421875, |
| "loss": 0.6286, |
| "rewards/accuracies": 0.606249988079071, |
| "rewards/chosen": -1.0075339078903198, |
| "rewards/margins": 0.198730930685997, |
| "rewards/rejected": -1.206264853477478, |
| "step": 210 |
| }, |
| { |
| "epoch": 0.46, |
| "learning_rate": 3.272542485937368e-07, |
| "logits/chosen": -3.768575668334961, |
| "logits/rejected": -3.7488741874694824, |
| "logps/chosen": -1258.8096923828125, |
| "logps/rejected": -1142.510986328125, |
| "loss": 0.642, |
| "rewards/accuracies": 0.6499999761581421, |
| "rewards/chosen": -1.4313360452651978, |
| "rewards/margins": 0.30053994059562683, |
| "rewards/rejected": -1.731876015663147, |
| "step": 220 |
| }, |
| { |
| "epoch": 0.48, |
| "learning_rate": 3.096924887558854e-07, |
| "logits/chosen": -3.841555118560791, |
| "logits/rejected": -3.8393311500549316, |
| "logps/chosen": -1235.6275634765625, |
| "logps/rejected": -1221.802490234375, |
| "loss": 0.6336, |
| "rewards/accuracies": 0.6312500238418579, |
| "rewards/chosen": -1.150726079940796, |
| "rewards/margins": 0.19467496871948242, |
| "rewards/rejected": -1.3454009294509888, |
| "step": 230 |
| }, |
| { |
| "epoch": 0.5, |
| "learning_rate": 2.9181224366319943e-07, |
| "logits/chosen": -3.7873482704162598, |
| "logits/rejected": -3.7755465507507324, |
| "logps/chosen": -1102.6480712890625, |
| "logps/rejected": -1071.907470703125, |
| "loss": 0.6386, |
| "rewards/accuracies": 0.625, |
| "rewards/chosen": -1.1753332614898682, |
| "rewards/margins": 0.2552703022956848, |
| "rewards/rejected": -1.4306036233901978, |
| "step": 240 |
| }, |
| { |
| "epoch": 0.52, |
| "learning_rate": 2.7370891215954565e-07, |
| "logits/chosen": -3.7938873767852783, |
| "logits/rejected": -3.7814221382141113, |
| "logps/chosen": -1199.1143798828125, |
| "logps/rejected": -1127.072509765625, |
| "loss": 0.6234, |
| "rewards/accuracies": 0.606249988079071, |
| "rewards/chosen": -1.2862367630004883, |
| "rewards/margins": 0.1962926685810089, |
| "rewards/rejected": -1.4825294017791748, |
| "step": 250 |
| }, |
| { |
| "epoch": 0.54, |
| "learning_rate": 2.55479083351317e-07, |
| "logits/chosen": -3.7459778785705566, |
| "logits/rejected": -3.7299225330352783, |
| "logps/chosen": -1272.9356689453125, |
| "logps/rejected": -1093.160888671875, |
| "loss": 0.6235, |
| "rewards/accuracies": 0.675000011920929, |
| "rewards/chosen": -1.6445872783660889, |
| "rewards/margins": 0.31901127099990845, |
| "rewards/rejected": -1.963598608970642, |
| "step": 260 |
| }, |
| { |
| "epoch": 0.56, |
| "learning_rate": 2.3722002126275822e-07, |
| "logits/chosen": -3.699172258377075, |
| "logits/rejected": -3.690387010574341, |
| "logps/chosen": -1218.9598388671875, |
| "logps/rejected": -1155.4327392578125, |
| "loss": 0.6172, |
| "rewards/accuracies": 0.643750011920929, |
| "rewards/chosen": -1.3695436716079712, |
| "rewards/margins": 0.258453905582428, |
| "rewards/rejected": -1.627997636795044, |
| "step": 270 |
| }, |
| { |
| "epoch": 0.59, |
| "learning_rate": 2.19029145890313e-07, |
| "logits/chosen": -3.7434639930725098, |
| "logits/rejected": -3.74462628364563, |
| "logps/chosen": -1194.427978515625, |
| "logps/rejected": -1074.471923828125, |
| "loss": 0.6292, |
| "rewards/accuracies": 0.574999988079071, |
| "rewards/chosen": -1.259026288986206, |
| "rewards/margins": 0.23618105053901672, |
| "rewards/rejected": -1.4952075481414795, |
| "step": 280 |
| }, |
| { |
| "epoch": 0.61, |
| "learning_rate": 2.0100351342479216e-07, |
| "logits/chosen": -3.748202085494995, |
| "logits/rejected": -3.749154567718506, |
| "logps/chosen": -1363.892333984375, |
| "logps/rejected": -1237.686767578125, |
| "loss": 0.6268, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": -2.0306642055511475, |
| "rewards/margins": 0.26983994245529175, |
| "rewards/rejected": -2.300504207611084, |
| "step": 290 |
| }, |
| { |
| "epoch": 0.63, |
| "learning_rate": 1.8323929841460178e-07, |
| "logits/chosen": -3.7096924781799316, |
| "logits/rejected": -3.692265272140503, |
| "logps/chosen": -1310.372802734375, |
| "logps/rejected": -1217.112548828125, |
| "loss": 0.6282, |
| "rewards/accuracies": 0.6312500238418579, |
| "rewards/chosen": -1.6397817134857178, |
| "rewards/margins": 0.23407919704914093, |
| "rewards/rejected": -1.8738609552383423, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.63, |
| "eval_logits/chosen": -3.781625270843506, |
| "eval_logits/rejected": -3.7836155891418457, |
| "eval_logps/chosen": -1263.44384765625, |
| "eval_logps/rejected": -1176.3936767578125, |
| "eval_loss": 0.6260908842086792, |
| "eval_rewards/accuracies": 0.6171875, |
| "eval_rewards/chosen": -1.375810146331787, |
| "eval_rewards/margins": 0.22865042090415955, |
| "eval_rewards/rejected": -1.604460597038269, |
| "eval_runtime": 140.2825, |
| "eval_samples_per_second": 14.257, |
| "eval_steps_per_second": 0.228, |
| "step": 300 |
| }, |
| { |
| "epoch": 0.65, |
| "learning_rate": 1.6583128063291573e-07, |
| "logits/chosen": -3.809044361114502, |
| "logits/rejected": -3.791278123855591, |
| "logps/chosen": -1239.725341796875, |
| "logps/rejected": -1062.396240234375, |
| "loss": 0.6125, |
| "rewards/accuracies": 0.7250000238418579, |
| "rewards/chosen": -1.3013659715652466, |
| "rewards/margins": 0.32230299711227417, |
| "rewards/rejected": -1.6236690282821655, |
| "step": 310 |
| }, |
| { |
| "epoch": 0.67, |
| "learning_rate": 1.488723393865766e-07, |
| "logits/chosen": -3.765423536300659, |
| "logits/rejected": -3.7561004161834717, |
| "logps/chosen": -1140.689697265625, |
| "logps/rejected": -1053.4720458984375, |
| "loss": 0.6058, |
| "rewards/accuracies": 0.71875, |
| "rewards/chosen": -1.3573421239852905, |
| "rewards/margins": 0.37179604172706604, |
| "rewards/rejected": -1.7291381359100342, |
| "step": 320 |
| }, |
| { |
| "epoch": 0.69, |
| "learning_rate": 1.3245295796480788e-07, |
| "logits/chosen": -3.785099506378174, |
| "logits/rejected": -3.775768995285034, |
| "logps/chosen": -1286.830078125, |
| "logps/rejected": -1169.2603759765625, |
| "loss": 0.5974, |
| "rewards/accuracies": 0.65625, |
| "rewards/chosen": -1.6952753067016602, |
| "rewards/margins": 0.3214367628097534, |
| "rewards/rejected": -2.016711950302124, |
| "step": 330 |
| }, |
| { |
| "epoch": 0.71, |
| "learning_rate": 1.1666074087171627e-07, |
| "logits/chosen": -3.743661880493164, |
| "logits/rejected": -3.743412494659424, |
| "logps/chosen": -1269.880615234375, |
| "logps/rejected": -1129.3505859375, |
| "loss": 0.6214, |
| "rewards/accuracies": 0.612500011920929, |
| "rewards/chosen": -1.5259571075439453, |
| "rewards/margins": 0.2673645317554474, |
| "rewards/rejected": -1.7933216094970703, |
| "step": 340 |
| }, |
| { |
| "epoch": 0.73, |
| "learning_rate": 1.0157994641835734e-07, |
| "logits/chosen": -3.6973376274108887, |
| "logits/rejected": -3.685258150100708, |
| "logps/chosen": -1198.2239990234375, |
| "logps/rejected": -1045.2313232421875, |
| "loss": 0.6104, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": -1.5120819807052612, |
| "rewards/margins": 0.24861010909080505, |
| "rewards/rejected": -1.7606918811798096, |
| "step": 350 |
| }, |
| { |
| "epoch": 0.75, |
| "learning_rate": 8.729103716819111e-08, |
| "logits/chosen": -3.763815402984619, |
| "logits/rejected": -3.751462459564209, |
| "logps/chosen": -1353.9212646484375, |
| "logps/rejected": -1129.874755859375, |
| "loss": 0.6195, |
| "rewards/accuracies": 0.706250011920929, |
| "rewards/chosen": -1.5439411401748657, |
| "rewards/margins": 0.3551940321922302, |
| "rewards/rejected": -1.8991349935531616, |
| "step": 360 |
| }, |
| { |
| "epoch": 0.77, |
| "learning_rate": 7.387025063449081e-08, |
| "logits/chosen": -3.702362060546875, |
| "logits/rejected": -3.679819107055664, |
| "logps/chosen": -1319.5167236328125, |
| "logps/rejected": -1136.005859375, |
| "loss": 0.6216, |
| "rewards/accuracies": 0.6499999761581421, |
| "rewards/chosen": -1.571559190750122, |
| "rewards/margins": 0.3320244550704956, |
| "rewards/rejected": -1.9035835266113281, |
| "step": 370 |
| }, |
| { |
| "epoch": 0.79, |
| "learning_rate": 6.138919252022435e-08, |
| "logits/chosen": -3.723046064376831, |
| "logits/rejected": -3.7033233642578125, |
| "logps/chosen": -1308.8182373046875, |
| "logps/rejected": -1194.8675537109375, |
| "loss": 0.6179, |
| "rewards/accuracies": 0.606249988079071, |
| "rewards/chosen": -1.5545730590820312, |
| "rewards/margins": 0.248977929353714, |
| "rewards/rejected": -1.8035507202148438, |
| "step": 380 |
| }, |
| { |
| "epoch": 0.82, |
| "learning_rate": 4.991445467064689e-08, |
| "logits/chosen": -3.7112629413604736, |
| "logits/rejected": -3.699855089187622, |
| "logps/chosen": -1272.218994140625, |
| "logps/rejected": -1214.7161865234375, |
| "loss": 0.6171, |
| "rewards/accuracies": 0.668749988079071, |
| "rewards/chosen": -1.7044214010238647, |
| "rewards/margins": 0.2680589556694031, |
| "rewards/rejected": -1.9724804162979126, |
| "step": 390 |
| }, |
| { |
| "epoch": 0.84, |
| "learning_rate": 3.9507259776993954e-08, |
| "logits/chosen": -3.7324581146240234, |
| "logits/rejected": -3.7278366088867188, |
| "logps/chosen": -1438.254150390625, |
| "logps/rejected": -1324.505615234375, |
| "loss": 0.6216, |
| "rewards/accuracies": 0.581250011920929, |
| "rewards/chosen": -1.5886811017990112, |
| "rewards/margins": 0.22386451065540314, |
| "rewards/rejected": -1.8125455379486084, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.84, |
| "eval_logits/chosen": -3.7396798133850098, |
| "eval_logits/rejected": -3.742741346359253, |
| "eval_logps/chosen": -1277.390869140625, |
| "eval_logps/rejected": -1193.6827392578125, |
| "eval_loss": 0.621441662311554, |
| "eval_rewards/accuracies": 0.65625, |
| "eval_rewards/chosen": -1.5152814388275146, |
| "eval_rewards/margins": 0.2620699107646942, |
| "eval_rewards/rejected": -1.7773513793945312, |
| "eval_runtime": 140.2367, |
| "eval_samples_per_second": 14.262, |
| "eval_steps_per_second": 0.228, |
| "step": 400 |
| }, |
| { |
| "epoch": 0.86, |
| "learning_rate": 3.022313472693447e-08, |
| "logits/chosen": -3.677812099456787, |
| "logits/rejected": -3.6684982776641846, |
| "logps/chosen": -1169.028076171875, |
| "logps/rejected": -1097.857666015625, |
| "loss": 0.63, |
| "rewards/accuracies": 0.606249988079071, |
| "rewards/chosen": -1.7634309530258179, |
| "rewards/margins": 0.23868277668952942, |
| "rewards/rejected": -2.0021133422851562, |
| "step": 410 |
| }, |
| { |
| "epoch": 0.88, |
| "learning_rate": 2.2111614344599684e-08, |
| "logits/chosen": -3.698087215423584, |
| "logits/rejected": -3.686516284942627, |
| "logps/chosen": -1219.0716552734375, |
| "logps/rejected": -1234.5465087890625, |
| "loss": 0.6267, |
| "rewards/accuracies": 0.643750011920929, |
| "rewards/chosen": -1.605316162109375, |
| "rewards/margins": 0.2227143496274948, |
| "rewards/rejected": -1.8280303478240967, |
| "step": 420 |
| }, |
| { |
| "epoch": 0.9, |
| "learning_rate": 1.521597710086439e-08, |
| "logits/chosen": -3.7486472129821777, |
| "logits/rejected": -3.74114727973938, |
| "logps/chosen": -1312.8095703125, |
| "logps/rejected": -1190.264892578125, |
| "loss": 0.6216, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": -1.5552055835723877, |
| "rewards/margins": 0.2965962290763855, |
| "rewards/rejected": -1.8518016338348389, |
| "step": 430 |
| }, |
| { |
| "epoch": 0.92, |
| "learning_rate": 9.57301420397924e-09, |
| "logits/chosen": -3.736569881439209, |
| "logits/rejected": -3.7111613750457764, |
| "logps/chosen": -1194.3740234375, |
| "logps/rejected": -1144.5274658203125, |
| "loss": 0.6122, |
| "rewards/accuracies": 0.6875, |
| "rewards/chosen": -1.60610032081604, |
| "rewards/margins": 0.2976963222026825, |
| "rewards/rejected": -1.9037965536117554, |
| "step": 440 |
| }, |
| { |
| "epoch": 0.94, |
| "learning_rate": 5.212833302556258e-09, |
| "logits/chosen": -3.6851935386657715, |
| "logits/rejected": -3.6797587871551514, |
| "logps/chosen": -1200.477783203125, |
| "logps/rejected": -1089.0234375, |
| "loss": 0.6297, |
| "rewards/accuracies": 0.6312500238418579, |
| "rewards/chosen": -1.5448455810546875, |
| "rewards/margins": 0.21296477317810059, |
| "rewards/rejected": -1.7578102350234985, |
| "step": 450 |
| }, |
| { |
| "epoch": 0.96, |
| "learning_rate": 2.158697848236607e-09, |
| "logits/chosen": -3.709555149078369, |
| "logits/rejected": -3.689227342605591, |
| "logps/chosen": -1132.578369140625, |
| "logps/rejected": -1029.9420166015625, |
| "loss": 0.6116, |
| "rewards/accuracies": 0.606249988079071, |
| "rewards/chosen": -1.6119111776351929, |
| "rewards/margins": 0.21641072630882263, |
| "rewards/rejected": -1.8283220529556274, |
| "step": 460 |
| }, |
| { |
| "epoch": 0.98, |
| "learning_rate": 4.269029751107489e-10, |
| "logits/chosen": -3.7561886310577393, |
| "logits/rejected": -3.7258613109588623, |
| "logps/chosen": -1350.369140625, |
| "logps/rejected": -1191.349853515625, |
| "loss": 0.6002, |
| "rewards/accuracies": 0.7437499761581421, |
| "rewards/chosen": -1.488236427307129, |
| "rewards/margins": 0.42441487312316895, |
| "rewards/rejected": -1.9126513004302979, |
| "step": 470 |
| }, |
| { |
| "epoch": 1.0, |
| "step": 478, |
| "total_flos": 0.0, |
| "train_loss": 0.6380818049778001, |
| "train_runtime": 9009.2405, |
| "train_samples_per_second": 6.786, |
| "train_steps_per_second": 0.053 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 478, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 1, |
| "save_steps": 1000, |
| "total_flos": 0.0, |
| "train_batch_size": 8, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
