| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 500, | |
| "global_step": 216, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 8.33483565345429, | |
| "learning_rate": 2.2727272727272725e-08, | |
| "logits/chosen": -0.9294053316116333, | |
| "logits/rejected": -1.2306766510009766, | |
| "logps/chosen": -79.32324981689453, | |
| "logps/rejected": -124.87494659423828, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 8.957121309516765, | |
| "learning_rate": 2.2727272727272726e-07, | |
| "logits/chosen": -0.5779850482940674, | |
| "logits/rejected": -1.1789488792419434, | |
| "logps/chosen": -83.05654907226562, | |
| "logps/rejected": -135.26220703125, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.4930555522441864, | |
| "rewards/chosen": 0.0003059397859033197, | |
| "rewards/margins": 0.00026924433768726885, | |
| "rewards/rejected": 3.6695546441478655e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 8.37389940695367, | |
| "learning_rate": 4.545454545454545e-07, | |
| "logits/chosen": -0.621688723564148, | |
| "logits/rejected": -1.1858972311019897, | |
| "logps/chosen": -95.2262954711914, | |
| "logps/rejected": -143.64205932617188, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.48124998807907104, | |
| "rewards/chosen": -0.0002137714036507532, | |
| "rewards/margins": -0.0006433525122702122, | |
| "rewards/rejected": 0.00042958109406754375, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 8.286301306160889, | |
| "learning_rate": 4.979050253066063e-07, | |
| "logits/chosen": -0.6890290975570679, | |
| "logits/rejected": -1.1448113918304443, | |
| "logps/chosen": -88.85906982421875, | |
| "logps/rejected": -132.80384826660156, | |
| "loss": 0.6917, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.0019540651701390743, | |
| "rewards/margins": 0.00375129166059196, | |
| "rewards/rejected": -0.0017972266068682075, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "grad_norm": 8.434509757435025, | |
| "learning_rate": 4.894543310469967e-07, | |
| "logits/chosen": -0.5220860242843628, | |
| "logits/rejected": -1.1176507472991943, | |
| "logps/chosen": -91.24484252929688, | |
| "logps/rejected": -126.5925521850586, | |
| "loss": 0.6902, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.0025604977272450924, | |
| "rewards/margins": 0.005932547152042389, | |
| "rewards/rejected": -0.003372048959136009, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 8.535183437584605, | |
| "learning_rate": 4.747379352713488e-07, | |
| "logits/chosen": -0.685869574546814, | |
| "logits/rejected": -1.2057311534881592, | |
| "logps/chosen": -96.0764389038086, | |
| "logps/rejected": -127.81497955322266, | |
| "loss": 0.6884, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": 0.003552153008058667, | |
| "rewards/margins": 0.010397722944617271, | |
| "rewards/rejected": -0.006845570169389248, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "grad_norm": 8.190633609066436, | |
| "learning_rate": 4.541409157643027e-07, | |
| "logits/chosen": -0.6526767015457153, | |
| "logits/rejected": -1.1397744417190552, | |
| "logps/chosen": -89.37004089355469, | |
| "logps/rejected": -124.7223129272461, | |
| "loss": 0.6851, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": 0.0007335222908295691, | |
| "rewards/margins": 0.016615714877843857, | |
| "rewards/rejected": -0.01588219217956066, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 8.847754167480884, | |
| "learning_rate": 4.282022261367073e-07, | |
| "logits/chosen": -0.4965807795524597, | |
| "logits/rejected": -1.2134766578674316, | |
| "logps/chosen": -86.97933959960938, | |
| "logps/rejected": -134.1288299560547, | |
| "loss": 0.6824, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": 0.0005543470033444464, | |
| "rewards/margins": 0.023061692714691162, | |
| "rewards/rejected": -0.02250734530389309, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "grad_norm": 8.841125645791609, | |
| "learning_rate": 3.9760059325148063e-07, | |
| "logits/chosen": -0.6324743032455444, | |
| "logits/rejected": -1.147215485572815, | |
| "logps/chosen": -91.2511978149414, | |
| "logps/rejected": -149.07229614257812, | |
| "loss": 0.6791, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": 0.0010622118134051561, | |
| "rewards/margins": 0.032594919204711914, | |
| "rewards/rejected": -0.03153270110487938, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "grad_norm": 8.826466564074398, | |
| "learning_rate": 3.6313675726113475e-07, | |
| "logits/chosen": -0.6752112507820129, | |
| "logits/rejected": -1.1697431802749634, | |
| "logps/chosen": -83.96028900146484, | |
| "logps/rejected": -135.6808624267578, | |
| "loss": 0.6766, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -0.007847856730222702, | |
| "rewards/margins": 0.033581167459487915, | |
| "rewards/rejected": -0.04142902046442032, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 8.986918539516395, | |
| "learning_rate": 3.2571251897448763e-07, | |
| "logits/chosen": -0.7318952083587646, | |
| "logits/rejected": -1.1481754779815674, | |
| "logps/chosen": -104.59893798828125, | |
| "logps/rejected": -132.4669952392578, | |
| "loss": 0.6729, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.011359195224940777, | |
| "rewards/margins": 0.043643511831760406, | |
| "rewards/rejected": -0.055002711713314056, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "grad_norm": 8.709487324526142, | |
| "learning_rate": 2.863071428113726e-07, | |
| "logits/chosen": -0.6006079316139221, | |
| "logits/rejected": -1.0975641012191772, | |
| "logps/chosen": -84.71573638916016, | |
| "logps/rejected": -149.96249389648438, | |
| "loss": 0.6686, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.0164839755743742, | |
| "rewards/margins": 0.04694540053606033, | |
| "rewards/rejected": -0.06342937797307968, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "grad_norm": 9.237752042690602, | |
| "learning_rate": 2.459517327993746e-07, | |
| "logits/chosen": -0.6878473162651062, | |
| "logits/rejected": -1.3123815059661865, | |
| "logps/chosen": -90.21903991699219, | |
| "logps/rejected": -151.0259552001953, | |
| "loss": 0.6615, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": -0.019475463777780533, | |
| "rewards/margins": 0.07015477865934372, | |
| "rewards/rejected": -0.08963023871183395, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "grad_norm": 9.286029411981152, | |
| "learning_rate": 2.0570225210519433e-07, | |
| "logits/chosen": -0.790011465549469, | |
| "logits/rejected": -1.2298551797866821, | |
| "logps/chosen": -93.1588363647461, | |
| "logps/rejected": -151.1782684326172, | |
| "loss": 0.661, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.030178239569067955, | |
| "rewards/margins": 0.07631012797355652, | |
| "rewards/rejected": -0.10648836940526962, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "grad_norm": 10.19246601057813, | |
| "learning_rate": 1.6661189208729489e-07, | |
| "logits/chosen": -0.6443529725074768, | |
| "logits/rejected": -1.133919596672058, | |
| "logps/chosen": -84.59071350097656, | |
| "logps/rejected": -138.85736083984375, | |
| "loss": 0.6579, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.04141147807240486, | |
| "rewards/margins": 0.07676441967487335, | |
| "rewards/rejected": -0.11817590147256851, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "grad_norm": 9.407025901804323, | |
| "learning_rate": 1.2970351387729872e-07, | |
| "logits/chosen": -0.73204106092453, | |
| "logits/rejected": -1.1557518243789673, | |
| "logps/chosen": -88.95694732666016, | |
| "logps/rejected": -131.26490783691406, | |
| "loss": 0.652, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -0.03920191153883934, | |
| "rewards/margins": 0.08074145019054413, | |
| "rewards/rejected": -0.11994334310293198, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "grad_norm": 8.654405006026039, | |
| "learning_rate": 9.594288359976815e-08, | |
| "logits/chosen": -0.7579759359359741, | |
| "logits/rejected": -1.263896107673645, | |
| "logps/chosen": -86.85200500488281, | |
| "logps/rejected": -138.72409057617188, | |
| "loss": 0.6501, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.047876037657260895, | |
| "rewards/margins": 0.09631510823965073, | |
| "rewards/rejected": -0.14419114589691162, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "grad_norm": 8.716288122934651, | |
| "learning_rate": 6.621340157319996e-08, | |
| "logits/chosen": -0.736527681350708, | |
| "logits/rejected": -1.3233287334442139, | |
| "logps/chosen": -99.1356430053711, | |
| "logps/rejected": -159.6076202392578, | |
| "loss": 0.6487, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -0.06683658063411713, | |
| "rewards/margins": 0.10212481021881104, | |
| "rewards/rejected": -0.16896137595176697, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "grad_norm": 9.622323797016332, | |
| "learning_rate": 4.1292986742682254e-08, | |
| "logits/chosen": -0.7446157932281494, | |
| "logits/rejected": -1.380061149597168, | |
| "logps/chosen": -96.81314849853516, | |
| "logps/rejected": -152.16970825195312, | |
| "loss": 0.6474, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.06702615320682526, | |
| "rewards/margins": 0.1122649684548378, | |
| "rewards/rejected": -0.17929109930992126, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "grad_norm": 10.2499777892521, | |
| "learning_rate": 2.183372119961499e-08, | |
| "logits/chosen": -0.695759117603302, | |
| "logits/rejected": -1.323012351989746, | |
| "logps/chosen": -100.9570541381836, | |
| "logps/rejected": -148.34474182128906, | |
| "loss": 0.6454, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.07095570862293243, | |
| "rewards/margins": 0.09609047323465347, | |
| "rewards/rejected": -0.1670461893081665, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 9.240245078846256, | |
| "learning_rate": 8.344787421847216e-09, | |
| "logits/chosen": -0.7457176446914673, | |
| "logits/rejected": -1.3048522472381592, | |
| "logps/chosen": -97.51776885986328, | |
| "logps/rejected": -142.3980712890625, | |
| "loss": 0.6493, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.07124415785074234, | |
| "rewards/margins": 0.09479977190494537, | |
| "rewards/rejected": -0.16604390740394592, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "grad_norm": 9.838380839286996, | |
| "learning_rate": 1.1791447083465133e-09, | |
| "logits/chosen": -0.8287758827209473, | |
| "logits/rejected": -1.403287649154663, | |
| "logps/chosen": -104.15522766113281, | |
| "logps/rejected": -157.41732788085938, | |
| "loss": 0.6432, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.08230911195278168, | |
| "rewards/margins": 0.10910389572381973, | |
| "rewards/rejected": -0.191413015127182, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 216, | |
| "total_flos": 0.0, | |
| "train_loss": 0.6676328546471066, | |
| "train_runtime": 2437.1919, | |
| "train_samples_per_second": 5.672, | |
| "train_steps_per_second": 0.089 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 216, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 500, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |