| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 50, | |
| "global_step": 352, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.028409090909090908, | |
| "grad_norm": 38.70626469652959, | |
| "learning_rate": 1.3888888888888888e-07, | |
| "logits/chosen": -2.8592591285705566, | |
| "logits/rejected": -2.64298152923584, | |
| "logps/chosen": -390.552490234375, | |
| "logps/rejected": -607.8676147460938, | |
| "loss": 0.6889, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": 0.002072929637506604, | |
| "rewards/margins": 0.013672290369868279, | |
| "rewards/rejected": -0.011599360033869743, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.056818181818181816, | |
| "grad_norm": 19.199796929875, | |
| "learning_rate": 2.7777777777777776e-07, | |
| "logits/chosen": -2.8421695232391357, | |
| "logits/rejected": -2.6945552825927734, | |
| "logps/chosen": -328.68048095703125, | |
| "logps/rejected": -775.0625, | |
| "loss": 0.5324, | |
| "rewards/accuracies": 0.987500011920929, | |
| "rewards/chosen": 0.08464664220809937, | |
| "rewards/margins": 0.6608749032020569, | |
| "rewards/rejected": -0.5762282609939575, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.08522727272727272, | |
| "grad_norm": 5.168742449938278, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.85664439201355, | |
| "logits/rejected": -2.673206329345703, | |
| "logps/chosen": -295.9004821777344, | |
| "logps/rejected": -1121.4083251953125, | |
| "loss": 0.2036, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.3638155460357666, | |
| "rewards/margins": 4.79835319519043, | |
| "rewards/rejected": -4.434537887573242, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.11363636363636363, | |
| "grad_norm": 1.5631392125120753, | |
| "learning_rate": 4.998023493068254e-07, | |
| "logits/chosen": -2.8252196311950684, | |
| "logits/rejected": -2.627849578857422, | |
| "logps/chosen": -318.1038818359375, | |
| "logps/rejected": -2185.43017578125, | |
| "loss": 0.0507, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": 0.28651586174964905, | |
| "rewards/margins": 15.061859130859375, | |
| "rewards/rejected": -14.775341987609863, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.14204545454545456, | |
| "grad_norm": 0.27100229342130666, | |
| "learning_rate": 4.975823666181255e-07, | |
| "logits/chosen": -2.792081356048584, | |
| "logits/rejected": -2.4893665313720703, | |
| "logps/chosen": -460.21923828125, | |
| "logps/rejected": -4558.29833984375, | |
| "loss": 0.0066, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -0.9758400917053223, | |
| "rewards/margins": 38.06774139404297, | |
| "rewards/rejected": -39.043575286865234, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.14204545454545456, | |
| "eval_logits/chosen": -2.907456398010254, | |
| "eval_logits/rejected": -2.4200961589813232, | |
| "eval_logps/chosen": -484.4838562011719, | |
| "eval_logps/rejected": -5210.53955078125, | |
| "eval_loss": 0.0037060165777802467, | |
| "eval_rewards/accuracies": 0.9959677457809448, | |
| "eval_rewards/chosen": -1.1750215291976929, | |
| "eval_rewards/margins": 44.522483825683594, | |
| "eval_rewards/rejected": -45.697505950927734, | |
| "eval_runtime": 193.8738, | |
| "eval_samples_per_second": 20.142, | |
| "eval_steps_per_second": 0.32, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.17045454545454544, | |
| "grad_norm": 0.5320174620153905, | |
| "learning_rate": 4.929173350101024e-07, | |
| "logits/chosen": -2.949294090270996, | |
| "logits/rejected": -2.318241596221924, | |
| "logps/chosen": -495.1114196777344, | |
| "logps/rejected": -5547.14013671875, | |
| "loss": 0.0038, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -1.3521640300750732, | |
| "rewards/margins": 47.690696716308594, | |
| "rewards/rejected": -49.0428581237793, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.19886363636363635, | |
| "grad_norm": 0.14876040672715532, | |
| "learning_rate": 4.858533249305336e-07, | |
| "logits/chosen": -2.8789420127868652, | |
| "logits/rejected": -1.8734362125396729, | |
| "logps/chosen": -535.8917846679688, | |
| "logps/rejected": -5914.61865234375, | |
| "loss": 0.0032, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.5673115253448486, | |
| "rewards/margins": 51.4515266418457, | |
| "rewards/rejected": -53.018836975097656, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.22727272727272727, | |
| "grad_norm": 6.953457462206189, | |
| "learning_rate": 4.764600984163808e-07, | |
| "logits/chosen": -2.9963574409484863, | |
| "logits/rejected": -1.690744161605835, | |
| "logps/chosen": -566.8291015625, | |
| "logps/rejected": -7217.30712890625, | |
| "loss": 0.0032, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.6050817966461182, | |
| "rewards/margins": 63.83066940307617, | |
| "rewards/rejected": -65.43574523925781, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.2556818181818182, | |
| "grad_norm": 0.03391351087291168, | |
| "learning_rate": 4.6483042014491527e-07, | |
| "logits/chosen": -2.9498190879821777, | |
| "logits/rejected": -1.7918399572372437, | |
| "logps/chosen": -580.243408203125, | |
| "logps/rejected": -6469.92578125, | |
| "loss": 0.0113, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -2.181300640106201, | |
| "rewards/margins": 56.14410400390625, | |
| "rewards/rejected": -58.32539749145508, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.2840909090909091, | |
| "grad_norm": 0.39266433528373734, | |
| "learning_rate": 4.510791413176912e-07, | |
| "logits/chosen": -2.934478521347046, | |
| "logits/rejected": -0.735981285572052, | |
| "logps/chosen": -570.57568359375, | |
| "logps/rejected": -7235.8173828125, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.0110573768615723, | |
| "rewards/margins": 63.289306640625, | |
| "rewards/rejected": -65.30036926269531, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.2840909090909091, | |
| "eval_logits/chosen": -2.7790369987487793, | |
| "eval_logits/rejected": -0.2035539597272873, | |
| "eval_logps/chosen": -556.77392578125, | |
| "eval_logps/rejected": -6926.5537109375, | |
| "eval_loss": 0.0003049192891921848, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -1.897922158241272, | |
| "eval_rewards/margins": 60.95973205566406, | |
| "eval_rewards/rejected": -62.85765838623047, | |
| "eval_runtime": 193.4087, | |
| "eval_samples_per_second": 20.19, | |
| "eval_steps_per_second": 0.321, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3125, | |
| "grad_norm": 0.0360149995259237, | |
| "learning_rate": 4.353420654246546e-07, | |
| "logits/chosen": -2.6916072368621826, | |
| "logits/rejected": -0.07679282128810883, | |
| "logps/chosen": -566.1495361328125, | |
| "logps/rejected": -6591.94287109375, | |
| "loss": 0.0025, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.907283067703247, | |
| "rewards/margins": 58.01946258544922, | |
| "rewards/rejected": -59.92673873901367, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.3409090909090909, | |
| "grad_norm": 0.05338636863642817, | |
| "learning_rate": 4.177746070897592e-07, | |
| "logits/chosen": -2.5057804584503174, | |
| "logits/rejected": 0.8570491671562195, | |
| "logps/chosen": -578.2913818359375, | |
| "logps/rejected": -7127.86328125, | |
| "loss": 0.004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.9677445888519287, | |
| "rewards/margins": 62.9386100769043, | |
| "rewards/rejected": -64.90635681152344, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.3693181818181818, | |
| "grad_norm": 0.28121189018180337, | |
| "learning_rate": 3.9855025724292763e-07, | |
| "logits/chosen": -2.5544893741607666, | |
| "logits/rejected": 1.7402244806289673, | |
| "logps/chosen": -602.7317504882812, | |
| "logps/rejected": -7376.546875, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.3228020668029785, | |
| "rewards/margins": 65.03727722167969, | |
| "rewards/rejected": -67.3600845336914, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.3977272727272727, | |
| "grad_norm": 0.7726944716269485, | |
| "learning_rate": 3.7785886977585555e-07, | |
| "logits/chosen": -2.5471339225769043, | |
| "logits/rejected": 1.9610217809677124, | |
| "logps/chosen": -578.9056396484375, | |
| "logps/rejected": -8475.4873046875, | |
| "loss": 0.0019, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.1286935806274414, | |
| "rewards/margins": 76.206298828125, | |
| "rewards/rejected": -78.3349838256836, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.42613636363636365, | |
| "grad_norm": 0.5096412009910095, | |
| "learning_rate": 3.5590478660213206e-07, | |
| "logits/chosen": -2.5447940826416016, | |
| "logits/rejected": 1.8056743144989014, | |
| "logps/chosen": -599.0574340820312, | |
| "logps/rejected": -7813.42333984375, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.2099416255950928, | |
| "rewards/margins": 68.80451965332031, | |
| "rewards/rejected": -71.01445007324219, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.42613636363636365, | |
| "eval_logits/chosen": -2.655405044555664, | |
| "eval_logits/rejected": 1.886979103088379, | |
| "eval_logps/chosen": -556.2059326171875, | |
| "eval_logps/rejected": -7429.90234375, | |
| "eval_loss": 0.0001985544804483652, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -1.892242193222046, | |
| "eval_rewards/margins": 65.99889373779297, | |
| "eval_rewards/rejected": -67.8911361694336, | |
| "eval_runtime": 192.529, | |
| "eval_samples_per_second": 20.283, | |
| "eval_steps_per_second": 0.322, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.45454545454545453, | |
| "grad_norm": 2.608953110629928, | |
| "learning_rate": 3.3290481963801696e-07, | |
| "logits/chosen": -2.655552387237549, | |
| "logits/rejected": 2.1644370555877686, | |
| "logps/chosen": -535.6851806640625, | |
| "logps/rejected": -7245.27734375, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.6554721593856812, | |
| "rewards/margins": 64.63240051269531, | |
| "rewards/rejected": -66.28787231445312, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.48295454545454547, | |
| "grad_norm": 0.0342477427475356, | |
| "learning_rate": 3.0908610963322626e-07, | |
| "logits/chosen": -2.3891050815582275, | |
| "logits/rejected": 2.3125836849212646, | |
| "logps/chosen": -601.9050903320312, | |
| "logps/rejected": -8139.94921875, | |
| "loss": 0.0053, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.1428682804107666, | |
| "rewards/margins": 71.97351837158203, | |
| "rewards/rejected": -74.11639404296875, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.5113636363636364, | |
| "grad_norm": 0.011995212283351854, | |
| "learning_rate": 2.846838829972671e-07, | |
| "logits/chosen": -2.39284610748291, | |
| "logits/rejected": 2.5177011489868164, | |
| "logps/chosen": -605.0557861328125, | |
| "logps/rejected": -7604.11328125, | |
| "loss": 0.0022, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.441514492034912, | |
| "rewards/margins": 67.46978759765625, | |
| "rewards/rejected": -69.91130065917969, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.5397727272727273, | |
| "grad_norm": 0.17953481614212255, | |
| "learning_rate": 2.5993912877423147e-07, | |
| "logits/chosen": -2.4631476402282715, | |
| "logits/rejected": 1.9089019298553467, | |
| "logps/chosen": -553.119140625, | |
| "logps/rejected": -8167.6572265625, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.075033664703369, | |
| "rewards/margins": 72.6854248046875, | |
| "rewards/rejected": -74.76045989990234, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.5681818181818182, | |
| "grad_norm": 0.7077913782197723, | |
| "learning_rate": 2.3509621870754504e-07, | |
| "logits/chosen": -2.350708484649658, | |
| "logits/rejected": 2.6935336589813232, | |
| "logps/chosen": -562.2257080078125, | |
| "logps/rejected": -7312.64453125, | |
| "loss": 0.0013, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.970958948135376, | |
| "rewards/margins": 64.8470687866211, | |
| "rewards/rejected": -66.81802368164062, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5681818181818182, | |
| "eval_logits/chosen": -2.2775018215179443, | |
| "eval_logits/rejected": 2.429776430130005, | |
| "eval_logps/chosen": -582.5308227539062, | |
| "eval_logps/rejected": -8019.7568359375, | |
| "eval_loss": 0.00011486691073514521, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -2.155491352081299, | |
| "eval_rewards/margins": 71.63419342041016, | |
| "eval_rewards/rejected": -73.78968811035156, | |
| "eval_runtime": 192.7674, | |
| "eval_samples_per_second": 20.258, | |
| "eval_steps_per_second": 0.322, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.5965909090909091, | |
| "grad_norm": 0.24855740640565924, | |
| "learning_rate": 2.1040049389819624e-07, | |
| "logits/chosen": -2.1356868743896484, | |
| "logits/rejected": 2.5835635662078857, | |
| "logps/chosen": -601.9248046875, | |
| "logps/rejected": -7937.8779296875, | |
| "loss": 0.0034, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.1128151416778564, | |
| "rewards/margins": 70.94691467285156, | |
| "rewards/rejected": -73.05973052978516, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.625, | |
| "grad_norm": 1.3430282073648296, | |
| "learning_rate": 1.8609584188988133e-07, | |
| "logits/chosen": -1.6919233798980713, | |
| "logits/rejected": 2.1380715370178223, | |
| "logps/chosen": -634.4395751953125, | |
| "logps/rejected": -7114.49462890625, | |
| "loss": 0.0034, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.6983890533447266, | |
| "rewards/margins": 61.95018768310547, | |
| "rewards/rejected": -64.64857482910156, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.6534090909090909, | |
| "grad_norm": 35.7863072981701, | |
| "learning_rate": 1.624222881090439e-07, | |
| "logits/chosen": -1.9383169412612915, | |
| "logits/rejected": 1.8241441249847412, | |
| "logps/chosen": -657.0816040039062, | |
| "logps/rejected": -8359.109375, | |
| "loss": 0.0111, | |
| "rewards/accuracies": 0.9937499761581421, | |
| "rewards/chosen": -2.6241703033447266, | |
| "rewards/margins": 74.1111068725586, | |
| "rewards/rejected": -76.73527526855469, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.6818181818181818, | |
| "grad_norm": 0.3968538526472747, | |
| "learning_rate": 1.3961362544602212e-07, | |
| "logits/chosen": -1.7102636098861694, | |
| "logits/rejected": 2.2134716510772705, | |
| "logps/chosen": -606.8711547851562, | |
| "logps/rejected": -7288.10302734375, | |
| "loss": 0.001, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.3085877895355225, | |
| "rewards/margins": 64.6317138671875, | |
| "rewards/rejected": -66.94029998779297, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.7102272727272727, | |
| "grad_norm": 0.345391461678324, | |
| "learning_rate": 1.1789510538684522e-07, | |
| "logits/chosen": -1.7613790035247803, | |
| "logits/rejected": 2.189953088760376, | |
| "logps/chosen": -568.7935791015625, | |
| "logps/rejected": -9133.8701171875, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.16759991645813, | |
| "rewards/margins": 81.87528228759766, | |
| "rewards/rejected": -84.04288482666016, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.7102272727272727, | |
| "eval_logits/chosen": -1.6277003288269043, | |
| "eval_logits/rejected": 2.342982769012451, | |
| "eval_logps/chosen": -586.8257446289062, | |
| "eval_logps/rejected": -7875.4423828125, | |
| "eval_loss": 9.437829430680722e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -2.198439836502075, | |
| "eval_rewards/margins": 70.1480941772461, | |
| "eval_rewards/rejected": -72.34652709960938, | |
| "eval_runtime": 192.1809, | |
| "eval_samples_per_second": 20.319, | |
| "eval_steps_per_second": 0.323, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.7386363636363636, | |
| "grad_norm": 0.30273821428387077, | |
| "learning_rate": 9.748121349736891e-08, | |
| "logits/chosen": -1.565211534500122, | |
| "logits/rejected": 2.2146430015563965, | |
| "logps/chosen": -623.1561279296875, | |
| "logps/rejected": -8230.65625, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.393279552459717, | |
| "rewards/margins": 73.13200378417969, | |
| "rewards/rejected": -75.52528381347656, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.7670454545454546, | |
| "grad_norm": 0.03514974255734628, | |
| "learning_rate": 7.857355122839673e-08, | |
| "logits/chosen": -1.7121340036392212, | |
| "logits/rejected": 2.091160535812378, | |
| "logps/chosen": -599.9885864257812, | |
| "logps/rejected": -7744.3994140625, | |
| "loss": 0.0007, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.191390037536621, | |
| "rewards/margins": 68.56300354003906, | |
| "rewards/rejected": -70.75439453125, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.7954545454545454, | |
| "grad_norm": 0.1441611887831923, | |
| "learning_rate": 6.135884496044244e-08, | |
| "logits/chosen": -1.670058012008667, | |
| "logits/rejected": 2.3099844455718994, | |
| "logps/chosen": -586.2249145507812, | |
| "logps/rejected": -7608.43212890625, | |
| "loss": 0.0011, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.070519208908081, | |
| "rewards/margins": 67.56490325927734, | |
| "rewards/rejected": -69.63542175292969, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.8238636363636364, | |
| "grad_norm": 0.02043160544641207, | |
| "learning_rate": 4.600710195020982e-08, | |
| "logits/chosen": -1.508570909500122, | |
| "logits/rejected": 2.35508131980896, | |
| "logps/chosen": -602.7320556640625, | |
| "logps/rejected": -7490.234375, | |
| "loss": 0.0026, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.135887622833252, | |
| "rewards/margins": 66.35386657714844, | |
| "rewards/rejected": -68.48976135253906, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.8522727272727273, | |
| "grad_norm": 0.6511993649543483, | |
| "learning_rate": 3.2669931390104374e-08, | |
| "logits/chosen": -1.581496000289917, | |
| "logits/rejected": 2.216219186782837, | |
| "logps/chosen": -553.2108154296875, | |
| "logps/rejected": -8122.2275390625, | |
| "loss": 0.0009, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.969562292098999, | |
| "rewards/margins": 73.0444564819336, | |
| "rewards/rejected": -75.0140151977539, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.8522727272727273, | |
| "eval_logits/chosen": -1.4073354005813599, | |
| "eval_logits/rejected": 2.322908639907837, | |
| "eval_logps/chosen": -579.916259765625, | |
| "eval_logps/rejected": -7770.10595703125, | |
| "eval_loss": 6.592821591766551e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -2.129345178604126, | |
| "eval_rewards/margins": 69.1638412475586, | |
| "eval_rewards/rejected": -71.29317474365234, | |
| "eval_runtime": 193.1056, | |
| "eval_samples_per_second": 20.222, | |
| "eval_steps_per_second": 0.321, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.8806818181818182, | |
| "grad_norm": 0.14250781664198195, | |
| "learning_rate": 2.147904716149135e-08, | |
| "logits/chosen": -1.3581085205078125, | |
| "logits/rejected": 2.2543134689331055, | |
| "logps/chosen": -562.1097412109375, | |
| "logps/rejected": -7898.74072265625, | |
| "loss": 0.0016, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.0004312992095947, | |
| "rewards/margins": 69.9819564819336, | |
| "rewards/rejected": -71.98239135742188, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "grad_norm": 0.015240907682121196, | |
| "learning_rate": 1.254496706805433e-08, | |
| "logits/chosen": -1.5214873552322388, | |
| "logits/rejected": 2.192413806915283, | |
| "logps/chosen": -591.5843505859375, | |
| "logps/rejected": -7921.0029296875, | |
| "loss": 0.0005, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.089113235473633, | |
| "rewards/margins": 70.60108947753906, | |
| "rewards/rejected": -72.69020080566406, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.9375, | |
| "grad_norm": 0.1617057843470206, | |
| "learning_rate": 5.955921395237318e-09, | |
| "logits/chosen": -1.4116395711898804, | |
| "logits/rejected": 2.1853275299072266, | |
| "logps/chosen": -551.7928466796875, | |
| "logps/rejected": -7996.5966796875, | |
| "loss": 0.0004, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.9652475118637085, | |
| "rewards/margins": 71.52317810058594, | |
| "rewards/rejected": -73.4884262084961, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.9659090909090909, | |
| "grad_norm": 0.18410596009005747, | |
| "learning_rate": 1.7769815745066474e-09, | |
| "logits/chosen": -1.6253635883331299, | |
| "logits/rejected": 2.031024217605591, | |
| "logps/chosen": -567.4468994140625, | |
| "logps/rejected": -7287.84130859375, | |
| "loss": 0.0006, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -1.9846607446670532, | |
| "rewards/margins": 64.80608367919922, | |
| "rewards/rejected": -66.79075622558594, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.9943181818181818, | |
| "grad_norm": 1.9094053325335274, | |
| "learning_rate": 4.9417557483610875e-11, | |
| "logits/chosen": -1.3966957330703735, | |
| "logits/rejected": 2.161142587661743, | |
| "logps/chosen": -596.6432495117188, | |
| "logps/rejected": -7771.0029296875, | |
| "loss": 0.0015, | |
| "rewards/accuracies": 1.0, | |
| "rewards/chosen": -2.2237548828125, | |
| "rewards/margins": 69.01069641113281, | |
| "rewards/rejected": -71.23445129394531, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.9943181818181818, | |
| "eval_logits/chosen": -1.3155968189239502, | |
| "eval_logits/rejected": 2.341789960861206, | |
| "eval_logps/chosen": -582.794677734375, | |
| "eval_logps/rejected": -7854.4189453125, | |
| "eval_loss": 6.306579598458484e-05, | |
| "eval_rewards/accuracies": 1.0, | |
| "eval_rewards/chosen": -2.1581296920776367, | |
| "eval_rewards/margins": 69.9781723022461, | |
| "eval_rewards/rejected": -72.13629150390625, | |
| "eval_runtime": 193.0578, | |
| "eval_samples_per_second": 20.227, | |
| "eval_steps_per_second": 0.321, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 352, | |
| "total_flos": 0.0, | |
| "train_loss": 0.04425002944373608, | |
| "train_runtime": 10145.5361, | |
| "train_samples_per_second": 4.435, | |
| "train_steps_per_second": 0.035 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 352, | |
| "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 | |
| } | |