| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.983957219251337, | |
| "eval_steps": 50, | |
| "global_step": 279, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.053475935828877004, | |
| "grad_norm": 57.38362641680358, | |
| "learning_rate": 5e-07, | |
| "logits/chosen": -2.734001874923706, | |
| "logits/rejected": -2.714400053024292, | |
| "logps/chosen": -259.45416259765625, | |
| "logps/rejected": -213.60086059570312, | |
| "loss": 0.69, | |
| "rewards/accuracies": 0.3687500059604645, | |
| "rewards/chosen": 0.022585459053516388, | |
| "rewards/margins": 0.012519368901848793, | |
| "rewards/rejected": 0.010066090151667595, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.10695187165775401, | |
| "grad_norm": 47.481455328039516, | |
| "learning_rate": 1e-06, | |
| "logits/chosen": -2.6481270790100098, | |
| "logits/rejected": -2.652583599090576, | |
| "logps/chosen": -257.953369140625, | |
| "logps/rejected": -188.35134887695312, | |
| "loss": 0.6387, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.550362229347229, | |
| "rewards/margins": 0.1817861795425415, | |
| "rewards/rejected": 0.3685761094093323, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.16042780748663102, | |
| "grad_norm": 57.27168970604102, | |
| "learning_rate": 9.991477798614637e-07, | |
| "logits/chosen": -2.517893075942993, | |
| "logits/rejected": -2.512808322906494, | |
| "logps/chosen": -238.94729614257812, | |
| "logps/rejected": -195.82278442382812, | |
| "loss": 0.6343, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 1.433282732963562, | |
| "rewards/margins": 0.7576474547386169, | |
| "rewards/rejected": 0.6756354570388794, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.21390374331550802, | |
| "grad_norm": 49.36931799258527, | |
| "learning_rate": 9.965940245625131e-07, | |
| "logits/chosen": -2.3814330101013184, | |
| "logits/rejected": -2.3650126457214355, | |
| "logps/chosen": -240.35299682617188, | |
| "logps/rejected": -224.2460174560547, | |
| "loss": 0.6444, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 1.3586658239364624, | |
| "rewards/margins": 0.8488560914993286, | |
| "rewards/rejected": 0.5098099112510681, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.26737967914438504, | |
| "grad_norm": 38.23771695176387, | |
| "learning_rate": 9.923474395499264e-07, | |
| "logits/chosen": -2.3290395736694336, | |
| "logits/rejected": -2.300835371017456, | |
| "logps/chosen": -240.2759246826172, | |
| "logps/rejected": -190.5952911376953, | |
| "loss": 0.5839, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": 1.567830204963684, | |
| "rewards/margins": 1.0694384574890137, | |
| "rewards/rejected": 0.4983917772769928, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.32085561497326204, | |
| "grad_norm": 38.55306482248171, | |
| "learning_rate": 9.86422500924775e-07, | |
| "logits/chosen": -2.385282039642334, | |
| "logits/rejected": -2.380516767501831, | |
| "logps/chosen": -237.0162353515625, | |
| "logps/rejected": -203.95974731445312, | |
| "loss": 0.5888, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 1.3996493816375732, | |
| "rewards/margins": 0.9638010263442993, | |
| "rewards/rejected": 0.43584829568862915, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.37433155080213903, | |
| "grad_norm": 40.05177774617033, | |
| "learning_rate": 9.788394060951227e-07, | |
| "logits/chosen": -2.4351038932800293, | |
| "logits/rejected": -2.440431833267212, | |
| "logps/chosen": -247.30227661132812, | |
| "logps/rejected": -201.399658203125, | |
| "loss": 0.6124, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 1.2308628559112549, | |
| "rewards/margins": 1.0766284465789795, | |
| "rewards/rejected": 0.15423443913459778, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.42780748663101603, | |
| "grad_norm": 38.64888093844348, | |
| "learning_rate": 9.696240049254742e-07, | |
| "logits/chosen": -2.4633519649505615, | |
| "logits/rejected": -2.466259241104126, | |
| "logps/chosen": -249.7180938720703, | |
| "logps/rejected": -194.3883056640625, | |
| "loss": 0.5959, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 1.1658284664154053, | |
| "rewards/margins": 1.0980390310287476, | |
| "rewards/rejected": 0.06778934597969055, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.48128342245989303, | |
| "grad_norm": 41.079430965451316, | |
| "learning_rate": 9.588077116176756e-07, | |
| "logits/chosen": -2.4081149101257324, | |
| "logits/rejected": -2.406311273574829, | |
| "logps/chosen": -238.9816436767578, | |
| "logps/rejected": -201.0675048828125, | |
| "loss": 0.6076, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 1.1393609046936035, | |
| "rewards/margins": 1.4659656286239624, | |
| "rewards/rejected": -0.32660484313964844, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.5347593582887701, | |
| "grad_norm": 46.70159372711324, | |
| "learning_rate": 9.464273976236516e-07, | |
| "logits/chosen": -2.4118716716766357, | |
| "logits/rejected": -2.408468246459961, | |
| "logps/chosen": -219.6370849609375, | |
| "logps/rejected": -186.2394561767578, | |
| "loss": 0.5857, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.8120909929275513, | |
| "rewards/margins": 0.906074047088623, | |
| "rewards/rejected": -0.09398309141397476, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.5347593582887701, | |
| "eval_logits/chosen": -2.4070043563842773, | |
| "eval_logits/rejected": -2.395042896270752, | |
| "eval_logps/chosen": -246.7726287841797, | |
| "eval_logps/rejected": -220.63446044921875, | |
| "eval_loss": 0.5587548613548279, | |
| "eval_rewards/accuracies": 0.7202380895614624, | |
| "eval_rewards/chosen": 0.8626245856285095, | |
| "eval_rewards/margins": 1.329900860786438, | |
| "eval_rewards/rejected": -0.46727630496025085, | |
| "eval_runtime": 180.0372, | |
| "eval_samples_per_second": 14.775, | |
| "eval_steps_per_second": 0.233, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.5882352941176471, | |
| "grad_norm": 38.079371712676014, | |
| "learning_rate": 9.325252659550308e-07, | |
| "logits/chosen": -2.387826919555664, | |
| "logits/rejected": -2.3770289421081543, | |
| "logps/chosen": -226.2966766357422, | |
| "logps/rejected": -208.65219116210938, | |
| "loss": 0.5977, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.6860524415969849, | |
| "rewards/margins": 0.8900126218795776, | |
| "rewards/rejected": -0.2039601057767868, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.6417112299465241, | |
| "grad_norm": 37.176849832612625, | |
| "learning_rate": 9.171487073181197e-07, | |
| "logits/chosen": -2.309826374053955, | |
| "logits/rejected": -2.319997549057007, | |
| "logps/chosen": -230.22189331054688, | |
| "logps/rejected": -206.3903350830078, | |
| "loss": 0.5777, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.7191376686096191, | |
| "rewards/margins": 1.5079154968261719, | |
| "rewards/rejected": -0.7887779474258423, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.6951871657754011, | |
| "grad_norm": 36.39927326888579, | |
| "learning_rate": 9.003501385646448e-07, | |
| "logits/chosen": -2.2107081413269043, | |
| "logits/rejected": -2.21248722076416, | |
| "logps/chosen": -238.040771484375, | |
| "logps/rejected": -207.3212127685547, | |
| "loss": 0.5854, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.5230123996734619, | |
| "rewards/margins": 1.2387675046920776, | |
| "rewards/rejected": -0.715755045413971, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.7486631016042781, | |
| "grad_norm": 34.98626076335809, | |
| "learning_rate": 8.821868240089676e-07, | |
| "logits/chosen": -2.1564557552337646, | |
| "logits/rejected": -2.1244616508483887, | |
| "logps/chosen": -232.23452758789062, | |
| "logps/rejected": -215.52719116210938, | |
| "loss": 0.5572, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": 1.0822027921676636, | |
| "rewards/margins": 1.5664498805999756, | |
| "rewards/rejected": -0.48424673080444336, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.8021390374331551, | |
| "grad_norm": 33.602724500864966, | |
| "learning_rate": 8.62720680220876e-07, | |
| "logits/chosen": -2.2200913429260254, | |
| "logits/rejected": -2.161189079284668, | |
| "logps/chosen": -232.7351837158203, | |
| "logps/rejected": -218.43045043945312, | |
| "loss": 0.5712, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.5824815630912781, | |
| "rewards/margins": 1.2068629264831543, | |
| "rewards/rejected": -0.6243813037872314, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.8556149732620321, | |
| "grad_norm": 32.25811213479902, | |
| "learning_rate": 8.420180649593929e-07, | |
| "logits/chosen": -2.223334789276123, | |
| "logits/rejected": -2.206092119216919, | |
| "logps/chosen": -237.7117156982422, | |
| "logps/rejected": -213.7598114013672, | |
| "loss": 0.5111, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": 0.454035222530365, | |
| "rewards/margins": 1.7620937824249268, | |
| "rewards/rejected": -1.3080583810806274, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "grad_norm": 28.217025801566006, | |
| "learning_rate": 8.201495509671036e-07, | |
| "logits/chosen": -2.205242872238159, | |
| "logits/rejected": -2.2124757766723633, | |
| "logps/chosen": -250.62673950195312, | |
| "logps/rejected": -233.39517211914062, | |
| "loss": 0.5286, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.5412265658378601, | |
| "rewards/margins": 1.507828712463379, | |
| "rewards/rejected": -0.9666021466255188, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 0.9625668449197861, | |
| "grad_norm": 33.081462718842836, | |
| "learning_rate": 7.971896853961042e-07, | |
| "logits/chosen": -2.286748170852661, | |
| "logits/rejected": -2.2566237449645996, | |
| "logps/chosen": -235.38076782226562, | |
| "logps/rejected": -223.08413696289062, | |
| "loss": 0.5372, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 1.0189321041107178, | |
| "rewards/margins": 1.5707125663757324, | |
| "rewards/rejected": -0.5517805814743042, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 1.0160427807486632, | |
| "grad_norm": 19.031150325303443, | |
| "learning_rate": 7.732167356856654e-07, | |
| "logits/chosen": -2.383920192718506, | |
| "logits/rejected": -2.3609161376953125, | |
| "logps/chosen": -249.6156768798828, | |
| "logps/rejected": -201.63851928710938, | |
| "loss": 0.3778, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 1.3106629848480225, | |
| "rewards/margins": 2.160562038421631, | |
| "rewards/rejected": -0.8498989939689636, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 1.0695187165775402, | |
| "grad_norm": 21.9932361270153, | |
| "learning_rate": 7.48312422757881e-07, | |
| "logits/chosen": -2.432481050491333, | |
| "logits/rejected": -2.416440010070801, | |
| "logps/chosen": -221.5184326171875, | |
| "logps/rejected": -238.3723602294922, | |
| "loss": 0.2515, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 1.44422447681427, | |
| "rewards/margins": 3.0625545978546143, | |
| "rewards/rejected": -1.6183300018310547, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.0695187165775402, | |
| "eval_logits/chosen": -2.3926949501037598, | |
| "eval_logits/rejected": -2.3801937103271484, | |
| "eval_logps/chosen": -243.76559448242188, | |
| "eval_logps/rejected": -223.52978515625, | |
| "eval_loss": 0.50016188621521, | |
| "eval_rewards/accuracies": 0.7827380895614624, | |
| "eval_rewards/chosen": 1.1633288860321045, | |
| "eval_rewards/margins": 1.9201369285583496, | |
| "eval_rewards/rejected": -0.7568081617355347, | |
| "eval_runtime": 179.117, | |
| "eval_samples_per_second": 14.851, | |
| "eval_steps_per_second": 0.234, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 1.1229946524064172, | |
| "grad_norm": 16.051097270946578, | |
| "learning_rate": 7.225616424408044e-07, | |
| "logits/chosen": -2.378419876098633, | |
| "logits/rejected": -2.397940158843994, | |
| "logps/chosen": -237.4298858642578, | |
| "logps/rejected": -201.9007110595703, | |
| "loss": 0.2112, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": 1.7581756114959717, | |
| "rewards/margins": 2.863471508026123, | |
| "rewards/rejected": -1.1052957773208618, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 1.1764705882352942, | |
| "grad_norm": 26.077993971821677, | |
| "learning_rate": 6.96052176068713e-07, | |
| "logits/chosen": -2.31712007522583, | |
| "logits/rejected": -2.3032708168029785, | |
| "logps/chosen": -250.89144897460938, | |
| "logps/rejected": -237.0896759033203, | |
| "loss": 0.2471, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 2.404763698577881, | |
| "rewards/margins": 3.7296714782714844, | |
| "rewards/rejected": -1.324907898902893, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.2299465240641712, | |
| "grad_norm": 22.630798527994564, | |
| "learning_rate": 6.688743912460229e-07, | |
| "logits/chosen": -2.259591579437256, | |
| "logits/rejected": -2.223191022872925, | |
| "logps/chosen": -237.0186309814453, | |
| "logps/rejected": -205.0946502685547, | |
| "loss": 0.296, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": 1.7101577520370483, | |
| "rewards/margins": 3.211549758911133, | |
| "rewards/rejected": -1.5013922452926636, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 1.2834224598930482, | |
| "grad_norm": 23.017342274496936, | |
| "learning_rate": 6.411209337949213e-07, | |
| "logits/chosen": -2.303983688354492, | |
| "logits/rejected": -2.2950103282928467, | |
| "logps/chosen": -216.56710815429688, | |
| "logps/rejected": -241.30038452148438, | |
| "loss": 0.269, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": 1.7581017017364502, | |
| "rewards/margins": 3.2936463356018066, | |
| "rewards/rejected": -1.5355446338653564, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.3368983957219251, | |
| "grad_norm": 20.732901772334376, | |
| "learning_rate": 6.128864119368233e-07, | |
| "logits/chosen": -2.3555662631988525, | |
| "logits/rejected": -2.3257126808166504, | |
| "logps/chosen": -238.39718627929688, | |
| "logps/rejected": -211.1797332763672, | |
| "loss": 0.2479, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": 1.7987607717514038, | |
| "rewards/margins": 3.354307174682617, | |
| "rewards/rejected": -1.555546522140503, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 1.3903743315508021, | |
| "grad_norm": 24.618263339060046, | |
| "learning_rate": 5.842670737842467e-07, | |
| "logits/chosen": -2.3552603721618652, | |
| "logits/rejected": -2.335536479949951, | |
| "logps/chosen": -234.12429809570312, | |
| "logps/rejected": -210.45089721679688, | |
| "loss": 0.2562, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.5379806756973267, | |
| "rewards/margins": 3.431626796722412, | |
| "rewards/rejected": -1.893646478652954, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 1.4438502673796791, | |
| "grad_norm": 19.84135287004455, | |
| "learning_rate": 5.553604792424922e-07, | |
| "logits/chosen": -2.387741804122925, | |
| "logits/rejected": -2.3648810386657715, | |
| "logps/chosen": -244.1914825439453, | |
| "logps/rejected": -243.6887664794922, | |
| "loss": 0.285, | |
| "rewards/accuracies": 0.90625, | |
| "rewards/chosen": 1.6935621500015259, | |
| "rewards/margins": 3.424959182739258, | |
| "rewards/rejected": -1.7313973903656006, | |
| "step": 135 | |
| }, | |
| { | |
| "epoch": 1.4973262032085561, | |
| "grad_norm": 24.865530479013078, | |
| "learning_rate": 5.262651674395798e-07, | |
| "logits/chosen": -2.4149279594421387, | |
| "logits/rejected": -2.4083454608917236, | |
| "logps/chosen": -240.76852416992188, | |
| "logps/rejected": -203.88772583007812, | |
| "loss": 0.2848, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.5183442831039429, | |
| "rewards/margins": 3.529254198074341, | |
| "rewards/rejected": -2.0109100341796875, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 1.5508021390374331, | |
| "grad_norm": 22.21336633780826, | |
| "learning_rate": 4.970803208181314e-07, | |
| "logits/chosen": -2.448906898498535, | |
| "logits/rejected": -2.449152946472168, | |
| "logps/chosen": -243.2611083984375, | |
| "logps/rejected": -222.88009643554688, | |
| "loss": 0.2809, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": 1.4146430492401123, | |
| "rewards/margins": 3.450530529022217, | |
| "rewards/rejected": -2.0358877182006836, | |
| "step": 145 | |
| }, | |
| { | |
| "epoch": 1.6042780748663101, | |
| "grad_norm": 24.85958859994276, | |
| "learning_rate": 4.679054270342702e-07, | |
| "logits/chosen": -2.437875986099243, | |
| "logits/rejected": -2.439277410507202, | |
| "logps/chosen": -229.2501983642578, | |
| "logps/rejected": -220.6576690673828, | |
| "loss": 0.2749, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 1.0872418880462646, | |
| "rewards/margins": 3.1538193225860596, | |
| "rewards/rejected": -2.066577434539795, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.6042780748663101, | |
| "eval_logits/chosen": -2.43774676322937, | |
| "eval_logits/rejected": -2.4268596172332764, | |
| "eval_logps/chosen": -246.59274291992188, | |
| "eval_logps/rejected": -230.45828247070312, | |
| "eval_loss": 0.5162904858589172, | |
| "eval_rewards/accuracies": 0.773809552192688, | |
| "eval_rewards/chosen": 0.8806134462356567, | |
| "eval_rewards/margins": 2.3302714824676514, | |
| "eval_rewards/rejected": -1.4496580362319946, | |
| "eval_runtime": 179.1524, | |
| "eval_samples_per_second": 14.848, | |
| "eval_steps_per_second": 0.234, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 1.6577540106951871, | |
| "grad_norm": 23.85929193996249, | |
| "learning_rate": 4.3883993981608567e-07, | |
| "logits/chosen": -2.421318531036377, | |
| "logits/rejected": -2.428043842315674, | |
| "logps/chosen": -219.21615600585938, | |
| "logps/rejected": -215.4480743408203, | |
| "loss": 0.2731, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": 1.5339256525039673, | |
| "rewards/margins": 3.147657871246338, | |
| "rewards/rejected": -1.6137323379516602, | |
| "step": 155 | |
| }, | |
| { | |
| "epoch": 1.7112299465240641, | |
| "grad_norm": 18.045046575777963, | |
| "learning_rate": 4.0998293993775234e-07, | |
| "logits/chosen": -2.357412815093994, | |
| "logits/rejected": -2.358879804611206, | |
| "logps/chosen": -230.360107421875, | |
| "logps/rejected": -245.9751434326172, | |
| "loss": 0.298, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": 1.3649165630340576, | |
| "rewards/margins": 3.5000126361846924, | |
| "rewards/rejected": -2.135096311569214, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 1.7647058823529411, | |
| "grad_norm": 18.083909549401817, | |
| "learning_rate": 3.814327974650066e-07, | |
| "logits/chosen": -2.36337947845459, | |
| "logits/rejected": -2.334078311920166, | |
| "logps/chosen": -238.8087158203125, | |
| "logps/rejected": -239.94656372070312, | |
| "loss": 0.2542, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.7868614196777344, | |
| "rewards/margins": 3.700504779815674, | |
| "rewards/rejected": -1.9136435985565186, | |
| "step": 165 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 26.808062566115737, | |
| "learning_rate": 3.532868364233416e-07, | |
| "logits/chosen": -2.359863758087158, | |
| "logits/rejected": -2.35335111618042, | |
| "logps/chosen": -247.17538452148438, | |
| "logps/rejected": -224.4107666015625, | |
| "loss": 0.2707, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": 1.2189900875091553, | |
| "rewards/margins": 2.990385055541992, | |
| "rewards/rejected": -1.7713947296142578, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 1.8716577540106951, | |
| "grad_norm": 25.84360767956436, | |
| "learning_rate": 3.256410030320304e-07, | |
| "logits/chosen": -2.382321834564209, | |
| "logits/rejected": -2.3686156272888184, | |
| "logps/chosen": -232.55703735351562, | |
| "logps/rejected": -249.15274047851562, | |
| "loss": 0.2832, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": 1.7483094930648804, | |
| "rewards/margins": 3.7192482948303223, | |
| "rewards/rejected": -1.9709389209747314, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 1.9251336898395723, | |
| "grad_norm": 29.502574876119475, | |
| "learning_rate": 2.985895386349233e-07, | |
| "logits/chosen": -2.3494954109191895, | |
| "logits/rejected": -2.344552755355835, | |
| "logps/chosen": -220.0012664794922, | |
| "logps/rejected": -192.16897583007812, | |
| "loss": 0.327, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.0759856700897217, | |
| "rewards/margins": 3.117873430252075, | |
| "rewards/rejected": -2.0418879985809326, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 1.9786096256684491, | |
| "grad_norm": 28.585594159089958, | |
| "learning_rate": 2.7222465844296514e-07, | |
| "logits/chosen": -2.36253023147583, | |
| "logits/rejected": -2.3399176597595215, | |
| "logps/chosen": -243.80984497070312, | |
| "logps/rejected": -237.32705688476562, | |
| "loss": 0.2707, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 1.482215166091919, | |
| "rewards/margins": 3.8792481422424316, | |
| "rewards/rejected": -2.397033214569092, | |
| "step": 185 | |
| }, | |
| { | |
| "epoch": 2.0320855614973263, | |
| "grad_norm": 16.99665080325436, | |
| "learning_rate": 2.466362371835544e-07, | |
| "logits/chosen": -2.339444398880005, | |
| "logits/rejected": -2.310081958770752, | |
| "logps/chosen": -269.8495178222656, | |
| "logps/rejected": -238.92733764648438, | |
| "loss": 0.2069, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.6897594928741455, | |
| "rewards/margins": 3.9975783824920654, | |
| "rewards/rejected": -2.307819128036499, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 2.085561497326203, | |
| "grad_norm": 12.592628034781791, | |
| "learning_rate": 2.2191150272833386e-07, | |
| "logits/chosen": -2.2964937686920166, | |
| "logits/rejected": -2.2724859714508057, | |
| "logps/chosen": -268.2159423828125, | |
| "logps/rejected": -242.285888671875, | |
| "loss": 0.1461, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 2.0836310386657715, | |
| "rewards/margins": 4.240233421325684, | |
| "rewards/rejected": -2.156602621078491, | |
| "step": 195 | |
| }, | |
| { | |
| "epoch": 2.1390374331550803, | |
| "grad_norm": 15.52633737897683, | |
| "learning_rate": 1.9813473874379395e-07, | |
| "logits/chosen": -2.2681386470794678, | |
| "logits/rejected": -2.25181245803833, | |
| "logps/chosen": -242.4980010986328, | |
| "logps/rejected": -218.2045440673828, | |
| "loss": 0.148, | |
| "rewards/accuracies": 0.9624999761581421, | |
| "rewards/chosen": 2.216174602508545, | |
| "rewards/margins": 4.149984359741211, | |
| "rewards/rejected": -1.9338098764419556, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.1390374331550803, | |
| "eval_logits/chosen": -2.2425496578216553, | |
| "eval_logits/rejected": -2.215097188949585, | |
| "eval_logps/chosen": -244.92294311523438, | |
| "eval_logps/rejected": -230.5529022216797, | |
| "eval_loss": 0.5243918299674988, | |
| "eval_rewards/accuracies": 0.788690447807312, | |
| "eval_rewards/chosen": 1.047592043876648, | |
| "eval_rewards/margins": 2.506711006164551, | |
| "eval_rewards/rejected": -1.4591190814971924, | |
| "eval_runtime": 179.1376, | |
| "eval_samples_per_second": 14.849, | |
| "eval_steps_per_second": 0.234, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 2.192513368983957, | |
| "grad_norm": 14.438053966786335, | |
| "learning_rate": 1.7538699737832237e-07, | |
| "logits/chosen": -2.2599101066589355, | |
| "logits/rejected": -2.229984998703003, | |
| "logps/chosen": -238.42465209960938, | |
| "logps/rejected": -246.22314453125, | |
| "loss": 0.1544, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 1.7270065546035767, | |
| "rewards/margins": 4.179945468902588, | |
| "rewards/rejected": -2.452939033508301, | |
| "step": 205 | |
| }, | |
| { | |
| "epoch": 2.2459893048128343, | |
| "grad_norm": 18.25472469644171, | |
| "learning_rate": 1.5374582296511053e-07, | |
| "logits/chosen": -2.2662346363067627, | |
| "logits/rejected": -2.2205100059509277, | |
| "logps/chosen": -246.0399932861328, | |
| "logps/rejected": -217.50338745117188, | |
| "loss": 0.1769, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.734651803970337, | |
| "rewards/margins": 3.9355132579803467, | |
| "rewards/rejected": -2.2008614540100098, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 2.299465240641711, | |
| "grad_norm": 16.555354862851956, | |
| "learning_rate": 1.3328498768278418e-07, | |
| "logits/chosen": -2.2428641319274902, | |
| "logits/rejected": -2.227715015411377, | |
| "logps/chosen": -231.2602081298828, | |
| "logps/rejected": -215.73104858398438, | |
| "loss": 0.1592, | |
| "rewards/accuracies": 0.949999988079071, | |
| "rewards/chosen": 1.829836130142212, | |
| "rewards/margins": 4.1750311851501465, | |
| "rewards/rejected": -2.3451952934265137, | |
| "step": 215 | |
| }, | |
| { | |
| "epoch": 2.3529411764705883, | |
| "grad_norm": 13.937201275972438, | |
| "learning_rate": 1.1407424007485928e-07, | |
| "logits/chosen": -2.239574909210205, | |
| "logits/rejected": -2.2249979972839355, | |
| "logps/chosen": -240.37655639648438, | |
| "logps/rejected": -245.929931640625, | |
| "loss": 0.1511, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 1.7110488414764404, | |
| "rewards/margins": 4.399131774902344, | |
| "rewards/rejected": -2.6880829334259033, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 2.406417112299465, | |
| "grad_norm": 16.98238781070165, | |
| "learning_rate": 9.617906728528679e-08, | |
| "logits/chosen": -2.242847204208374, | |
| "logits/rejected": -2.234100580215454, | |
| "logps/chosen": -236.35546875, | |
| "logps/rejected": -238.10684204101562, | |
| "loss": 0.1541, | |
| "rewards/accuracies": 0.9375, | |
| "rewards/chosen": 1.8751004934310913, | |
| "rewards/margins": 4.450439929962158, | |
| "rewards/rejected": -2.5753393173217773, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 2.4598930481283423, | |
| "grad_norm": 14.431070206060122, | |
| "learning_rate": 7.966047182060226e-08, | |
| "logits/chosen": -2.2399837970733643, | |
| "logits/rejected": -2.221710681915283, | |
| "logps/chosen": -238.238037109375, | |
| "logps/rejected": -228.6750946044922, | |
| "loss": 0.1542, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.7544193267822266, | |
| "rewards/margins": 4.158696174621582, | |
| "rewards/rejected": -2.4042768478393555, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 2.5133689839572195, | |
| "grad_norm": 16.667898124775515, | |
| "learning_rate": 6.457476359966684e-08, | |
| "logits/chosen": -2.2352004051208496, | |
| "logits/rejected": -2.2400095462799072, | |
| "logps/chosen": -238.9654998779297, | |
| "logps/rejected": -248.92007446289062, | |
| "loss": 0.1548, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 1.9369821548461914, | |
| "rewards/margins": 4.680064678192139, | |
| "rewards/rejected": -2.7430825233459473, | |
| "step": 235 | |
| }, | |
| { | |
| "epoch": 2.5668449197860963, | |
| "grad_norm": 21.474673970211064, | |
| "learning_rate": 5.097336799988067e-08, | |
| "logits/chosen": -2.2469589710235596, | |
| "logits/rejected": -2.239057779312134, | |
| "logps/chosen": -235.53195190429688, | |
| "logps/rejected": -235.65548706054688, | |
| "loss": 0.1703, | |
| "rewards/accuracies": 0.956250011920929, | |
| "rewards/chosen": 1.796133041381836, | |
| "rewards/margins": 4.509002208709717, | |
| "rewards/rejected": -2.712869167327881, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.620320855614973, | |
| "grad_norm": 17.249826502843938, | |
| "learning_rate": 3.8902650554212826e-08, | |
| "logits/chosen": -2.276364326477051, | |
| "logits/rejected": -2.23805570602417, | |
| "logps/chosen": -246.699462890625, | |
| "logps/rejected": -236.81015014648438, | |
| "loss": 0.1391, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 1.7032339572906494, | |
| "rewards/margins": 4.512608051300049, | |
| "rewards/rejected": -2.8093743324279785, | |
| "step": 245 | |
| }, | |
| { | |
| "epoch": 2.6737967914438503, | |
| "grad_norm": 17.650950251508487, | |
| "learning_rate": 2.8403758896638707e-08, | |
| "logits/chosen": -2.2679107189178467, | |
| "logits/rejected": -2.241079330444336, | |
| "logps/chosen": -230.755615234375, | |
| "logps/rejected": -233.1773223876953, | |
| "loss": 0.1437, | |
| "rewards/accuracies": 0.981249988079071, | |
| "rewards/chosen": 1.7524242401123047, | |
| "rewards/margins": 4.39614200592041, | |
| "rewards/rejected": -2.6437177658081055, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.6737967914438503, | |
| "eval_logits/chosen": -2.2617194652557373, | |
| "eval_logits/rejected": -2.2342660427093506, | |
| "eval_logps/chosen": -246.77085876464844, | |
| "eval_logps/rejected": -233.40264892578125, | |
| "eval_loss": 0.5224205851554871, | |
| "eval_rewards/accuracies": 0.7797619104385376, | |
| "eval_rewards/chosen": 0.8628033995628357, | |
| "eval_rewards/margins": 2.6068973541259766, | |
| "eval_rewards/rejected": -1.7440941333770752, | |
| "eval_runtime": 179.5867, | |
| "eval_samples_per_second": 14.812, | |
| "eval_steps_per_second": 0.234, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 2.7272727272727275, | |
| "grad_norm": 16.15420279517317, | |
| "learning_rate": 1.951248249476961e-08, | |
| "logits/chosen": -2.267923355102539, | |
| "logits/rejected": -2.260650634765625, | |
| "logps/chosen": -228.34219360351562, | |
| "logps/rejected": -212.5850372314453, | |
| "loss": 0.1601, | |
| "rewards/accuracies": 0.9750000238418579, | |
| "rewards/chosen": 1.7928234338760376, | |
| "rewards/margins": 4.317173957824707, | |
| "rewards/rejected": -2.524350643157959, | |
| "step": 255 | |
| }, | |
| { | |
| "epoch": 2.7807486631016043, | |
| "grad_norm": 15.002058410881267, | |
| "learning_rate": 1.2259130647833626e-08, | |
| "logits/chosen": -2.2616398334503174, | |
| "logits/rejected": -2.253014087677002, | |
| "logps/chosen": -233.8542938232422, | |
| "logps/rejected": -215.76025390625, | |
| "loss": 0.1438, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": 1.7512426376342773, | |
| "rewards/margins": 3.940762996673584, | |
| "rewards/rejected": -2.1895203590393066, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 2.834224598930481, | |
| "grad_norm": 15.129024986561038, | |
| "learning_rate": 6.668429165893996e-09, | |
| "logits/chosen": -2.273944616317749, | |
| "logits/rejected": -2.251878023147583, | |
| "logps/chosen": -222.91928100585938, | |
| "logps/rejected": -237.1503143310547, | |
| "loss": 0.1463, | |
| "rewards/accuracies": 0.9437500238418579, | |
| "rewards/chosen": 1.4963939189910889, | |
| "rewards/margins": 4.055937767028809, | |
| "rewards/rejected": -2.559544086456299, | |
| "step": 265 | |
| }, | |
| { | |
| "epoch": 2.8877005347593583, | |
| "grad_norm": 15.544498852964693, | |
| "learning_rate": 2.759436082516664e-09, | |
| "logits/chosen": -2.2754452228546143, | |
| "logits/rejected": -2.2349867820739746, | |
| "logps/chosen": -230.5224151611328, | |
| "logps/rejected": -232.70382690429688, | |
| "loss": 0.1313, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": 1.5887566804885864, | |
| "rewards/margins": 4.587313652038574, | |
| "rewards/rejected": -2.9985568523406982, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 2.9411764705882355, | |
| "grad_norm": 21.658825130572435, | |
| "learning_rate": 5.454766882097006e-10, | |
| "logits/chosen": -2.280028820037842, | |
| "logits/rejected": -2.2496137619018555, | |
| "logps/chosen": -256.89654541015625, | |
| "logps/rejected": -258.9067077636719, | |
| "loss": 0.1593, | |
| "rewards/accuracies": 0.9312499761581421, | |
| "rewards/chosen": 2.17570161819458, | |
| "rewards/margins": 4.6133832931518555, | |
| "rewards/rejected": -2.4376816749572754, | |
| "step": 275 | |
| }, | |
| { | |
| "epoch": 2.983957219251337, | |
| "step": 279, | |
| "total_flos": 3289753017384960.0, | |
| "train_loss": 0.3371215338775334, | |
| "train_runtime": 10383.8324, | |
| "train_samples_per_second": 6.914, | |
| "train_steps_per_second": 0.027 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 279, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "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": 3289753017384960.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |