| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.0, | |
| "eval_steps": 100, | |
| "global_step": 574, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.003484320557491289, | |
| "grad_norm": 289.1508698329696, | |
| "learning_rate": 3.4482758620689654e-09, | |
| "logits/chosen": -2.5345611572265625, | |
| "logits/rejected": -2.581700563430786, | |
| "logps/chosen": -60.002105712890625, | |
| "logps/rejected": -99.98374938964844, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.03484320557491289, | |
| "grad_norm": 277.24379710378327, | |
| "learning_rate": 3.448275862068966e-08, | |
| "logits/chosen": -2.5635900497436523, | |
| "logits/rejected": -2.562225818634033, | |
| "logps/chosen": -59.68327331542969, | |
| "logps/rejected": -73.39383697509766, | |
| "loss": 0.6974, | |
| "rewards/accuracies": 0.1875, | |
| "rewards/chosen": -0.01181174349039793, | |
| "rewards/margins": -0.0053515201434493065, | |
| "rewards/rejected": -0.006460222881287336, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.06968641114982578, | |
| "grad_norm": 357.5434709312531, | |
| "learning_rate": 6.896551724137931e-08, | |
| "logits/chosen": -2.6049184799194336, | |
| "logits/rejected": -2.5635530948638916, | |
| "logps/chosen": -104.09075927734375, | |
| "logps/rejected": -94.9400634765625, | |
| "loss": 0.6899, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": 0.01815590262413025, | |
| "rewards/margins": 0.04113120958209038, | |
| "rewards/rejected": -0.02297530695796013, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.10452961672473868, | |
| "grad_norm": 337.4139606266084, | |
| "learning_rate": 1.0344827586206897e-07, | |
| "logits/chosen": -2.5913662910461426, | |
| "logits/rejected": -2.5713112354278564, | |
| "logps/chosen": -82.475341796875, | |
| "logps/rejected": -91.52882385253906, | |
| "loss": 0.6923, | |
| "rewards/accuracies": 0.26875001192092896, | |
| "rewards/chosen": 0.01061153132468462, | |
| "rewards/margins": 0.004934412427246571, | |
| "rewards/rejected": 0.005677118897438049, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.13937282229965156, | |
| "grad_norm": 282.9755513914394, | |
| "learning_rate": 1.3793103448275863e-07, | |
| "logits/chosen": -2.4979004859924316, | |
| "logits/rejected": -2.495636463165283, | |
| "logps/chosen": -77.93832397460938, | |
| "logps/rejected": -73.0629653930664, | |
| "loss": 0.6791, | |
| "rewards/accuracies": 0.20624999701976776, | |
| "rewards/chosen": -0.00995340384542942, | |
| "rewards/margins": 0.020863836631178856, | |
| "rewards/rejected": -0.030817240476608276, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.17421602787456447, | |
| "grad_norm": 245.53657846887563, | |
| "learning_rate": 1.7241379310344825e-07, | |
| "logits/chosen": -2.533763885498047, | |
| "logits/rejected": -2.537358522415161, | |
| "logps/chosen": -63.84333419799805, | |
| "logps/rejected": -76.09790802001953, | |
| "loss": 0.6717, | |
| "rewards/accuracies": 0.2562499940395355, | |
| "rewards/chosen": 0.047906167805194855, | |
| "rewards/margins": 0.04513678327202797, | |
| "rewards/rejected": 0.002769380807876587, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.20905923344947736, | |
| "grad_norm": 247.7041408944673, | |
| "learning_rate": 1.9922480620155037e-07, | |
| "logits/chosen": -2.506993532180786, | |
| "logits/rejected": -2.5004165172576904, | |
| "logps/chosen": -72.5146255493164, | |
| "logps/rejected": -67.91732025146484, | |
| "loss": 0.6647, | |
| "rewards/accuracies": 0.33125001192092896, | |
| "rewards/chosen": 0.34594517946243286, | |
| "rewards/margins": 0.1003229171037674, | |
| "rewards/rejected": 0.24562230706214905, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.24390243902439024, | |
| "grad_norm": 249.16248968780815, | |
| "learning_rate": 1.9534883720930232e-07, | |
| "logits/chosen": -2.528493881225586, | |
| "logits/rejected": -2.5241379737854004, | |
| "logps/chosen": -62.91923904418945, | |
| "logps/rejected": -67.45994567871094, | |
| "loss": 0.6534, | |
| "rewards/accuracies": 0.2874999940395355, | |
| "rewards/chosen": 0.6584776043891907, | |
| "rewards/margins": 0.17514197528362274, | |
| "rewards/rejected": 0.4833356738090515, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.2787456445993031, | |
| "grad_norm": 307.9296174602441, | |
| "learning_rate": 1.9147286821705426e-07, | |
| "logits/chosen": -2.4762024879455566, | |
| "logits/rejected": -2.4663474559783936, | |
| "logps/chosen": -74.05433654785156, | |
| "logps/rejected": -76.48423767089844, | |
| "loss": 0.6613, | |
| "rewards/accuracies": 0.3187499940395355, | |
| "rewards/chosen": 0.805270791053772, | |
| "rewards/margins": 0.2899132966995239, | |
| "rewards/rejected": 0.5153574347496033, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.313588850174216, | |
| "grad_norm": 236.73851878615983, | |
| "learning_rate": 1.8759689922480618e-07, | |
| "logits/chosen": -2.4865849018096924, | |
| "logits/rejected": -2.5006980895996094, | |
| "logps/chosen": -64.770263671875, | |
| "logps/rejected": -69.27301788330078, | |
| "loss": 0.6551, | |
| "rewards/accuracies": 0.3062500059604645, | |
| "rewards/chosen": 0.6060104966163635, | |
| "rewards/margins": 0.18549713492393494, | |
| "rewards/rejected": 0.420513391494751, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.34843205574912894, | |
| "grad_norm": 302.9071215603059, | |
| "learning_rate": 1.8372093023255813e-07, | |
| "logits/chosen": -2.4684031009674072, | |
| "logits/rejected": -2.4694905281066895, | |
| "logps/chosen": -73.97297668457031, | |
| "logps/rejected": -80.28447723388672, | |
| "loss": 0.6562, | |
| "rewards/accuracies": 0.30000001192092896, | |
| "rewards/chosen": 0.649095356464386, | |
| "rewards/margins": 0.26550906896591187, | |
| "rewards/rejected": 0.3835863173007965, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.34843205574912894, | |
| "eval_logits/chosen": -2.5565242767333984, | |
| "eval_logits/rejected": -2.540536880493164, | |
| "eval_logps/chosen": -73.13079833984375, | |
| "eval_logps/rejected": -80.36016845703125, | |
| "eval_loss": 0.6328476667404175, | |
| "eval_rewards/accuracies": 0.3551587164402008, | |
| "eval_rewards/chosen": 0.6802141070365906, | |
| "eval_rewards/margins": 0.296122670173645, | |
| "eval_rewards/rejected": 0.3840915858745575, | |
| "eval_runtime": 113.5393, | |
| "eval_samples_per_second": 17.615, | |
| "eval_steps_per_second": 0.555, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.3832752613240418, | |
| "grad_norm": 321.32489311642775, | |
| "learning_rate": 1.7984496124031007e-07, | |
| "logits/chosen": -2.497744560241699, | |
| "logits/rejected": -2.4591681957244873, | |
| "logps/chosen": -72.22078704833984, | |
| "logps/rejected": -62.5744514465332, | |
| "loss": 0.6334, | |
| "rewards/accuracies": 0.3062500059604645, | |
| "rewards/chosen": 0.474620521068573, | |
| "rewards/margins": 0.256552129983902, | |
| "rewards/rejected": 0.21806840598583221, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.4181184668989547, | |
| "grad_norm": 224.1076153559292, | |
| "learning_rate": 1.7596899224806202e-07, | |
| "logits/chosen": -2.520296573638916, | |
| "logits/rejected": -2.488671064376831, | |
| "logps/chosen": -76.48248291015625, | |
| "logps/rejected": -66.29643249511719, | |
| "loss": 0.619, | |
| "rewards/accuracies": 0.29374998807907104, | |
| "rewards/chosen": 0.48954588174819946, | |
| "rewards/margins": 0.30197009444236755, | |
| "rewards/rejected": 0.1875758022069931, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.4529616724738676, | |
| "grad_norm": 339.42643355240523, | |
| "learning_rate": 1.7209302325581394e-07, | |
| "logits/chosen": -2.5603480339050293, | |
| "logits/rejected": -2.5410096645355225, | |
| "logps/chosen": -83.0821533203125, | |
| "logps/rejected": -87.61650085449219, | |
| "loss": 0.6341, | |
| "rewards/accuracies": 0.3812499940395355, | |
| "rewards/chosen": 0.3993155062198639, | |
| "rewards/margins": 0.4533071517944336, | |
| "rewards/rejected": -0.053991664201021194, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.4878048780487805, | |
| "grad_norm": 307.50496227541765, | |
| "learning_rate": 1.6821705426356588e-07, | |
| "logits/chosen": -2.459794282913208, | |
| "logits/rejected": -2.447605609893799, | |
| "logps/chosen": -80.36363220214844, | |
| "logps/rejected": -70.87696838378906, | |
| "loss": 0.616, | |
| "rewards/accuracies": 0.35624998807907104, | |
| "rewards/chosen": 0.3909004330635071, | |
| "rewards/margins": 0.4472619891166687, | |
| "rewards/rejected": -0.05636156350374222, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.5226480836236934, | |
| "grad_norm": 305.3753178775468, | |
| "learning_rate": 1.6434108527131783e-07, | |
| "logits/chosen": -2.520991086959839, | |
| "logits/rejected": -2.47597074508667, | |
| "logps/chosen": -78.84916687011719, | |
| "logps/rejected": -79.61804962158203, | |
| "loss": 0.6534, | |
| "rewards/accuracies": 0.29374998807907104, | |
| "rewards/chosen": 0.31758052110671997, | |
| "rewards/margins": 0.35121217370033264, | |
| "rewards/rejected": -0.03363170847296715, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.5574912891986062, | |
| "grad_norm": 267.54053358304463, | |
| "learning_rate": 1.6046511627906975e-07, | |
| "logits/chosen": -2.4771218299865723, | |
| "logits/rejected": -2.498429298400879, | |
| "logps/chosen": -63.93762969970703, | |
| "logps/rejected": -71.9593734741211, | |
| "loss": 0.6535, | |
| "rewards/accuracies": 0.28125, | |
| "rewards/chosen": 0.29360949993133545, | |
| "rewards/margins": 0.2868410050868988, | |
| "rewards/rejected": 0.0067685008980333805, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.5923344947735192, | |
| "grad_norm": 274.1930461210022, | |
| "learning_rate": 1.565891472868217e-07, | |
| "logits/chosen": -2.4762015342712402, | |
| "logits/rejected": -2.461243152618408, | |
| "logps/chosen": -68.60774230957031, | |
| "logps/rejected": -76.81765747070312, | |
| "loss": 0.6071, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": 0.17982278764247894, | |
| "rewards/margins": 0.43163204193115234, | |
| "rewards/rejected": -0.251809298992157, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.627177700348432, | |
| "grad_norm": 274.26901021006046, | |
| "learning_rate": 1.5271317829457364e-07, | |
| "logits/chosen": -2.5262434482574463, | |
| "logits/rejected": -2.5153231620788574, | |
| "logps/chosen": -91.26216125488281, | |
| "logps/rejected": -86.61434936523438, | |
| "loss": 0.6537, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": 0.29750674962997437, | |
| "rewards/margins": 0.3494582176208496, | |
| "rewards/rejected": -0.05195152759552002, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.662020905923345, | |
| "grad_norm": 207.64618032691078, | |
| "learning_rate": 1.4883720930232558e-07, | |
| "logits/chosen": -2.523658037185669, | |
| "logits/rejected": -2.513747453689575, | |
| "logps/chosen": -70.37347412109375, | |
| "logps/rejected": -81.337646484375, | |
| "loss": 0.6272, | |
| "rewards/accuracies": 0.3125, | |
| "rewards/chosen": 0.33900195360183716, | |
| "rewards/margins": 0.24373552203178406, | |
| "rewards/rejected": 0.0952664315700531, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.6968641114982579, | |
| "grad_norm": 364.4684926175276, | |
| "learning_rate": 1.449612403100775e-07, | |
| "logits/chosen": -2.5569560527801514, | |
| "logits/rejected": -2.564101219177246, | |
| "logps/chosen": -89.63420104980469, | |
| "logps/rejected": -91.90037536621094, | |
| "loss": 0.6601, | |
| "rewards/accuracies": 0.3375000059604645, | |
| "rewards/chosen": 0.5110756754875183, | |
| "rewards/margins": 0.4359167516231537, | |
| "rewards/rejected": 0.07515887171030045, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.6968641114982579, | |
| "eval_logits/chosen": -2.5332281589508057, | |
| "eval_logits/rejected": -2.516913414001465, | |
| "eval_logps/chosen": -73.89338684082031, | |
| "eval_logps/rejected": -81.30782318115234, | |
| "eval_loss": 0.6409674882888794, | |
| "eval_rewards/accuracies": 0.3452380895614624, | |
| "eval_rewards/chosen": 0.29892128705978394, | |
| "eval_rewards/margins": 0.38865822553634644, | |
| "eval_rewards/rejected": -0.08973691612482071, | |
| "eval_runtime": 113.5083, | |
| "eval_samples_per_second": 17.62, | |
| "eval_steps_per_second": 0.555, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.7317073170731707, | |
| "grad_norm": 418.8344130807906, | |
| "learning_rate": 1.4108527131782945e-07, | |
| "logits/chosen": -2.5373802185058594, | |
| "logits/rejected": -2.510119915008545, | |
| "logps/chosen": -68.75162506103516, | |
| "logps/rejected": -63.84955978393555, | |
| "loss": 0.6439, | |
| "rewards/accuracies": 0.39375001192092896, | |
| "rewards/chosen": 0.2837230861186981, | |
| "rewards/margins": 0.5062949657440186, | |
| "rewards/rejected": -0.22257189452648163, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.7665505226480837, | |
| "grad_norm": 227.78900163940662, | |
| "learning_rate": 1.372093023255814e-07, | |
| "logits/chosen": -2.5648019313812256, | |
| "logits/rejected": -2.5423765182495117, | |
| "logps/chosen": -72.84169006347656, | |
| "logps/rejected": -71.51380920410156, | |
| "loss": 0.644, | |
| "rewards/accuracies": 0.26875001192092896, | |
| "rewards/chosen": 0.30824023485183716, | |
| "rewards/margins": 0.23355619609355927, | |
| "rewards/rejected": 0.0746840387582779, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.8013937282229965, | |
| "grad_norm": 280.8923204773476, | |
| "learning_rate": 1.333333333333333e-07, | |
| "logits/chosen": -2.56278920173645, | |
| "logits/rejected": -2.5390899181365967, | |
| "logps/chosen": -89.15130615234375, | |
| "logps/rejected": -88.63081359863281, | |
| "loss": 0.6323, | |
| "rewards/accuracies": 0.40625, | |
| "rewards/chosen": 0.4246680736541748, | |
| "rewards/margins": 0.6656498908996582, | |
| "rewards/rejected": -0.24098177254199982, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.8362369337979094, | |
| "grad_norm": 250.23928935872348, | |
| "learning_rate": 1.2945736434108528e-07, | |
| "logits/chosen": -2.5548999309539795, | |
| "logits/rejected": -2.5180442333221436, | |
| "logps/chosen": -86.33263397216797, | |
| "logps/rejected": -80.53160095214844, | |
| "loss": 0.6558, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": 0.20448681712150574, | |
| "rewards/margins": 0.4022372364997864, | |
| "rewards/rejected": -0.19775035977363586, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.8710801393728222, | |
| "grad_norm": 281.88116685137084, | |
| "learning_rate": 1.255813953488372e-07, | |
| "logits/chosen": -2.5691909790039062, | |
| "logits/rejected": -2.53218412399292, | |
| "logps/chosen": -95.07816314697266, | |
| "logps/rejected": -90.45535278320312, | |
| "loss": 0.5834, | |
| "rewards/accuracies": 0.40625, | |
| "rewards/chosen": 0.1855524331331253, | |
| "rewards/margins": 0.44013509154319763, | |
| "rewards/rejected": -0.25458264350891113, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.9059233449477352, | |
| "grad_norm": 217.05487038029386, | |
| "learning_rate": 1.2170542635658915e-07, | |
| "logits/chosen": -2.4733526706695557, | |
| "logits/rejected": -2.4898457527160645, | |
| "logps/chosen": -58.55949783325195, | |
| "logps/rejected": -65.79776763916016, | |
| "loss": 0.6288, | |
| "rewards/accuracies": 0.3062500059604645, | |
| "rewards/chosen": 0.24493035674095154, | |
| "rewards/margins": 0.329669326543808, | |
| "rewards/rejected": -0.08473895490169525, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.9407665505226481, | |
| "grad_norm": 271.80312819553575, | |
| "learning_rate": 1.1782945736434109e-07, | |
| "logits/chosen": -2.5729527473449707, | |
| "logits/rejected": -2.5736825466156006, | |
| "logps/chosen": -68.29542541503906, | |
| "logps/rejected": -83.07144927978516, | |
| "loss": 0.6039, | |
| "rewards/accuracies": 0.34375, | |
| "rewards/chosen": 0.2923750579357147, | |
| "rewards/margins": 0.562214732170105, | |
| "rewards/rejected": -0.26983970403671265, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.975609756097561, | |
| "grad_norm": 238.59512203733584, | |
| "learning_rate": 1.1395348837209302e-07, | |
| "logits/chosen": -2.4834656715393066, | |
| "logits/rejected": -2.458590030670166, | |
| "logps/chosen": -67.38432312011719, | |
| "logps/rejected": -71.17386627197266, | |
| "loss": 0.6017, | |
| "rewards/accuracies": 0.32499998807907104, | |
| "rewards/chosen": 0.22502359747886658, | |
| "rewards/margins": 0.39475032687187195, | |
| "rewards/rejected": -0.16972672939300537, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 1.0104529616724738, | |
| "grad_norm": 76.34654611961486, | |
| "learning_rate": 1.1007751937984495e-07, | |
| "logits/chosen": -2.4974358081817627, | |
| "logits/rejected": -2.4653944969177246, | |
| "logps/chosen": -70.81477355957031, | |
| "logps/rejected": -65.4640121459961, | |
| "loss": 0.5461, | |
| "rewards/accuracies": 0.3812499940395355, | |
| "rewards/chosen": 0.6264529228210449, | |
| "rewards/margins": 0.9015191197395325, | |
| "rewards/rejected": -0.27506622672080994, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 1.0452961672473868, | |
| "grad_norm": 39.056755757958314, | |
| "learning_rate": 1.062015503875969e-07, | |
| "logits/chosen": -2.51816987991333, | |
| "logits/rejected": -2.5042262077331543, | |
| "logps/chosen": -64.20988464355469, | |
| "logps/rejected": -70.89491271972656, | |
| "loss": 0.4195, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": 1.1695148944854736, | |
| "rewards/margins": 2.031907320022583, | |
| "rewards/rejected": -0.862392246723175, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.0452961672473868, | |
| "eval_logits/chosen": -2.5354163646698, | |
| "eval_logits/rejected": -2.519293785095215, | |
| "eval_logps/chosen": -73.24293518066406, | |
| "eval_logps/rejected": -80.80972290039062, | |
| "eval_loss": 0.6371246576309204, | |
| "eval_rewards/accuracies": 0.3531745970249176, | |
| "eval_rewards/chosen": 0.6241527199745178, | |
| "eval_rewards/margins": 0.4648365378379822, | |
| "eval_rewards/rejected": 0.15931615233421326, | |
| "eval_runtime": 113.4957, | |
| "eval_samples_per_second": 17.622, | |
| "eval_steps_per_second": 0.555, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 1.0801393728222997, | |
| "grad_norm": 33.225277343615126, | |
| "learning_rate": 1.0232558139534883e-07, | |
| "logits/chosen": -2.4929442405700684, | |
| "logits/rejected": -2.498056650161743, | |
| "logps/chosen": -65.01827239990234, | |
| "logps/rejected": -77.01332092285156, | |
| "loss": 0.4196, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": 1.309188723564148, | |
| "rewards/margins": 2.192389965057373, | |
| "rewards/rejected": -0.8832012414932251, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 1.1149825783972125, | |
| "grad_norm": 73.59283724981043, | |
| "learning_rate": 9.844961240310076e-08, | |
| "logits/chosen": -2.544821262359619, | |
| "logits/rejected": -2.5316672325134277, | |
| "logps/chosen": -71.5367431640625, | |
| "logps/rejected": -79.99516296386719, | |
| "loss": 0.4155, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": 1.2115765810012817, | |
| "rewards/margins": 2.1014885902404785, | |
| "rewards/rejected": -0.8899120092391968, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 1.1498257839721253, | |
| "grad_norm": 61.862973963070324, | |
| "learning_rate": 9.457364341085271e-08, | |
| "logits/chosen": -2.5284900665283203, | |
| "logits/rejected": -2.502081871032715, | |
| "logps/chosen": -81.13185119628906, | |
| "logps/rejected": -80.43305969238281, | |
| "loss": 0.3845, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 1.4948151111602783, | |
| "rewards/margins": 2.3979744911193848, | |
| "rewards/rejected": -0.9031593203544617, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 1.1846689895470384, | |
| "grad_norm": 38.944715898756485, | |
| "learning_rate": 9.069767441860464e-08, | |
| "logits/chosen": -2.5041089057922363, | |
| "logits/rejected": -2.510469436645508, | |
| "logps/chosen": -80.71260833740234, | |
| "logps/rejected": -97.7466049194336, | |
| "loss": 0.4026, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 1.5175840854644775, | |
| "rewards/margins": 2.749465227127075, | |
| "rewards/rejected": -1.2318809032440186, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 1.2195121951219512, | |
| "grad_norm": 64.94626084684288, | |
| "learning_rate": 8.682170542635659e-08, | |
| "logits/chosen": -2.5372118949890137, | |
| "logits/rejected": -2.503645896911621, | |
| "logps/chosen": -66.34137725830078, | |
| "logps/rejected": -67.35131072998047, | |
| "loss": 0.4149, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": 1.1278231143951416, | |
| "rewards/margins": 1.9639180898666382, | |
| "rewards/rejected": -0.8360951542854309, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 1.254355400696864, | |
| "grad_norm": 72.87191418772136, | |
| "learning_rate": 8.294573643410853e-08, | |
| "logits/chosen": -2.529754877090454, | |
| "logits/rejected": -2.4995744228363037, | |
| "logps/chosen": -69.45609283447266, | |
| "logps/rejected": -67.36495971679688, | |
| "loss": 0.3984, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": 1.1970014572143555, | |
| "rewards/margins": 2.0473990440368652, | |
| "rewards/rejected": -0.8503974080085754, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 1.289198606271777, | |
| "grad_norm": 78.06787405031129, | |
| "learning_rate": 7.906976744186046e-08, | |
| "logits/chosen": -2.498607635498047, | |
| "logits/rejected": -2.5195744037628174, | |
| "logps/chosen": -68.30110168457031, | |
| "logps/rejected": -75.86735534667969, | |
| "loss": 0.418, | |
| "rewards/accuracies": 0.45625001192092896, | |
| "rewards/chosen": 1.382476568222046, | |
| "rewards/margins": 2.1491870880126953, | |
| "rewards/rejected": -0.7667104005813599, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 1.32404181184669, | |
| "grad_norm": 35.524129566350084, | |
| "learning_rate": 7.51937984496124e-08, | |
| "logits/chosen": -2.507997989654541, | |
| "logits/rejected": -2.5067758560180664, | |
| "logps/chosen": -82.98981475830078, | |
| "logps/rejected": -88.70475769042969, | |
| "loss": 0.3909, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 1.942350149154663, | |
| "rewards/margins": 3.0615932941436768, | |
| "rewards/rejected": -1.1192431449890137, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 1.3588850174216027, | |
| "grad_norm": 71.79408635042864, | |
| "learning_rate": 7.131782945736434e-08, | |
| "logits/chosen": -2.5821242332458496, | |
| "logits/rejected": -2.5696043968200684, | |
| "logps/chosen": -64.58931732177734, | |
| "logps/rejected": -73.18416595458984, | |
| "loss": 0.3949, | |
| "rewards/accuracies": 0.4312500059604645, | |
| "rewards/chosen": 1.203021764755249, | |
| "rewards/margins": 2.1647653579711914, | |
| "rewards/rejected": -0.9617435336112976, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 1.3937282229965158, | |
| "grad_norm": 40.565767402769495, | |
| "learning_rate": 6.744186046511627e-08, | |
| "logits/chosen": -2.553471088409424, | |
| "logits/rejected": -2.5233709812164307, | |
| "logps/chosen": -83.41809844970703, | |
| "logps/rejected": -100.60682678222656, | |
| "loss": 0.3956, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": 1.3644764423370361, | |
| "rewards/margins": 2.466003179550171, | |
| "rewards/rejected": -1.1015267372131348, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.3937282229965158, | |
| "eval_logits/chosen": -2.5541367530822754, | |
| "eval_logits/rejected": -2.5378055572509766, | |
| "eval_logps/chosen": -73.62639617919922, | |
| "eval_logps/rejected": -81.42265319824219, | |
| "eval_loss": 0.645966649055481, | |
| "eval_rewards/accuracies": 0.363095223903656, | |
| "eval_rewards/chosen": 0.43241986632347107, | |
| "eval_rewards/margins": 0.5795699954032898, | |
| "eval_rewards/rejected": -0.14715011417865753, | |
| "eval_runtime": 113.5638, | |
| "eval_samples_per_second": 17.611, | |
| "eval_steps_per_second": 0.555, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.4285714285714286, | |
| "grad_norm": 26.444115538892063, | |
| "learning_rate": 6.356589147286822e-08, | |
| "logits/chosen": -2.534757614135742, | |
| "logits/rejected": -2.5250518321990967, | |
| "logps/chosen": -79.89534759521484, | |
| "logps/rejected": -81.98127746582031, | |
| "loss": 0.3911, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": 1.35897958278656, | |
| "rewards/margins": 2.1614632606506348, | |
| "rewards/rejected": -0.8024836778640747, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 1.4634146341463414, | |
| "grad_norm": 50.792504489879086, | |
| "learning_rate": 5.968992248062015e-08, | |
| "logits/chosen": -2.5849547386169434, | |
| "logits/rejected": -2.5869174003601074, | |
| "logps/chosen": -73.73815155029297, | |
| "logps/rejected": -87.13761901855469, | |
| "loss": 0.4085, | |
| "rewards/accuracies": 0.4312500059604645, | |
| "rewards/chosen": 1.1775071620941162, | |
| "rewards/margins": 2.3105506896972656, | |
| "rewards/rejected": -1.1330437660217285, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 1.4982578397212545, | |
| "grad_norm": 62.104061513696834, | |
| "learning_rate": 5.5813953488372087e-08, | |
| "logits/chosen": -2.5392727851867676, | |
| "logits/rejected": -2.528538465499878, | |
| "logps/chosen": -66.37205505371094, | |
| "logps/rejected": -74.73827362060547, | |
| "loss": 0.4034, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 1.1451423168182373, | |
| "rewards/margins": 2.2697441577911377, | |
| "rewards/rejected": -1.12460196018219, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.533101045296167, | |
| "grad_norm": 99.60660196952468, | |
| "learning_rate": 5.193798449612403e-08, | |
| "logits/chosen": -2.525665283203125, | |
| "logits/rejected": -2.5361111164093018, | |
| "logps/chosen": -67.78907775878906, | |
| "logps/rejected": -75.18824005126953, | |
| "loss": 0.4097, | |
| "rewards/accuracies": 0.4312500059604645, | |
| "rewards/chosen": 1.094794511795044, | |
| "rewards/margins": 1.9390010833740234, | |
| "rewards/rejected": -0.8442065119743347, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 1.5679442508710801, | |
| "grad_norm": 52.21837049460921, | |
| "learning_rate": 4.806201550387597e-08, | |
| "logits/chosen": -2.573503255844116, | |
| "logits/rejected": -2.538872718811035, | |
| "logps/chosen": -88.26780700683594, | |
| "logps/rejected": -89.10135650634766, | |
| "loss": 0.4059, | |
| "rewards/accuracies": 0.518750011920929, | |
| "rewards/chosen": 1.2870919704437256, | |
| "rewards/margins": 2.856309413909912, | |
| "rewards/rejected": -1.5692174434661865, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 1.6027874564459932, | |
| "grad_norm": 32.28937168577243, | |
| "learning_rate": 4.418604651162791e-08, | |
| "logits/chosen": -2.5357863903045654, | |
| "logits/rejected": -2.5155162811279297, | |
| "logps/chosen": -73.77734375, | |
| "logps/rejected": -79.86143493652344, | |
| "loss": 0.42, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 1.2010835409164429, | |
| "rewards/margins": 2.6294467449188232, | |
| "rewards/rejected": -1.4283630847930908, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.6376306620209058, | |
| "grad_norm": 38.87268593820624, | |
| "learning_rate": 4.031007751937984e-08, | |
| "logits/chosen": -2.5288286209106445, | |
| "logits/rejected": -2.5216517448425293, | |
| "logps/chosen": -59.2172966003418, | |
| "logps/rejected": -73.46331024169922, | |
| "loss": 0.422, | |
| "rewards/accuracies": 0.39375001192092896, | |
| "rewards/chosen": 1.042550802230835, | |
| "rewards/margins": 1.938773512840271, | |
| "rewards/rejected": -0.8962229490280151, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 1.6724738675958188, | |
| "grad_norm": 86.29055693743794, | |
| "learning_rate": 3.643410852713178e-08, | |
| "logits/chosen": -2.5448553562164307, | |
| "logits/rejected": -2.5282416343688965, | |
| "logps/chosen": -52.4884147644043, | |
| "logps/rejected": -53.635406494140625, | |
| "loss": 0.4111, | |
| "rewards/accuracies": 0.3499999940395355, | |
| "rewards/chosen": 0.7865492105484009, | |
| "rewards/margins": 1.6136754751205444, | |
| "rewards/rejected": -0.8271263837814331, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 1.7073170731707317, | |
| "grad_norm": 101.66995149882506, | |
| "learning_rate": 3.2558139534883724e-08, | |
| "logits/chosen": -2.5356173515319824, | |
| "logits/rejected": -2.5221338272094727, | |
| "logps/chosen": -71.38683319091797, | |
| "logps/rejected": -72.329345703125, | |
| "loss": 0.4539, | |
| "rewards/accuracies": 0.39375001192092896, | |
| "rewards/chosen": 0.9174805879592896, | |
| "rewards/margins": 2.05107045173645, | |
| "rewards/rejected": -1.1335899829864502, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 1.7421602787456445, | |
| "grad_norm": 62.46391948863256, | |
| "learning_rate": 2.868217054263566e-08, | |
| "logits/chosen": -2.4681618213653564, | |
| "logits/rejected": -2.4625699520111084, | |
| "logps/chosen": -73.69934844970703, | |
| "logps/rejected": -86.02604675292969, | |
| "loss": 0.3945, | |
| "rewards/accuracies": 0.45625001192092896, | |
| "rewards/chosen": 1.2942867279052734, | |
| "rewards/margins": 2.4456610679626465, | |
| "rewards/rejected": -1.1513745784759521, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.7421602787456445, | |
| "eval_logits/chosen": -2.5708680152893066, | |
| "eval_logits/rejected": -2.5542969703674316, | |
| "eval_logps/chosen": -73.87686920166016, | |
| "eval_logps/rejected": -81.72207641601562, | |
| "eval_loss": 0.6465447545051575, | |
| "eval_rewards/accuracies": 0.3710317313671112, | |
| "eval_rewards/chosen": 0.3071807324886322, | |
| "eval_rewards/margins": 0.6040430665016174, | |
| "eval_rewards/rejected": -0.29686233401298523, | |
| "eval_runtime": 114.1539, | |
| "eval_samples_per_second": 17.52, | |
| "eval_steps_per_second": 0.552, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.7770034843205575, | |
| "grad_norm": 45.13747528668081, | |
| "learning_rate": 2.4806201550387595e-08, | |
| "logits/chosen": -2.5347788333892822, | |
| "logits/rejected": -2.5145461559295654, | |
| "logps/chosen": -68.51091003417969, | |
| "logps/rejected": -71.65690612792969, | |
| "loss": 0.3946, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 1.1227554082870483, | |
| "rewards/margins": 2.1591665744781494, | |
| "rewards/rejected": -1.036410927772522, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 1.8118466898954704, | |
| "grad_norm": 69.84605729348092, | |
| "learning_rate": 2.0930232558139537e-08, | |
| "logits/chosen": -2.549290418624878, | |
| "logits/rejected": -2.5436315536499023, | |
| "logps/chosen": -72.00182342529297, | |
| "logps/rejected": -79.60043334960938, | |
| "loss": 0.4047, | |
| "rewards/accuracies": 0.46875, | |
| "rewards/chosen": 1.1346954107284546, | |
| "rewards/margins": 2.3285088539123535, | |
| "rewards/rejected": -1.1938132047653198, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 1.8466898954703832, | |
| "grad_norm": 47.800583904519996, | |
| "learning_rate": 1.7054263565891472e-08, | |
| "logits/chosen": -2.526092767715454, | |
| "logits/rejected": -2.5203917026519775, | |
| "logps/chosen": -68.9186782836914, | |
| "logps/rejected": -73.74215698242188, | |
| "loss": 0.4153, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 1.208732008934021, | |
| "rewards/margins": 2.3980166912078857, | |
| "rewards/rejected": -1.1892845630645752, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 1.8815331010452963, | |
| "grad_norm": 503.5623137913587, | |
| "learning_rate": 1.3178294573643412e-08, | |
| "logits/chosen": -2.5318820476531982, | |
| "logits/rejected": -2.5522282123565674, | |
| "logps/chosen": -61.30474090576172, | |
| "logps/rejected": -75.95423889160156, | |
| "loss": 0.426, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 1.213020920753479, | |
| "rewards/margins": 2.5194857120513916, | |
| "rewards/rejected": -1.3064649105072021, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 1.916376306620209, | |
| "grad_norm": 39.90012660671112, | |
| "learning_rate": 9.302325581395349e-09, | |
| "logits/chosen": -2.523984909057617, | |
| "logits/rejected": -2.5038905143737793, | |
| "logps/chosen": -87.86249542236328, | |
| "logps/rejected": -89.35391235351562, | |
| "loss": 0.4004, | |
| "rewards/accuracies": 0.4937500059604645, | |
| "rewards/chosen": 1.4861438274383545, | |
| "rewards/margins": 2.720036506652832, | |
| "rewards/rejected": -1.233892560005188, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 1.951219512195122, | |
| "grad_norm": 18.05863362962586, | |
| "learning_rate": 5.4263565891472866e-09, | |
| "logits/chosen": -2.5144786834716797, | |
| "logits/rejected": -2.5275046825408936, | |
| "logps/chosen": -61.50837326049805, | |
| "logps/rejected": -72.43628692626953, | |
| "loss": 0.424, | |
| "rewards/accuracies": 0.44999998807907104, | |
| "rewards/chosen": 1.105821132659912, | |
| "rewards/margins": 2.175888776779175, | |
| "rewards/rejected": -1.0700676441192627, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 1.986062717770035, | |
| "grad_norm": 49.2882801078698, | |
| "learning_rate": 1.5503875968992247e-09, | |
| "logits/chosen": -2.589893102645874, | |
| "logits/rejected": -2.5597004890441895, | |
| "logps/chosen": -63.0964241027832, | |
| "logps/rejected": -64.02296447753906, | |
| "loss": 0.4131, | |
| "rewards/accuracies": 0.4312500059604645, | |
| "rewards/chosen": 1.058401346206665, | |
| "rewards/margins": 1.7885780334472656, | |
| "rewards/rejected": -0.7301766872406006, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 574, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5264799162901237, | |
| "train_runtime": 6499.6728, | |
| "train_samples_per_second": 5.643, | |
| "train_steps_per_second": 0.088 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 574, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "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 | |
| } | |