| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9988623435722411, | |
| "eval_steps": 10000000, | |
| "global_step": 439, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 113.90219393876704, | |
| "learning_rate": 2.2727272727272727e-09, | |
| "logits/chosen": -1.6768856048583984, | |
| "logits/rejected": -1.7259055376052856, | |
| "logps/chosen": -1.2793102264404297, | |
| "logps/rejected": -1.2162058353424072, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 129.29173344209107, | |
| "learning_rate": 2.2727272727272725e-08, | |
| "logits/chosen": -1.702974796295166, | |
| "logits/rejected": -1.6685913801193237, | |
| "logps/chosen": -1.213309645652771, | |
| "logps/rejected": -1.2203179597854614, | |
| "loss": 0.6934, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": -4.0181395888794214e-05, | |
| "rewards/margins": -0.004325871355831623, | |
| "rewards/rejected": 0.0042856899090111256, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "grad_norm": 129.013817774111, | |
| "learning_rate": 4.545454545454545e-08, | |
| "logits/chosen": -1.7797605991363525, | |
| "logits/rejected": -1.7345327138900757, | |
| "logps/chosen": -1.144714593887329, | |
| "logps/rejected": -1.1853687763214111, | |
| "loss": 0.6906, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -0.014920851215720177, | |
| "rewards/margins": 0.0041695646941661835, | |
| "rewards/rejected": -0.01909041777253151, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "grad_norm": 125.39732746183036, | |
| "learning_rate": 6.818181818181817e-08, | |
| "logits/chosen": -1.7443883419036865, | |
| "logits/rejected": -1.6757313013076782, | |
| "logps/chosen": -1.1939805746078491, | |
| "logps/rejected": -1.2470417022705078, | |
| "loss": 0.6755, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.11079101264476776, | |
| "rewards/margins": 0.058313049376010895, | |
| "rewards/rejected": -0.16910405457019806, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "grad_norm": 99.57108978782348, | |
| "learning_rate": 9.09090909090909e-08, | |
| "logits/chosen": -1.731239914894104, | |
| "logits/rejected": -1.6651241779327393, | |
| "logps/chosen": -1.2352855205535889, | |
| "logps/rejected": -1.3083709478378296, | |
| "loss": 0.6603, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.2833693027496338, | |
| "rewards/margins": 0.20761942863464355, | |
| "rewards/rejected": -0.49098867177963257, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "grad_norm": 94.95534987876165, | |
| "learning_rate": 9.994307990108962e-08, | |
| "logits/chosen": -1.6967604160308838, | |
| "logits/rejected": -1.63302743434906, | |
| "logps/chosen": -1.2550745010375977, | |
| "logps/rejected": -1.304324984550476, | |
| "loss": 0.6261, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.4698738157749176, | |
| "rewards/margins": 0.29158082604408264, | |
| "rewards/rejected": -0.7614546418190002, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 134.17306631598268, | |
| "learning_rate": 9.959570405988094e-08, | |
| "logits/chosen": -1.7350832223892212, | |
| "logits/rejected": -1.6556373834609985, | |
| "logps/chosen": -1.1744005680084229, | |
| "logps/rejected": -1.2519078254699707, | |
| "loss": 0.6292, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.6530232429504395, | |
| "rewards/margins": 0.25578898191452026, | |
| "rewards/rejected": -0.9088121652603149, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 82.25189789646025, | |
| "learning_rate": 9.893476820924666e-08, | |
| "logits/chosen": -1.8206403255462646, | |
| "logits/rejected": -1.7381073236465454, | |
| "logps/chosen": -1.2853193283081055, | |
| "logps/rejected": -1.3663524389266968, | |
| "loss": 0.5893, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.7247735261917114, | |
| "rewards/margins": 0.34368234872817993, | |
| "rewards/rejected": -1.0684559345245361, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "grad_norm": 111.68081612756365, | |
| "learning_rate": 9.796445099843647e-08, | |
| "logits/chosen": -1.8098180294036865, | |
| "logits/rejected": -1.7271239757537842, | |
| "logps/chosen": -1.2911165952682495, | |
| "logps/rejected": -1.3880374431610107, | |
| "loss": 0.6284, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.8414822816848755, | |
| "rewards/margins": 0.3639078140258789, | |
| "rewards/rejected": -1.2053899765014648, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "grad_norm": 98.75254486198716, | |
| "learning_rate": 9.669088708527066e-08, | |
| "logits/chosen": -1.7565476894378662, | |
| "logits/rejected": -1.6922813653945923, | |
| "logps/chosen": -1.3374165296554565, | |
| "logps/rejected": -1.3871676921844482, | |
| "loss": 0.5903, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.0171922445297241, | |
| "rewards/margins": 0.42706432938575745, | |
| "rewards/rejected": -1.4442565441131592, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "grad_norm": 110.42472659036405, | |
| "learning_rate": 9.512212835085849e-08, | |
| "logits/chosen": -1.7919187545776367, | |
| "logits/rejected": -1.7078930139541626, | |
| "logps/chosen": -1.284099817276001, | |
| "logps/rejected": -1.3786065578460693, | |
| "loss": 0.5835, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.1564443111419678, | |
| "rewards/margins": 0.5179702639579773, | |
| "rewards/rejected": -1.6744146347045898, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "grad_norm": 94.2164442304763, | |
| "learning_rate": 9.326809299301306e-08, | |
| "logits/chosen": -1.7935024499893188, | |
| "logits/rejected": -1.6928462982177734, | |
| "logps/chosen": -1.3194783926010132, | |
| "logps/rejected": -1.4348446130752563, | |
| "loss": 0.5774, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.188363790512085, | |
| "rewards/margins": 0.63763028383255, | |
| "rewards/rejected": -1.8259938955307007, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "grad_norm": 124.23898564185464, | |
| "learning_rate": 9.114050282021158e-08, | |
| "logits/chosen": -1.7868391275405884, | |
| "logits/rejected": -1.7275841236114502, | |
| "logps/chosen": -1.271259069442749, | |
| "logps/rejected": -1.3735918998718262, | |
| "loss": 0.563, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.2369645833969116, | |
| "rewards/margins": 0.5494047403335571, | |
| "rewards/rejected": -1.7863690853118896, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "grad_norm": 94.58029537839892, | |
| "learning_rate": 8.875280914254802e-08, | |
| "logits/chosen": -1.7857173681259155, | |
| "logits/rejected": -1.6955511569976807, | |
| "logps/chosen": -1.3196834325790405, | |
| "logps/rejected": -1.4175035953521729, | |
| "loss": 0.5654, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.2685333490371704, | |
| "rewards/margins": 0.6939533948898315, | |
| "rewards/rejected": -1.9624868631362915, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "grad_norm": 78.26072920533862, | |
| "learning_rate": 8.612010772821971e-08, | |
| "logits/chosen": -1.8125206232070923, | |
| "logits/rejected": -1.767704725265503, | |
| "logps/chosen": -1.3238108158111572, | |
| "logps/rejected": -1.362679123878479, | |
| "loss": 0.5665, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.0993187427520752, | |
| "rewards/margins": 0.6377816200256348, | |
| "rewards/rejected": -1.73710036277771, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "grad_norm": 79.54962530262117, | |
| "learning_rate": 8.325904336322055e-08, | |
| "logits/chosen": -1.784720778465271, | |
| "logits/rejected": -1.728867530822754, | |
| "logps/chosen": -1.2866287231445312, | |
| "logps/rejected": -1.4028968811035156, | |
| "loss": 0.565, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.192639708518982, | |
| "rewards/margins": 0.6909436583518982, | |
| "rewards/rejected": -1.883583426475525, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "grad_norm": 81.2500983198601, | |
| "learning_rate": 8.01877046176447e-08, | |
| "logits/chosen": -1.7332093715667725, | |
| "logits/rejected": -1.6599019765853882, | |
| "logps/chosen": -1.3141100406646729, | |
| "logps/rejected": -1.4209729433059692, | |
| "loss": 0.5662, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.358100175857544, | |
| "rewards/margins": 0.5547691583633423, | |
| "rewards/rejected": -1.9128694534301758, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "grad_norm": 82.5330575937106, | |
| "learning_rate": 7.692550948392249e-08, | |
| "logits/chosen": -1.7886979579925537, | |
| "logits/rejected": -1.7225795984268188, | |
| "logps/chosen": -1.324988603591919, | |
| "logps/rejected": -1.400198221206665, | |
| "loss": 0.5593, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -1.1473289728164673, | |
| "rewards/margins": 0.647649884223938, | |
| "rewards/rejected": -1.7949787378311157, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "grad_norm": 90.45090476516394, | |
| "learning_rate": 7.349308261002021e-08, | |
| "logits/chosen": -1.7442381381988525, | |
| "logits/rejected": -1.6850671768188477, | |
| "logps/chosen": -1.318302869796753, | |
| "logps/rejected": -1.4245436191558838, | |
| "loss": 0.5533, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.2121667861938477, | |
| "rewards/margins": 0.5659047961235046, | |
| "rewards/rejected": -1.778071641921997, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "grad_norm": 84.71027965029923, | |
| "learning_rate": 6.991212490377531e-08, | |
| "logits/chosen": -1.7995620965957642, | |
| "logits/rejected": -1.745843529701233, | |
| "logps/chosen": -1.3425184488296509, | |
| "logps/rejected": -1.4334449768066406, | |
| "loss": 0.5291, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.1241543292999268, | |
| "rewards/margins": 0.6938456296920776, | |
| "rewards/rejected": -1.818000078201294, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "grad_norm": 104.41708218825802, | |
| "learning_rate": 6.620527633276978e-08, | |
| "logits/chosen": -1.7381435632705688, | |
| "logits/rejected": -1.662602424621582, | |
| "logps/chosen": -1.3073325157165527, | |
| "logps/rejected": -1.4739927053451538, | |
| "loss": 0.5404, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.1287935972213745, | |
| "rewards/margins": 0.7987143993377686, | |
| "rewards/rejected": -1.927507758140564, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 98.12458997212138, | |
| "learning_rate": 6.239597278716581e-08, | |
| "logits/chosen": -1.808566689491272, | |
| "logits/rejected": -1.7453174591064453, | |
| "logps/chosen": -1.3598778247833252, | |
| "logps/rejected": -1.4206109046936035, | |
| "loss": 0.5355, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.1454766988754272, | |
| "rewards/margins": 0.813967227935791, | |
| "rewards/rejected": -1.9594440460205078, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "grad_norm": 73.95327688888813, | |
| "learning_rate": 5.8508297910462456e-08, | |
| "logits/chosen": -1.7581428289413452, | |
| "logits/rejected": -1.6758205890655518, | |
| "logps/chosen": -1.2785987854003906, | |
| "logps/rejected": -1.4212281703948975, | |
| "loss": 0.5267, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -1.1508338451385498, | |
| "rewards/margins": 0.8485604524612427, | |
| "rewards/rejected": -1.999394178390503, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "grad_norm": 101.92396024664481, | |
| "learning_rate": 5.456683083494731e-08, | |
| "logits/chosen": -1.7508561611175537, | |
| "logits/rejected": -1.7111085653305054, | |
| "logps/chosen": -1.2999995946884155, | |
| "logps/rejected": -1.3745485544204712, | |
| "loss": 0.554, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -1.0619090795516968, | |
| "rewards/margins": 0.5080693364143372, | |
| "rewards/rejected": -1.5699783563613892, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "grad_norm": 110.90195766775575, | |
| "learning_rate": 5.059649078450834e-08, | |
| "logits/chosen": -1.745862603187561, | |
| "logits/rejected": -1.69782292842865, | |
| "logps/chosen": -1.2861733436584473, | |
| "logps/rejected": -1.4080326557159424, | |
| "loss": 0.5271, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.1781983375549316, | |
| "rewards/margins": 0.5982070565223694, | |
| "rewards/rejected": -1.7764055728912354, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "grad_norm": 90.78412484714156, | |
| "learning_rate": 4.6622379527277186e-08, | |
| "logits/chosen": -1.7498047351837158, | |
| "logits/rejected": -1.6966612339019775, | |
| "logps/chosen": -1.263024091720581, | |
| "logps/rejected": -1.3622833490371704, | |
| "loss": 0.5322, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.241353988647461, | |
| "rewards/margins": 0.6617964506149292, | |
| "rewards/rejected": -1.9031503200531006, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "grad_norm": 80.47587413810997, | |
| "learning_rate": 4.26696226741691e-08, | |
| "logits/chosen": -1.766274094581604, | |
| "logits/rejected": -1.6926915645599365, | |
| "logps/chosen": -1.3750711679458618, | |
| "logps/rejected": -1.4678473472595215, | |
| "loss": 0.5309, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.4224761724472046, | |
| "rewards/margins": 0.7280942797660828, | |
| "rewards/rejected": -2.1505703926086426, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "grad_norm": 81.2536856430445, | |
| "learning_rate": 3.876321082668098e-08, | |
| "logits/chosen": -1.8292741775512695, | |
| "logits/rejected": -1.7543790340423584, | |
| "logps/chosen": -1.3468070030212402, | |
| "logps/rejected": -1.4356333017349243, | |
| "loss": 0.5228, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -1.0831630229949951, | |
| "rewards/margins": 0.7951737642288208, | |
| "rewards/rejected": -1.8783369064331055, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "grad_norm": 87.95396768106268, | |
| "learning_rate": 3.492784157826244e-08, | |
| "logits/chosen": -1.754434585571289, | |
| "logits/rejected": -1.654488205909729, | |
| "logps/chosen": -1.331969141960144, | |
| "logps/rejected": -1.4491215944290161, | |
| "loss": 0.5257, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.0909301042556763, | |
| "rewards/margins": 0.8176229596138, | |
| "rewards/rejected": -1.908552885055542, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 75.75936548706386, | |
| "learning_rate": 3.118776336817812e-08, | |
| "logits/chosen": -1.800217628479004, | |
| "logits/rejected": -1.7326066493988037, | |
| "logps/chosen": -1.3206775188446045, | |
| "logps/rejected": -1.4107261896133423, | |
| "loss": 0.5236, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.1742886304855347, | |
| "rewards/margins": 0.8283244967460632, | |
| "rewards/rejected": -2.002613067626953, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "grad_norm": 86.61923660491236, | |
| "learning_rate": 2.7566622175067443e-08, | |
| "logits/chosen": -1.781106948852539, | |
| "logits/rejected": -1.7118265628814697, | |
| "logps/chosen": -1.3446478843688965, | |
| "logps/rejected": -1.4952938556671143, | |
| "loss": 0.5304, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.3606449365615845, | |
| "rewards/margins": 0.7265914082527161, | |
| "rewards/rejected": -2.0872364044189453, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 104.45230542444779, | |
| "learning_rate": 2.408731201945432e-08, | |
| "logits/chosen": -1.7766481637954712, | |
| "logits/rejected": -1.7281579971313477, | |
| "logps/chosen": -1.3146270513534546, | |
| "logps/rejected": -1.4018762111663818, | |
| "loss": 0.5176, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.2073173522949219, | |
| "rewards/margins": 0.670944333076477, | |
| "rewards/rejected": -1.8782615661621094, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "grad_norm": 97.89401056240122, | |
| "learning_rate": 2.0771830220378112e-08, | |
| "logits/chosen": -1.7322553396224976, | |
| "logits/rejected": -1.6741504669189453, | |
| "logps/chosen": -1.347646951675415, | |
| "logps/rejected": -1.4063564538955688, | |
| "loss": 0.5268, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.2476258277893066, | |
| "rewards/margins": 0.6655985713005066, | |
| "rewards/rejected": -1.913224458694458, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "grad_norm": 79.38110491324964, | |
| "learning_rate": 1.7641138321260257e-08, | |
| "logits/chosen": -1.7678320407867432, | |
| "logits/rejected": -1.692814826965332, | |
| "logps/chosen": -1.3020561933517456, | |
| "logps/rejected": -1.4196479320526123, | |
| "loss": 0.5181, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -1.1820148229599, | |
| "rewards/margins": 0.900565505027771, | |
| "rewards/rejected": -2.082580327987671, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "grad_norm": 79.242696260741, | |
| "learning_rate": 1.4715029564277793e-08, | |
| "logits/chosen": -1.8388326168060303, | |
| "logits/rejected": -1.7862517833709717, | |
| "logps/chosen": -1.306144118309021, | |
| "logps/rejected": -1.4227768182754517, | |
| "loss": 0.5342, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.0148370265960693, | |
| "rewards/margins": 0.7956900596618652, | |
| "rewards/rejected": -1.8105270862579346, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "grad_norm": 86.49780334887537, | |
| "learning_rate": 1.2012003751113343e-08, | |
| "logits/chosen": -1.822198510169983, | |
| "logits/rejected": -1.756326675415039, | |
| "logps/chosen": -1.2951385974884033, | |
| "logps/rejected": -1.4268760681152344, | |
| "loss": 0.5045, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.329688310623169, | |
| "rewards/margins": 0.815467357635498, | |
| "rewards/rejected": -2.145155668258667, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 82.9744571713248, | |
| "learning_rate": 9.549150281252633e-09, | |
| "logits/chosen": -1.7854712009429932, | |
| "logits/rejected": -1.7307994365692139, | |
| "logps/chosen": -1.2910486459732056, | |
| "logps/rejected": -1.4215586185455322, | |
| "loss": 0.508, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.1891343593597412, | |
| "rewards/margins": 0.7877954244613647, | |
| "rewards/rejected": -1.9769299030303955, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "grad_norm": 77.94995731245406, | |
| "learning_rate": 7.3420401072985306e-09, | |
| "logits/chosen": -1.8163235187530518, | |
| "logits/rejected": -1.7626298666000366, | |
| "logps/chosen": -1.3424322605133057, | |
| "logps/rejected": -1.483293890953064, | |
| "loss": 0.5133, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.2781879901885986, | |
| "rewards/margins": 0.8339790105819702, | |
| "rewards/rejected": -2.1121671199798584, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "grad_norm": 99.08582811249858, | |
| "learning_rate": 5.404627290395369e-09, | |
| "logits/chosen": -1.781881332397461, | |
| "logits/rejected": -1.712882399559021, | |
| "logps/chosen": -1.3003751039505005, | |
| "logps/rejected": -1.4186222553253174, | |
| "loss": 0.5136, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.1076927185058594, | |
| "rewards/margins": 0.7825103998184204, | |
| "rewards/rejected": -1.8902031183242798, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "grad_norm": 142.606766968544, | |
| "learning_rate": 3.74916077816162e-09, | |
| "logits/chosen": -1.7914397716522217, | |
| "logits/rejected": -1.7269777059555054, | |
| "logps/chosen": -1.2829267978668213, | |
| "logps/rejected": -1.3663567304611206, | |
| "loss": 0.5283, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -1.3313188552856445, | |
| "rewards/margins": 0.6730116605758667, | |
| "rewards/rejected": -2.0043303966522217, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "grad_norm": 87.75942228209004, | |
| "learning_rate": 2.386106962899165e-09, | |
| "logits/chosen": -1.7152767181396484, | |
| "logits/rejected": -1.6405121088027954, | |
| "logps/chosen": -1.3587360382080078, | |
| "logps/rejected": -1.4870920181274414, | |
| "loss": 0.523, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.3231509923934937, | |
| "rewards/margins": 0.7201642394065857, | |
| "rewards/rejected": -2.0433154106140137, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "grad_norm": 95.93932206444032, | |
| "learning_rate": 1.3240835096913706e-09, | |
| "logits/chosen": -1.7500959634780884, | |
| "logits/rejected": -1.6428474187850952, | |
| "logps/chosen": -1.280632734298706, | |
| "logps/rejected": -1.440298080444336, | |
| "loss": 0.5334, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -1.0908193588256836, | |
| "rewards/margins": 0.9025093913078308, | |
| "rewards/rejected": -1.9933288097381592, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "grad_norm": 185.3215699605587, | |
| "learning_rate": 5.698048727497462e-10, | |
| "logits/chosen": -1.7732937335968018, | |
| "logits/rejected": -1.6997044086456299, | |
| "logps/chosen": -1.3186805248260498, | |
| "logps/rejected": -1.473042607307434, | |
| "loss": 0.5207, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -1.2738254070281982, | |
| "rewards/margins": 0.8740528225898743, | |
| "rewards/rejected": -2.1478781700134277, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 85.90784633584286, | |
| "learning_rate": 1.2803984447259387e-10, | |
| "logits/chosen": -1.7828824520111084, | |
| "logits/rejected": -1.7222639322280884, | |
| "logps/chosen": -1.310931921005249, | |
| "logps/rejected": -1.4457513093948364, | |
| "loss": 0.5116, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.2859268188476562, | |
| "rewards/margins": 0.8716639280319214, | |
| "rewards/rejected": -2.157590866088867, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 439, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5566485422347293, | |
| "train_runtime": 6841.2124, | |
| "train_samples_per_second": 8.22, | |
| "train_steps_per_second": 0.064 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 439, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |