| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9994767137624281, | |
| "eval_steps": 100, | |
| "global_step": 955, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0010465724751439038, | |
| "grad_norm": 127.36319300622706, | |
| "learning_rate": 5.208333333333333e-09, | |
| "logits/chosen": -2.9235777854919434, | |
| "logits/rejected": -2.7912116050720215, | |
| "logps/chosen": -380.68548583984375, | |
| "logps/rejected": -358.4158630371094, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.010465724751439037, | |
| "grad_norm": 107.9740860657244, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -2.595984935760498, | |
| "logits/rejected": -2.56972074508667, | |
| "logps/chosen": -256.59259033203125, | |
| "logps/rejected": -234.86553955078125, | |
| "loss": 0.6936, | |
| "rewards/accuracies": 0.4513888955116272, | |
| "rewards/chosen": 0.00012609982513822615, | |
| "rewards/margins": -5.737816172768362e-05, | |
| "rewards/rejected": 0.00018347776494920254, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.020931449502878074, | |
| "grad_norm": 123.40563647833262, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -2.6136789321899414, | |
| "logits/rejected": -2.5759799480438232, | |
| "logps/chosen": -283.09326171875, | |
| "logps/rejected": -282.3014221191406, | |
| "loss": 0.6899, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": 0.0011462386464700103, | |
| "rewards/margins": 0.004408253822475672, | |
| "rewards/rejected": -0.0032620157580822706, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03139717425431711, | |
| "grad_norm": 121.3818362170214, | |
| "learning_rate": 1.5624999999999999e-07, | |
| "logits/chosen": -2.6919732093811035, | |
| "logits/rejected": -2.667656660079956, | |
| "logps/chosen": -270.1833801269531, | |
| "logps/rejected": -276.7445068359375, | |
| "loss": 0.683, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.01242474652826786, | |
| "rewards/margins": 0.02042328752577305, | |
| "rewards/rejected": -0.007998539134860039, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04186289900575615, | |
| "grad_norm": 115.66860622860622, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -2.665289878845215, | |
| "logits/rejected": -2.5902962684631348, | |
| "logps/chosen": -290.4902648925781, | |
| "logps/rejected": -282.1328125, | |
| "loss": 0.6686, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": 0.04880904406309128, | |
| "rewards/margins": 0.05465905740857124, | |
| "rewards/rejected": -0.005850008223205805, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.052328623757195186, | |
| "grad_norm": 113.05355403051037, | |
| "learning_rate": 2.604166666666667e-07, | |
| "logits/chosen": -2.672785997390747, | |
| "logits/rejected": -2.589231014251709, | |
| "logps/chosen": -266.2040100097656, | |
| "logps/rejected": -236.48739624023438, | |
| "loss": 0.641, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.14585763216018677, | |
| "rewards/margins": 0.13357332348823547, | |
| "rewards/rejected": 0.012284321710467339, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06279434850863422, | |
| "grad_norm": 106.6001336411059, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.6317365169525146, | |
| "logits/rejected": -2.596348285675049, | |
| "logps/chosen": -300.1541442871094, | |
| "logps/rejected": -274.40191650390625, | |
| "loss": 0.6134, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.3813043236732483, | |
| "rewards/margins": 0.24904949963092804, | |
| "rewards/rejected": 0.13225486874580383, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07326007326007326, | |
| "grad_norm": 105.44232045239558, | |
| "learning_rate": 3.645833333333333e-07, | |
| "logits/chosen": -2.5406157970428467, | |
| "logits/rejected": -2.5337507724761963, | |
| "logps/chosen": -256.82513427734375, | |
| "logps/rejected": -259.8224792480469, | |
| "loss": 0.5958, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.2853667438030243, | |
| "rewards/margins": 0.44680237770080566, | |
| "rewards/rejected": -0.16143563389778137, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.0837257980115123, | |
| "grad_norm": 118.11259244631633, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.583888292312622, | |
| "logits/rejected": -2.516160249710083, | |
| "logps/chosen": -271.23541259765625, | |
| "logps/rejected": -252.84814453125, | |
| "loss": 0.5843, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.4816661775112152, | |
| "rewards/margins": 0.6344618797302246, | |
| "rewards/rejected": -0.15279565751552582, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.09419152276295134, | |
| "grad_norm": 90.59677124687741, | |
| "learning_rate": 4.6874999999999996e-07, | |
| "logits/chosen": -2.549140691757202, | |
| "logits/rejected": -2.518716335296631, | |
| "logps/chosen": -249.97048950195312, | |
| "logps/rejected": -263.16864013671875, | |
| "loss": 0.5806, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.5636419057846069, | |
| "rewards/margins": 0.4112882614135742, | |
| "rewards/rejected": 0.1523536890745163, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "grad_norm": 89.93368824706333, | |
| "learning_rate": 4.999732492681437e-07, | |
| "logits/chosen": -2.5437731742858887, | |
| "logits/rejected": -2.4748952388763428, | |
| "logps/chosen": -290.6165771484375, | |
| "logps/rejected": -277.58087158203125, | |
| "loss": 0.5446, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.9238502383232117, | |
| "rewards/margins": 0.6838744878768921, | |
| "rewards/rejected": 0.23997588455677032, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "eval_logits/chosen": -2.5742745399475098, | |
| "eval_logits/rejected": -2.516094923019409, | |
| "eval_logps/chosen": -271.84344482421875, | |
| "eval_logps/rejected": -256.6897888183594, | |
| "eval_loss": 0.5752763748168945, | |
| "eval_rewards/accuracies": 0.7242063283920288, | |
| "eval_rewards/chosen": 1.0110565423965454, | |
| "eval_rewards/margins": 0.6581434011459351, | |
| "eval_rewards/rejected": 0.35291311144828796, | |
| "eval_runtime": 206.982, | |
| "eval_samples_per_second": 9.663, | |
| "eval_steps_per_second": 0.304, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1151229722658294, | |
| "grad_norm": 91.21630112817637, | |
| "learning_rate": 4.996723692767926e-07, | |
| "logits/chosen": -2.5750699043273926, | |
| "logits/rejected": -2.528003215789795, | |
| "logps/chosen": -234.3852081298828, | |
| "logps/rejected": -236.4423370361328, | |
| "loss": 0.5648, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.705768346786499, | |
| "rewards/margins": 0.6249028444290161, | |
| "rewards/rejected": 0.08086539804935455, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12558869701726844, | |
| "grad_norm": 106.85996872671201, | |
| "learning_rate": 4.990375746213598e-07, | |
| "logits/chosen": -2.524847984313965, | |
| "logits/rejected": -2.4740397930145264, | |
| "logps/chosen": -258.1869201660156, | |
| "logps/rejected": -292.03643798828125, | |
| "loss": 0.5378, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.43742185831069946, | |
| "rewards/margins": 0.7921646237373352, | |
| "rewards/rejected": -0.3547428548336029, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.1360544217687075, | |
| "grad_norm": 113.00637141782764, | |
| "learning_rate": 4.980697142834314e-07, | |
| "logits/chosen": -2.520475387573242, | |
| "logits/rejected": -2.4845001697540283, | |
| "logps/chosen": -278.5224304199219, | |
| "logps/rejected": -261.1920166015625, | |
| "loss": 0.5472, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.26921743154525757, | |
| "rewards/margins": 0.9498673677444458, | |
| "rewards/rejected": -0.6806498765945435, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.14652014652014653, | |
| "grad_norm": 98.25615957326066, | |
| "learning_rate": 4.967700826904229e-07, | |
| "logits/chosen": -2.4557089805603027, | |
| "logits/rejected": -2.4235308170318604, | |
| "logps/chosen": -284.69012451171875, | |
| "logps/rejected": -292.3227233886719, | |
| "loss": 0.5192, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.6424055099487305, | |
| "rewards/margins": 0.9809519648551941, | |
| "rewards/rejected": -0.3385465741157532, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15698587127158556, | |
| "grad_norm": 99.56093992482657, | |
| "learning_rate": 4.951404179843962e-07, | |
| "logits/chosen": -2.601367473602295, | |
| "logits/rejected": -2.5492353439331055, | |
| "logps/chosen": -286.1108093261719, | |
| "logps/rejected": -252.67074584960938, | |
| "loss": 0.5511, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.17216625809669495, | |
| "rewards/margins": 0.7409141659736633, | |
| "rewards/rejected": -0.568747878074646, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.1674515960230246, | |
| "grad_norm": 115.6693954933169, | |
| "learning_rate": 4.931828996974498e-07, | |
| "logits/chosen": -2.568312883377075, | |
| "logits/rejected": -2.4805312156677246, | |
| "logps/chosen": -291.2640380859375, | |
| "logps/rejected": -259.77508544921875, | |
| "loss": 0.5535, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": 0.4559265077114105, | |
| "rewards/margins": 0.8324508666992188, | |
| "rewards/rejected": -0.3765243887901306, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.17791732077446362, | |
| "grad_norm": 108.75023514338689, | |
| "learning_rate": 4.909001458367866e-07, | |
| "logits/chosen": -2.556863784790039, | |
| "logits/rejected": -2.4760687351226807, | |
| "logps/chosen": -274.7050476074219, | |
| "logps/rejected": -246.76708984375, | |
| "loss": 0.5534, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.37477999925613403, | |
| "rewards/margins": 0.9715734720230103, | |
| "rewards/rejected": -0.5967934727668762, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.18838304552590268, | |
| "grad_norm": 99.91541524541664, | |
| "learning_rate": 4.882952093833627e-07, | |
| "logits/chosen": -2.543196201324463, | |
| "logits/rejected": -2.501505136489868, | |
| "logps/chosen": -242.51748657226562, | |
| "logps/rejected": -259.04241943359375, | |
| "loss": 0.5136, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.2591323256492615, | |
| "rewards/margins": 0.9666780233383179, | |
| "rewards/rejected": -1.2258104085922241, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.1988487702773417, | |
| "grad_norm": 115.88419375588242, | |
| "learning_rate": 4.853715742087946e-07, | |
| "logits/chosen": -2.4405717849731445, | |
| "logits/rejected": -2.4257616996765137, | |
| "logps/chosen": -240.6494140625, | |
| "logps/rejected": -264.52215576171875, | |
| "loss": 0.5006, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.46298712491989136, | |
| "rewards/margins": 0.9989321827888489, | |
| "rewards/rejected": -1.4619193077087402, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "grad_norm": 105.94881559115358, | |
| "learning_rate": 4.821331504159906e-07, | |
| "logits/chosen": -2.5927906036376953, | |
| "logits/rejected": -2.517674684524536, | |
| "logps/chosen": -309.7959289550781, | |
| "logps/rejected": -265.2268371582031, | |
| "loss": 0.5475, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": 0.08473672717809677, | |
| "rewards/margins": 0.8970876932144165, | |
| "rewards/rejected": -0.812350869178772, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "eval_logits/chosen": -2.5923333168029785, | |
| "eval_logits/rejected": -2.537993907928467, | |
| "eval_logps/chosen": -277.6067810058594, | |
| "eval_logps/rejected": -265.0432434082031, | |
| "eval_loss": 0.5464460849761963, | |
| "eval_rewards/accuracies": 0.7638888955116272, | |
| "eval_rewards/chosen": 0.4347245991230011, | |
| "eval_rewards/margins": 0.9171593189239502, | |
| "eval_rewards/rejected": -0.48243483901023865, | |
| "eval_runtime": 206.9145, | |
| "eval_samples_per_second": 9.666, | |
| "eval_steps_per_second": 0.304, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21978021978021978, | |
| "grad_norm": 118.18724049066705, | |
| "learning_rate": 4.785842691097342e-07, | |
| "logits/chosen": -2.5883002281188965, | |
| "logits/rejected": -2.524261951446533, | |
| "logps/chosen": -278.22625732421875, | |
| "logps/rejected": -275.1640625, | |
| "loss": 0.5643, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": 0.45807909965515137, | |
| "rewards/margins": 0.7460596561431885, | |
| "rewards/rejected": -0.2879805266857147, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.2302459445316588, | |
| "grad_norm": 156.31813233608617, | |
| "learning_rate": 4.7472967660421603e-07, | |
| "logits/chosen": -2.5972952842712402, | |
| "logits/rejected": -2.5259041786193848, | |
| "logps/chosen": -284.1155700683594, | |
| "logps/rejected": -242.37039184570312, | |
| "loss": 0.5191, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.30692246556282043, | |
| "rewards/margins": 0.9284420013427734, | |
| "rewards/rejected": -0.6215195655822754, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.24071166928309787, | |
| "grad_norm": 103.00603061075961, | |
| "learning_rate": 4.705745280752585e-07, | |
| "logits/chosen": -2.527318000793457, | |
| "logits/rejected": -2.505685329437256, | |
| "logps/chosen": -287.8115539550781, | |
| "logps/rejected": -275.6258850097656, | |
| "loss": 0.5225, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.09247669577598572, | |
| "rewards/margins": 0.7993501424789429, | |
| "rewards/rejected": -0.7068735361099243, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.25117739403453687, | |
| "grad_norm": 187.82863477728128, | |
| "learning_rate": 4.6612438066572555e-07, | |
| "logits/chosen": -2.536621570587158, | |
| "logits/rejected": -2.525778293609619, | |
| "logps/chosen": -265.84332275390625, | |
| "logps/rejected": -262.84210205078125, | |
| "loss": 0.517, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.17310741543769836, | |
| "rewards/margins": 1.0395452976226807, | |
| "rewards/rejected": -0.8664379119873047, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.2616431187859759, | |
| "grad_norm": 98.47328118938334, | |
| "learning_rate": 4.6138518605333664e-07, | |
| "logits/chosen": -2.5257980823516846, | |
| "logits/rejected": -2.50608229637146, | |
| "logps/chosen": -275.45831298828125, | |
| "logps/rejected": -265.85247802734375, | |
| "loss": 0.4916, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": 0.23112091422080994, | |
| "rewards/margins": 1.057886004447937, | |
| "rewards/rejected": -0.8267651796340942, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.272108843537415, | |
| "grad_norm": 132.81269847647957, | |
| "learning_rate": 4.5636328249082514e-07, | |
| "logits/chosen": -2.6113829612731934, | |
| "logits/rejected": -2.5672502517700195, | |
| "logps/chosen": -288.1708068847656, | |
| "logps/rejected": -290.2517395019531, | |
| "loss": 0.5332, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.08042006194591522, | |
| "rewards/margins": 0.8779838681221008, | |
| "rewards/rejected": -0.7975638508796692, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.282574568288854, | |
| "grad_norm": 109.69739978094053, | |
| "learning_rate": 4.510653863290871e-07, | |
| "logits/chosen": -2.5312352180480957, | |
| "logits/rejected": -2.514791965484619, | |
| "logps/chosen": -294.85821533203125, | |
| "logps/rejected": -259.7037658691406, | |
| "loss": 0.4994, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.4837077260017395, | |
| "rewards/margins": 1.097794771194458, | |
| "rewards/rejected": -0.614086925983429, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.29304029304029305, | |
| "grad_norm": 112.60908808433207, | |
| "learning_rate": 4.4549858303465737e-07, | |
| "logits/chosen": -2.577178478240967, | |
| "logits/rejected": -2.527773380279541, | |
| "logps/chosen": -293.6539001464844, | |
| "logps/rejected": -294.2825927734375, | |
| "loss": 0.5343, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.25686317682266235, | |
| "rewards/margins": 1.3584675788879395, | |
| "rewards/rejected": -1.1016044616699219, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.3035060177917321, | |
| "grad_norm": 136.22152391063145, | |
| "learning_rate": 4.396703177135261e-07, | |
| "logits/chosen": -2.5472209453582764, | |
| "logits/rejected": -2.470147132873535, | |
| "logps/chosen": -272.1517028808594, | |
| "logps/rejected": -246.69384765625, | |
| "loss": 0.5047, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.06049398332834244, | |
| "rewards/margins": 1.0268157720565796, | |
| "rewards/rejected": -0.9663218259811401, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "grad_norm": 103.27642265277515, | |
| "learning_rate": 4.335883851539693e-07, | |
| "logits/chosen": -2.571641206741333, | |
| "logits/rejected": -2.5657002925872803, | |
| "logps/chosen": -269.2393493652344, | |
| "logps/rejected": -277.0874938964844, | |
| "loss": 0.5359, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -0.024735305458307266, | |
| "rewards/margins": 0.9094675183296204, | |
| "rewards/rejected": -0.9342027902603149, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "eval_logits/chosen": -2.559617757797241, | |
| "eval_logits/rejected": -2.5065526962280273, | |
| "eval_logps/chosen": -281.25714111328125, | |
| "eval_logps/rejected": -270.3889465332031, | |
| "eval_loss": 0.5473023653030396, | |
| "eval_rewards/accuracies": 0.7579365372657776, | |
| "eval_rewards/chosen": 0.0696859061717987, | |
| "eval_rewards/margins": 1.0866899490356445, | |
| "eval_rewards/rejected": -1.017004132270813, | |
| "eval_runtime": 206.5677, | |
| "eval_samples_per_second": 9.682, | |
| "eval_steps_per_second": 0.305, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32443746729461015, | |
| "grad_norm": 113.48472631776741, | |
| "learning_rate": 4.272609194017105e-07, | |
| "logits/chosen": -2.5980334281921387, | |
| "logits/rejected": -2.5018410682678223, | |
| "logps/chosen": -284.5079650878906, | |
| "logps/rejected": -245.64804077148438, | |
| "loss": 0.5341, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.07721372693777084, | |
| "rewards/margins": 1.208656907081604, | |
| "rewards/rejected": -1.2858705520629883, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.3349031920460492, | |
| "grad_norm": 94.90226304989493, | |
| "learning_rate": 4.2069638288135547e-07, | |
| "logits/chosen": -2.53539776802063, | |
| "logits/rejected": -2.496711254119873, | |
| "logps/chosen": -254.16464233398438, | |
| "logps/rejected": -263.13336181640625, | |
| "loss": 0.5831, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.1118149608373642, | |
| "rewards/margins": 1.1433535814285278, | |
| "rewards/rejected": -1.2551685571670532, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.3453689167974882, | |
| "grad_norm": 111.02501519646566, | |
| "learning_rate": 4.139035550786494e-07, | |
| "logits/chosen": -2.503291606903076, | |
| "logits/rejected": -2.466763973236084, | |
| "logps/chosen": -260.0853271484375, | |
| "logps/rejected": -246.05831909179688, | |
| "loss": 0.5226, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.3735058307647705, | |
| "rewards/margins": 1.021803379058838, | |
| "rewards/rejected": -1.395309329032898, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.35583464154892724, | |
| "grad_norm": 120.37840185839579, | |
| "learning_rate": 4.0689152079869306e-07, | |
| "logits/chosen": -2.419663667678833, | |
| "logits/rejected": -2.3767175674438477, | |
| "logps/chosen": -270.08642578125, | |
| "logps/rejected": -246.24227905273438, | |
| "loss": 0.6048, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.054160237312316895, | |
| "rewards/margins": 0.966072678565979, | |
| "rewards/rejected": -1.0202327966690063, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.3663003663003663, | |
| "grad_norm": 135.62587519571315, | |
| "learning_rate": 3.99669658015821e-07, | |
| "logits/chosen": -2.512122631072998, | |
| "logits/rejected": -2.480046510696411, | |
| "logps/chosen": -289.42449951171875, | |
| "logps/rejected": -262.649169921875, | |
| "loss": 0.537, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.17944186925888062, | |
| "rewards/margins": 0.8316653370857239, | |
| "rewards/rejected": -0.6522234678268433, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.37676609105180536, | |
| "grad_norm": 95.8864031595669, | |
| "learning_rate": 3.92247625331392e-07, | |
| "logits/chosen": -2.4578399658203125, | |
| "logits/rejected": -2.4470133781433105, | |
| "logps/chosen": -281.417236328125, | |
| "logps/rejected": -271.0404357910156, | |
| "loss": 0.4816, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.4489572048187256, | |
| "rewards/margins": 1.2753773927688599, | |
| "rewards/rejected": -0.8264200091362, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.3872318158032444, | |
| "grad_norm": 100.50656645709512, | |
| "learning_rate": 3.846353490562664e-07, | |
| "logits/chosen": -2.5093467235565186, | |
| "logits/rejected": -2.4688096046447754, | |
| "logps/chosen": -260.3323669433594, | |
| "logps/rejected": -242.1940155029297, | |
| "loss": 0.5303, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.2971992492675781, | |
| "rewards/margins": 0.9509322047233582, | |
| "rewards/rejected": -0.65373295545578, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.3976975405546834, | |
| "grad_norm": 106.58256559416274, | |
| "learning_rate": 3.768430099352445e-07, | |
| "logits/chosen": -2.487880229949951, | |
| "logits/rejected": -2.454012870788574, | |
| "logps/chosen": -255.39425659179688, | |
| "logps/rejected": -257.49078369140625, | |
| "loss": 0.5166, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.07468878477811813, | |
| "rewards/margins": 1.0435364246368408, | |
| "rewards/rejected": -1.118225336074829, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.40816326530612246, | |
| "grad_norm": 101.08281656389065, | |
| "learning_rate": 3.6888102953122304e-07, | |
| "logits/chosen": -2.5439352989196777, | |
| "logits/rejected": -2.506777286529541, | |
| "logps/chosen": -287.0313720703125, | |
| "logps/rejected": -284.038818359375, | |
| "loss": 0.5404, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.5431143045425415, | |
| "rewards/margins": 1.0249578952789307, | |
| "rewards/rejected": -1.5680720806121826, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "grad_norm": 295.883680569105, | |
| "learning_rate": 3.607600562872785e-07, | |
| "logits/chosen": -2.568805456161499, | |
| "logits/rejected": -2.5805678367614746, | |
| "logps/chosen": -274.0014953613281, | |
| "logps/rejected": -281.6232604980469, | |
| "loss": 0.5228, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.4060361385345459, | |
| "rewards/margins": 0.9878374338150024, | |
| "rewards/rejected": -1.3938738107681274, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "eval_logits/chosen": -2.6470746994018555, | |
| "eval_logits/rejected": -2.59334397315979, | |
| "eval_logps/chosen": -284.2651672363281, | |
| "eval_logps/rejected": -273.2837219238281, | |
| "eval_loss": 0.5321290493011475, | |
| "eval_rewards/accuracies": 0.7539682388305664, | |
| "eval_rewards/chosen": -0.23111547529697418, | |
| "eval_rewards/margins": 1.0753650665283203, | |
| "eval_rewards/rejected": -1.3064805269241333, | |
| "eval_runtime": 206.5734, | |
| "eval_samples_per_second": 9.682, | |
| "eval_steps_per_second": 0.305, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4290947148090005, | |
| "grad_norm": 103.72761100708047, | |
| "learning_rate": 3.5249095128531856e-07, | |
| "logits/chosen": -2.588397741317749, | |
| "logits/rejected": -2.5449459552764893, | |
| "logps/chosen": -274.9619140625, | |
| "logps/rejected": -259.3703308105469, | |
| "loss": 0.5364, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.32226118445396423, | |
| "rewards/margins": 0.9323698878288269, | |
| "rewards/rejected": -1.2546310424804688, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.43956043956043955, | |
| "grad_norm": 143.18448896096157, | |
| "learning_rate": 3.4408477372034736e-07, | |
| "logits/chosen": -2.6011955738067627, | |
| "logits/rejected": -2.511427402496338, | |
| "logps/chosen": -281.31475830078125, | |
| "logps/rejected": -264.31573486328125, | |
| "loss": 0.5605, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.08763083815574646, | |
| "rewards/margins": 1.3145806789398193, | |
| "rewards/rejected": -1.226949691772461, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.4500261643118786, | |
| "grad_norm": 105.39381768622819, | |
| "learning_rate": 3.3555276610977276e-07, | |
| "logits/chosen": -2.582556962966919, | |
| "logits/rejected": -2.563215732574463, | |
| "logps/chosen": -251.7899932861328, | |
| "logps/rejected": -261.0627746582031, | |
| "loss": 0.5028, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": 0.19878244400024414, | |
| "rewards/margins": 1.2880423069000244, | |
| "rewards/rejected": -1.0892596244812012, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.4604918890633176, | |
| "grad_norm": 97.95625671060284, | |
| "learning_rate": 3.269063392575352e-07, | |
| "logits/chosen": -2.5603010654449463, | |
| "logits/rejected": -2.5315120220184326, | |
| "logps/chosen": -294.2579040527344, | |
| "logps/rejected": -261.4208679199219, | |
| "loss": 0.5781, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": 0.100133016705513, | |
| "rewards/margins": 0.6995735168457031, | |
| "rewards/rejected": -0.5994404554367065, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.47095761381475665, | |
| "grad_norm": 115.32814881542778, | |
| "learning_rate": 3.1815705699316964e-07, | |
| "logits/chosen": -2.5920796394348145, | |
| "logits/rejected": -2.5575454235076904, | |
| "logps/chosen": -281.134033203125, | |
| "logps/rejected": -289.2639465332031, | |
| "loss": 0.5217, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.12630310654640198, | |
| "rewards/margins": 1.2271454334259033, | |
| "rewards/rejected": -1.1008423566818237, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.48142333856619574, | |
| "grad_norm": 118.44955170701016, | |
| "learning_rate": 3.0931662070620794e-07, | |
| "logits/chosen": -2.5641932487487793, | |
| "logits/rejected": -2.5097055435180664, | |
| "logps/chosen": -257.33160400390625, | |
| "logps/rejected": -255.80221557617188, | |
| "loss": 0.5178, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.01436941884458065, | |
| "rewards/margins": 1.0947532653808594, | |
| "rewards/rejected": -1.0803838968276978, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.49188906331763477, | |
| "grad_norm": 95.30482890686108, | |
| "learning_rate": 3.003968536966078e-07, | |
| "logits/chosen": -2.6018574237823486, | |
| "logits/rejected": -2.559516429901123, | |
| "logps/chosen": -282.80145263671875, | |
| "logps/rejected": -259.4977111816406, | |
| "loss": 0.4768, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": 0.09861959517002106, | |
| "rewards/margins": 1.0542025566101074, | |
| "rewards/rejected": -0.9555829763412476, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5023547880690737, | |
| "grad_norm": 105.76775426891116, | |
| "learning_rate": 2.9140968536213693e-07, | |
| "logits/chosen": -2.546515464782715, | |
| "logits/rejected": -2.4800972938537598, | |
| "logps/chosen": -268.1258850097656, | |
| "logps/rejected": -265.9749450683594, | |
| "loss": 0.4923, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.09133951365947723, | |
| "rewards/margins": 1.3749868869781494, | |
| "rewards/rejected": -1.466326355934143, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.5128205128205128, | |
| "grad_norm": 99.04770574478776, | |
| "learning_rate": 2.823671352438608e-07, | |
| "logits/chosen": -2.5936622619628906, | |
| "logits/rejected": -2.4636528491973877, | |
| "logps/chosen": -310.77911376953125, | |
| "logps/rejected": -268.76055908203125, | |
| "loss": 0.5039, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.11457942426204681, | |
| "rewards/margins": 1.352496862411499, | |
| "rewards/rejected": -1.237917423248291, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "grad_norm": 89.42678222589365, | |
| "learning_rate": 2.73281296951072e-07, | |
| "logits/chosen": -2.482358694076538, | |
| "logits/rejected": -2.4588165283203125, | |
| "logps/chosen": -258.74224853515625, | |
| "logps/rejected": -238.3992156982422, | |
| "loss": 0.5217, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.006418359465897083, | |
| "rewards/margins": 1.0917348861694336, | |
| "rewards/rejected": -1.0981531143188477, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "eval_logits/chosen": -2.5772807598114014, | |
| "eval_logits/rejected": -2.5195107460021973, | |
| "eval_logps/chosen": -281.8111267089844, | |
| "eval_logps/rejected": -272.2919006347656, | |
| "eval_loss": 0.526000440120697, | |
| "eval_rewards/accuracies": 0.7876983880996704, | |
| "eval_rewards/chosen": 0.014284107834100723, | |
| "eval_rewards/margins": 1.2215818166732788, | |
| "eval_rewards/rejected": -1.2072975635528564, | |
| "eval_runtime": 206.5403, | |
| "eval_samples_per_second": 9.683, | |
| "eval_steps_per_second": 0.305, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.533751962323391, | |
| "grad_norm": 95.80547597223038, | |
| "learning_rate": 2.641643219871597e-07, | |
| "logits/chosen": -2.5779619216918945, | |
| "logits/rejected": -2.567397356033325, | |
| "logps/chosen": -281.93707275390625, | |
| "logps/rejected": -275.4586486816406, | |
| "loss": 0.4822, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.19970571994781494, | |
| "rewards/margins": 1.36869215965271, | |
| "rewards/rejected": -1.1689863204956055, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.54421768707483, | |
| "grad_norm": 88.16712559993171, | |
| "learning_rate": 2.550284034980507e-07, | |
| "logits/chosen": -2.5944199562072754, | |
| "logits/rejected": -2.571303129196167, | |
| "logps/chosen": -274.64373779296875, | |
| "logps/rejected": -263.65179443359375, | |
| "loss": 0.5305, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.2565365731716156, | |
| "rewards/margins": 1.113057255744934, | |
| "rewards/rejected": -1.369593858718872, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.554683411826269, | |
| "grad_norm": 130.8715598080221, | |
| "learning_rate": 2.4588575996495794e-07, | |
| "logits/chosen": -2.616055965423584, | |
| "logits/rejected": -2.5736825466156006, | |
| "logps/chosen": -271.31500244140625, | |
| "logps/rejected": -268.7001037597656, | |
| "loss": 0.5681, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.25552767515182495, | |
| "rewards/margins": 1.114168405532837, | |
| "rewards/rejected": -1.3696961402893066, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.565149136577708, | |
| "grad_norm": 105.92360698073946, | |
| "learning_rate": 2.367486188632446e-07, | |
| "logits/chosen": -2.6313700675964355, | |
| "logits/rejected": -2.5749824047088623, | |
| "logps/chosen": -290.64703369140625, | |
| "logps/rejected": -272.3555908203125, | |
| "loss": 0.517, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.02643487975001335, | |
| "rewards/margins": 1.1456679105758667, | |
| "rewards/rejected": -1.172102689743042, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.5756148613291471, | |
| "grad_norm": 125.07827114011454, | |
| "learning_rate": 2.276292003092593e-07, | |
| "logits/chosen": -2.5747551918029785, | |
| "logits/rejected": -2.5299413204193115, | |
| "logps/chosen": -273.75311279296875, | |
| "logps/rejected": -243.8401641845703, | |
| "loss": 0.566, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -0.24802279472351074, | |
| "rewards/margins": 1.2050869464874268, | |
| "rewards/rejected": -1.453109860420227, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5860805860805861, | |
| "grad_norm": 119.35091354587506, | |
| "learning_rate": 2.185397007170141e-07, | |
| "logits/chosen": -2.539616584777832, | |
| "logits/rejected": -2.483868360519409, | |
| "logps/chosen": -284.36090087890625, | |
| "logps/rejected": -269.1393127441406, | |
| "loss": 0.5529, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.15324077010154724, | |
| "rewards/margins": 1.269974708557129, | |
| "rewards/rejected": -1.4232155084609985, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.5965463108320251, | |
| "grad_norm": 121.35724227176418, | |
| "learning_rate": 2.094922764865619e-07, | |
| "logits/chosen": -2.620558977127075, | |
| "logits/rejected": -2.552006244659424, | |
| "logps/chosen": -282.87921142578125, | |
| "logps/rejected": -279.60931396484375, | |
| "loss": 0.5355, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.3395797610282898, | |
| "rewards/margins": 1.0040457248687744, | |
| "rewards/rejected": -1.343625545501709, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.6070120355834642, | |
| "grad_norm": 114.95741828344511, | |
| "learning_rate": 2.0049902774588797e-07, | |
| "logits/chosen": -2.604160785675049, | |
| "logits/rejected": -2.53875994682312, | |
| "logps/chosen": -254.87548828125, | |
| "logps/rejected": -224.6983184814453, | |
| "loss": 0.5335, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.15681779384613037, | |
| "rewards/margins": 1.0275707244873047, | |
| "rewards/rejected": -1.1843883991241455, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.6174777603349032, | |
| "grad_norm": 116.63197017075697, | |
| "learning_rate": 1.9157198216806238e-07, | |
| "logits/chosen": -2.5892062187194824, | |
| "logits/rejected": -2.523660659790039, | |
| "logps/chosen": -282.70709228515625, | |
| "logps/rejected": -283.9153137207031, | |
| "loss": 0.5121, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.4720216691493988, | |
| "rewards/margins": 1.0485048294067383, | |
| "rewards/rejected": -1.52052640914917, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "grad_norm": 90.93991377080742, | |
| "learning_rate": 1.8272307888529274e-07, | |
| "logits/chosen": -2.502859592437744, | |
| "logits/rejected": -2.44142746925354, | |
| "logps/chosen": -288.04864501953125, | |
| "logps/rejected": -274.6844787597656, | |
| "loss": 0.517, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.5497933626174927, | |
| "rewards/margins": 0.9119491577148438, | |
| "rewards/rejected": -1.461742639541626, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "eval_logits/chosen": -2.5743534564971924, | |
| "eval_logits/rejected": -2.518336296081543, | |
| "eval_logps/chosen": -284.8755187988281, | |
| "eval_logps/rejected": -274.78082275390625, | |
| "eval_loss": 0.5262279510498047, | |
| "eval_rewards/accuracies": 0.7698412537574768, | |
| "eval_rewards/chosen": -0.29215142130851746, | |
| "eval_rewards/margins": 1.1640383005142212, | |
| "eval_rewards/rejected": -1.4561898708343506, | |
| "eval_runtime": 206.851, | |
| "eval_samples_per_second": 9.669, | |
| "eval_steps_per_second": 0.305, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6384092098377813, | |
| "grad_norm": 109.87505024379428, | |
| "learning_rate": 1.7396415252139288e-07, | |
| "logits/chosen": -2.54246187210083, | |
| "logits/rejected": -2.4643430709838867, | |
| "logps/chosen": -308.33099365234375, | |
| "logps/rejected": -259.29583740234375, | |
| "loss": 0.5252, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.2741188406944275, | |
| "rewards/margins": 1.187243103981018, | |
| "rewards/rejected": -1.4613618850708008, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.6488749345892203, | |
| "grad_norm": 92.62549956674485, | |
| "learning_rate": 1.6530691736402316e-07, | |
| "logits/chosen": -2.4948151111602783, | |
| "logits/rejected": -2.4823691844940186, | |
| "logps/chosen": -271.5950927734375, | |
| "logps/rejected": -249.6556396484375, | |
| "loss": 0.4543, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -0.05750391632318497, | |
| "rewards/margins": 1.2674978971481323, | |
| "rewards/rejected": -1.3250019550323486, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.6593406593406593, | |
| "grad_norm": 115.38512973551775, | |
| "learning_rate": 1.5676295169786864e-07, | |
| "logits/chosen": -2.588836193084717, | |
| "logits/rejected": -2.5063748359680176, | |
| "logps/chosen": -272.53277587890625, | |
| "logps/rejected": -246.722900390625, | |
| "loss": 0.4912, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.21716764569282532, | |
| "rewards/margins": 1.2922747135162354, | |
| "rewards/rejected": -1.0751070976257324, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.6698063840920984, | |
| "grad_norm": 98.58039473996509, | |
| "learning_rate": 1.483436823197092e-07, | |
| "logits/chosen": -2.5013747215270996, | |
| "logits/rejected": -2.4492290019989014, | |
| "logps/chosen": -267.44415283203125, | |
| "logps/rejected": -240.54727172851562, | |
| "loss": 0.4593, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.04387538880109787, | |
| "rewards/margins": 1.1440480947494507, | |
| "rewards/rejected": -1.1879234313964844, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.6802721088435374, | |
| "grad_norm": 108.19817847647745, | |
| "learning_rate": 1.4006036925609243e-07, | |
| "logits/chosen": -2.545503616333008, | |
| "logits/rejected": -2.5265164375305176, | |
| "logps/chosen": -264.9919738769531, | |
| "logps/rejected": -279.6017150878906, | |
| "loss": 0.5124, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.26580336689949036, | |
| "rewards/margins": 1.0504064559936523, | |
| "rewards/rejected": -0.7846031188964844, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6907378335949764, | |
| "grad_norm": 103.96694126480867, | |
| "learning_rate": 1.319240907040458e-07, | |
| "logits/chosen": -2.545365571975708, | |
| "logits/rejected": -2.4856820106506348, | |
| "logps/chosen": -279.41656494140625, | |
| "logps/rejected": -267.95294189453125, | |
| "loss": 0.5078, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.28060275316238403, | |
| "rewards/margins": 1.3304187059402466, | |
| "rewards/rejected": -1.0498158931732178, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.7012035583464155, | |
| "grad_norm": 126.63379779239241, | |
| "learning_rate": 1.239457282149695e-07, | |
| "logits/chosen": -2.5796685218811035, | |
| "logits/rejected": -2.5014591217041016, | |
| "logps/chosen": -288.4901428222656, | |
| "logps/rejected": -296.0643615722656, | |
| "loss": 0.4813, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.004715624265372753, | |
| "rewards/margins": 1.2996948957443237, | |
| "rewards/rejected": -1.2949790954589844, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.7116692830978545, | |
| "grad_norm": 104.43312059048667, | |
| "learning_rate": 1.1613595214152711e-07, | |
| "logits/chosen": -2.5504367351531982, | |
| "logits/rejected": -2.536426067352295, | |
| "logps/chosen": -310.7039489746094, | |
| "logps/rejected": -300.8048400878906, | |
| "loss": 0.5083, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.10008208453655243, | |
| "rewards/margins": 1.3875254392623901, | |
| "rewards/rejected": -1.2874433994293213, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.7221350078492935, | |
| "grad_norm": 86.78408797627893, | |
| "learning_rate": 1.0850520736699362e-07, | |
| "logits/chosen": -2.5602102279663086, | |
| "logits/rejected": -2.476207971572876, | |
| "logps/chosen": -272.1448974609375, | |
| "logps/rejected": -247.1821746826172, | |
| "loss": 0.4535, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.14905604720115662, | |
| "rewards/margins": 1.301754355430603, | |
| "rewards/rejected": -1.4508103132247925, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "grad_norm": 101.551211026177, | |
| "learning_rate": 1.0106369933615042e-07, | |
| "logits/chosen": -2.5397322177886963, | |
| "logits/rejected": -2.493654251098633, | |
| "logps/chosen": -256.18463134765625, | |
| "logps/rejected": -252.2621612548828, | |
| "loss": 0.4766, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.2093852460384369, | |
| "rewards/margins": 1.1600673198699951, | |
| "rewards/rejected": -1.369452714920044, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "eval_logits/chosen": -2.575119972229004, | |
| "eval_logits/rejected": -2.5193893909454346, | |
| "eval_logps/chosen": -282.1366271972656, | |
| "eval_logps/rejected": -273.1544494628906, | |
| "eval_loss": 0.527851402759552, | |
| "eval_rewards/accuracies": 0.7797619104385376, | |
| "eval_rewards/chosen": -0.018264232203364372, | |
| "eval_rewards/margins": 1.2752907276153564, | |
| "eval_rewards/rejected": -1.2935550212860107, | |
| "eval_runtime": 206.8622, | |
| "eval_samples_per_second": 9.668, | |
| "eval_steps_per_second": 0.305, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7430664573521716, | |
| "grad_norm": 195.86678751064096, | |
| "learning_rate": 9.382138040640714e-08, | |
| "logits/chosen": -2.5042312145233154, | |
| "logits/rejected": -2.4359371662139893, | |
| "logps/chosen": -261.2381896972656, | |
| "logps/rejected": -257.90142822265625, | |
| "loss": 0.5532, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.27821090817451477, | |
| "rewards/margins": 1.218544602394104, | |
| "rewards/rejected": -1.4967554807662964, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.7535321821036107, | |
| "grad_norm": 97.34497217574805, | |
| "learning_rate": 8.678793653740632e-08, | |
| "logits/chosen": -2.5442166328430176, | |
| "logits/rejected": -2.506136417388916, | |
| "logps/chosen": -283.2102966308594, | |
| "logps/rejected": -260.00079345703125, | |
| "loss": 0.5108, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": 0.05546964332461357, | |
| "rewards/margins": 1.3087044954299927, | |
| "rewards/rejected": -1.25323486328125, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.7639979068550498, | |
| "grad_norm": 112.03077715841201, | |
| "learning_rate": 7.997277433690983e-08, | |
| "logits/chosen": -2.5620975494384766, | |
| "logits/rejected": -2.50207781791687, | |
| "logps/chosen": -268.3021545410156, | |
| "logps/rejected": -249.5152587890625, | |
| "loss": 0.4936, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.2925592064857483, | |
| "rewards/margins": 0.9584562182426453, | |
| "rewards/rejected": -1.251015543937683, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.7744636316064888, | |
| "grad_norm": 104.0919829579398, | |
| "learning_rate": 7.338500848029602e-08, | |
| "logits/chosen": -2.5124528408050537, | |
| "logits/rejected": -2.4893691539764404, | |
| "logps/chosen": -255.79995727539062, | |
| "logps/rejected": -245.2656707763672, | |
| "loss": 0.557, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.2949948012828827, | |
| "rewards/margins": 1.0080726146697998, | |
| "rewards/rejected": -1.3030673265457153, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.7849293563579278, | |
| "grad_norm": 99.74133982899394, | |
| "learning_rate": 6.70334495204884e-08, | |
| "logits/chosen": -2.4846601486206055, | |
| "logits/rejected": -2.44720458984375, | |
| "logps/chosen": -250.17013549804688, | |
| "logps/rejected": -258.7621765136719, | |
| "loss": 0.4753, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.3299402594566345, | |
| "rewards/margins": 1.1778974533081055, | |
| "rewards/rejected": -1.5078377723693848, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7953950811093669, | |
| "grad_norm": 73.11096892518727, | |
| "learning_rate": 6.092659210462231e-08, | |
| "logits/chosen": -2.5241293907165527, | |
| "logits/rejected": -2.492896318435669, | |
| "logps/chosen": -244.87533569335938, | |
| "logps/rejected": -258.718017578125, | |
| "loss": 0.5085, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.351382315158844, | |
| "rewards/margins": 1.2206757068634033, | |
| "rewards/rejected": -1.5720579624176025, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.8058608058608059, | |
| "grad_norm": 75.76619723672131, | |
| "learning_rate": 5.507260361320737e-08, | |
| "logits/chosen": -2.559986114501953, | |
| "logits/rejected": -2.5477547645568848, | |
| "logps/chosen": -305.89483642578125, | |
| "logps/rejected": -336.5893859863281, | |
| "loss": 0.4732, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.25339916348457336, | |
| "rewards/margins": 1.2115254402160645, | |
| "rewards/rejected": -1.4649248123168945, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.8163265306122449, | |
| "grad_norm": 105.8401947481118, | |
| "learning_rate": 4.947931323697982e-08, | |
| "logits/chosen": -2.517803430557251, | |
| "logits/rejected": -2.4777727127075195, | |
| "logps/chosen": -319.5518493652344, | |
| "logps/rejected": -279.0348815917969, | |
| "loss": 0.5237, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.0607873797416687, | |
| "rewards/margins": 1.0895849466323853, | |
| "rewards/rejected": -1.1503721475601196, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.826792255363684, | |
| "grad_norm": 113.35027564232854, | |
| "learning_rate": 4.415420150605398e-08, | |
| "logits/chosen": -2.5854082107543945, | |
| "logits/rejected": -2.4739928245544434, | |
| "logps/chosen": -280.592529296875, | |
| "logps/rejected": -259.7514343261719, | |
| "loss": 0.5226, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.20229856669902802, | |
| "rewards/margins": 1.2445086240768433, | |
| "rewards/rejected": -1.4468071460723877, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "grad_norm": 115.9328163130989, | |
| "learning_rate": 3.9104390285376374e-08, | |
| "logits/chosen": -2.4794726371765137, | |
| "logits/rejected": -2.463261365890503, | |
| "logps/chosen": -268.74591064453125, | |
| "logps/rejected": -277.9842834472656, | |
| "loss": 0.4894, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.2617167830467224, | |
| "rewards/margins": 1.2013636827468872, | |
| "rewards/rejected": -1.4630804061889648, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "eval_logits/chosen": -2.5850720405578613, | |
| "eval_logits/rejected": -2.531074047088623, | |
| "eval_logps/chosen": -282.5210876464844, | |
| "eval_logps/rejected": -272.81268310546875, | |
| "eval_loss": 0.5257476568222046, | |
| "eval_rewards/accuracies": 0.7777777910232544, | |
| "eval_rewards/chosen": -0.05670681595802307, | |
| "eval_rewards/margins": 1.2026678323745728, | |
| "eval_rewards/rejected": -1.259374737739563, | |
| "eval_runtime": 207.4453, | |
| "eval_samples_per_second": 9.641, | |
| "eval_steps_per_second": 0.304, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.847723704866562, | |
| "grad_norm": 87.80187909830617, | |
| "learning_rate": 3.433663324986208e-08, | |
| "logits/chosen": -2.5537991523742676, | |
| "logits/rejected": -2.495201826095581, | |
| "logps/chosen": -292.41900634765625, | |
| "logps/rejected": -258.72625732421875, | |
| "loss": 0.5129, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.12009086459875107, | |
| "rewards/margins": 1.1076496839523315, | |
| "rewards/rejected": -1.2277406454086304, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.858189429618001, | |
| "grad_norm": 113.52567438362996, | |
| "learning_rate": 2.9857306851953897e-08, | |
| "logits/chosen": -2.5736160278320312, | |
| "logits/rejected": -2.5441737174987793, | |
| "logps/chosen": -305.28387451171875, | |
| "logps/rejected": -295.17120361328125, | |
| "loss": 0.508, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.05176614597439766, | |
| "rewards/margins": 1.149815320968628, | |
| "rewards/rejected": -1.201581358909607, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.8686551543694401, | |
| "grad_norm": 94.35513967454145, | |
| "learning_rate": 2.567240179368185e-08, | |
| "logits/chosen": -2.5325002670288086, | |
| "logits/rejected": -2.5152266025543213, | |
| "logps/chosen": -259.08319091796875, | |
| "logps/rejected": -262.1970520019531, | |
| "loss": 0.5463, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.09561798721551895, | |
| "rewards/margins": 1.1467230319976807, | |
| "rewards/rejected": -1.242341160774231, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.8791208791208791, | |
| "grad_norm": 129.97386022506407, | |
| "learning_rate": 2.1787515014630357e-08, | |
| "logits/chosen": -2.540358304977417, | |
| "logits/rejected": -2.4877097606658936, | |
| "logps/chosen": -282.8505859375, | |
| "logps/rejected": -278.73236083984375, | |
| "loss": 0.4725, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.05927306413650513, | |
| "rewards/margins": 1.2071725130081177, | |
| "rewards/rejected": -1.266445517539978, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.8895866038723181, | |
| "grad_norm": 79.91861676789057, | |
| "learning_rate": 1.820784220652766e-08, | |
| "logits/chosen": -2.480196714401245, | |
| "logits/rejected": -2.413062572479248, | |
| "logps/chosen": -282.0146179199219, | |
| "logps/rejected": -246.9482421875, | |
| "loss": 0.4622, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.1173391118645668, | |
| "rewards/margins": 1.2294926643371582, | |
| "rewards/rejected": -1.3468319177627563, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9000523286237572, | |
| "grad_norm": 108.90343136375496, | |
| "learning_rate": 1.4938170864468636e-08, | |
| "logits/chosen": -2.517470359802246, | |
| "logits/rejected": -2.458211898803711, | |
| "logps/chosen": -289.65374755859375, | |
| "logps/rejected": -274.5342102050781, | |
| "loss": 0.4958, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.1372273564338684, | |
| "rewards/margins": 1.1825356483459473, | |
| "rewards/rejected": -1.3197630643844604, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.9105180533751962, | |
| "grad_norm": 153.358042630482, | |
| "learning_rate": 1.1982873884064465e-08, | |
| "logits/chosen": -2.578099012374878, | |
| "logits/rejected": -2.503145933151245, | |
| "logps/chosen": -271.9369201660156, | |
| "logps/rejected": -258.82427978515625, | |
| "loss": 0.5166, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.014210554771125317, | |
| "rewards/margins": 1.2173117399215698, | |
| "rewards/rejected": -1.2315223217010498, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.9209837781266352, | |
| "grad_norm": 88.58371346729444, | |
| "learning_rate": 9.345903713082304e-09, | |
| "logits/chosen": -2.5621767044067383, | |
| "logits/rejected": -2.545060396194458, | |
| "logps/chosen": -303.8287658691406, | |
| "logps/rejected": -306.15374755859375, | |
| "loss": 0.4932, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": 0.02834125980734825, | |
| "rewards/margins": 1.1778244972229004, | |
| "rewards/rejected": -1.149483323097229, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.9314495028780743, | |
| "grad_norm": 123.70760581405742, | |
| "learning_rate": 7.030787065396865e-09, | |
| "logits/chosen": -2.5063490867614746, | |
| "logits/rejected": -2.4980926513671875, | |
| "logps/chosen": -288.7598571777344, | |
| "logps/rejected": -291.5402526855469, | |
| "loss": 0.5139, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.11085305362939835, | |
| "rewards/margins": 1.0220136642456055, | |
| "rewards/rejected": -1.132866621017456, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "grad_norm": 123.75723552086023, | |
| "learning_rate": 5.04062020432286e-09, | |
| "logits/chosen": -2.5003843307495117, | |
| "logits/rejected": -2.4942610263824463, | |
| "logps/chosen": -281.16094970703125, | |
| "logps/rejected": -294.83892822265625, | |
| "loss": 0.4722, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.08161640912294388, | |
| "rewards/margins": 1.2573249340057373, | |
| "rewards/rejected": -1.3389413356781006, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "eval_logits/chosen": -2.590116024017334, | |
| "eval_logits/rejected": -2.5362370014190674, | |
| "eval_logps/chosen": -282.11407470703125, | |
| "eval_logps/rejected": -272.7223205566406, | |
| "eval_loss": 0.5279979109764099, | |
| "eval_rewards/accuracies": 0.7797619104385376, | |
| "eval_rewards/chosen": -0.01600908488035202, | |
| "eval_rewards/margins": 1.2343299388885498, | |
| "eval_rewards/rejected": -1.2503389120101929, | |
| "eval_runtime": 207.5627, | |
| "eval_samples_per_second": 9.636, | |
| "eval_steps_per_second": 0.304, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9523809523809523, | |
| "grad_norm": 117.42904817244555, | |
| "learning_rate": 3.3780648016376866e-09, | |
| "logits/chosen": -2.4798412322998047, | |
| "logits/rejected": -2.4555060863494873, | |
| "logps/chosen": -224.1981658935547, | |
| "logps/rejected": -224.94448852539062, | |
| "loss": 0.4853, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.0745500773191452, | |
| "rewards/margins": 1.1322568655014038, | |
| "rewards/rejected": -1.206807017326355, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.9628466771323915, | |
| "grad_norm": 118.07550601245399, | |
| "learning_rate": 2.0453443778310766e-09, | |
| "logits/chosen": -2.59162974357605, | |
| "logits/rejected": -2.536531925201416, | |
| "logps/chosen": -287.6225280761719, | |
| "logps/rejected": -264.23828125, | |
| "loss": 0.5345, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.09747320413589478, | |
| "rewards/margins": 1.166621208190918, | |
| "rewards/rejected": -1.264094591140747, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.9733124018838305, | |
| "grad_norm": 76.103065820101, | |
| "learning_rate": 1.0442413283435758e-09, | |
| "logits/chosen": -2.567437171936035, | |
| "logits/rejected": -2.467308521270752, | |
| "logps/chosen": -315.1402893066406, | |
| "logps/rejected": -279.08343505859375, | |
| "loss": 0.4831, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -0.03391196206212044, | |
| "rewards/margins": 1.449866533279419, | |
| "rewards/rejected": -1.4837785959243774, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.9837781266352695, | |
| "grad_norm": 105.2080573269562, | |
| "learning_rate": 3.760945397705828e-10, | |
| "logits/chosen": -2.5792009830474854, | |
| "logits/rejected": -2.542511463165283, | |
| "logps/chosen": -261.95806884765625, | |
| "logps/rejected": -271.60552978515625, | |
| "loss": 0.4913, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": 0.12502527236938477, | |
| "rewards/margins": 1.313679575920105, | |
| "rewards/rejected": -1.1886541843414307, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.9942438513867086, | |
| "grad_norm": 121.4273155412146, | |
| "learning_rate": 4.17975992204056e-11, | |
| "logits/chosen": -2.5649020671844482, | |
| "logits/rejected": -2.5450055599212646, | |
| "logps/chosen": -303.64208984375, | |
| "logps/rejected": -295.95794677734375, | |
| "loss": 0.4961, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.08054888248443604, | |
| "rewards/margins": 1.018601417541504, | |
| "rewards/rejected": -1.0991504192352295, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.9994767137624281, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 0.528570184657711, | |
| "train_runtime": 17823.4182, | |
| "train_samples_per_second": 3.43, | |
| "train_steps_per_second": 0.054 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 955, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |