| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9994340690435767, | |
| "eval_steps": 100, | |
| "global_step": 883, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 5.617977528089887e-09, | |
| "logits/chosen": -2.8261122703552246, | |
| "logits/rejected": -2.782524824142456, | |
| "logps/chosen": -386.01312255859375, | |
| "logps/rejected": -174.26467895507812, | |
| "loss": 0.2845, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 5.617977528089887e-08, | |
| "logits/chosen": -2.770503044128418, | |
| "logits/rejected": -2.7067270278930664, | |
| "logps/chosen": -337.03875732421875, | |
| "logps/rejected": -169.81399536132812, | |
| "loss": 0.2812, | |
| "rewards/accuracies": 0.3541666567325592, | |
| "rewards/chosen": -0.0004154888156335801, | |
| "rewards/margins": -0.0005714126164093614, | |
| "rewards/rejected": 0.00015592378622386605, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.1235955056179774e-07, | |
| "logits/chosen": -2.804004192352295, | |
| "logits/rejected": -2.800266742706299, | |
| "logps/chosen": -306.04486083984375, | |
| "logps/rejected": -189.47840881347656, | |
| "loss": 0.2839, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": 0.001905800774693489, | |
| "rewards/margins": 0.003199492348358035, | |
| "rewards/rejected": -0.0012936916900798678, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.6853932584269663e-07, | |
| "logits/chosen": -2.7842857837677, | |
| "logits/rejected": -2.7480905055999756, | |
| "logps/chosen": -340.1336669921875, | |
| "logps/rejected": -176.36483764648438, | |
| "loss": 0.2736, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": 0.013033352792263031, | |
| "rewards/margins": 0.02086206153035164, | |
| "rewards/rejected": -0.007828707806766033, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.2471910112359549e-07, | |
| "logits/chosen": -2.7320685386657715, | |
| "logits/rejected": -2.694530963897705, | |
| "logps/chosen": -318.46612548828125, | |
| "logps/rejected": -170.09690856933594, | |
| "loss": 0.2779, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": 0.011926891282200813, | |
| "rewards/margins": 0.05010765790939331, | |
| "rewards/rejected": -0.038180768489837646, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 2.8089887640449437e-07, | |
| "logits/chosen": -2.638516902923584, | |
| "logits/rejected": -2.6276156902313232, | |
| "logps/chosen": -326.9003601074219, | |
| "logps/rejected": -187.60606384277344, | |
| "loss": 0.2637, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.04087567329406738, | |
| "rewards/margins": 0.14330647885799408, | |
| "rewards/rejected": -0.1024308055639267, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 3.3707865168539325e-07, | |
| "logits/chosen": -2.584765672683716, | |
| "logits/rejected": -2.537776231765747, | |
| "logps/chosen": -350.67724609375, | |
| "logps/rejected": -225.81283569335938, | |
| "loss": 0.2493, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.013082382269203663, | |
| "rewards/margins": 0.19787843525409698, | |
| "rewards/rejected": -0.21096083521842957, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 3.9325842696629214e-07, | |
| "logits/chosen": -2.5709831714630127, | |
| "logits/rejected": -2.5257468223571777, | |
| "logps/chosen": -342.55291748046875, | |
| "logps/rejected": -237.45059204101562, | |
| "loss": 0.2242, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.19872836768627167, | |
| "rewards/margins": 0.2264036238193512, | |
| "rewards/rejected": -0.42513203620910645, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 4.4943820224719097e-07, | |
| "logits/chosen": -2.46580171585083, | |
| "logits/rejected": -2.4357619285583496, | |
| "logps/chosen": -338.7400817871094, | |
| "logps/rejected": -255.5446319580078, | |
| "loss": 0.187, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.2514013946056366, | |
| "rewards/margins": 0.39069053530693054, | |
| "rewards/rejected": -0.6420919895172119, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.999980431020109e-07, | |
| "logits/chosen": -2.492363452911377, | |
| "logits/rejected": -2.4689993858337402, | |
| "logps/chosen": -373.74151611328125, | |
| "logps/rejected": -278.97564697265625, | |
| "loss": 0.1471, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.2644743025302887, | |
| "rewards/margins": 0.5433769822120667, | |
| "rewards/rejected": -0.8078513145446777, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 4.997632524101301e-07, | |
| "logits/chosen": -2.5200695991516113, | |
| "logits/rejected": -2.493776798248291, | |
| "logps/chosen": -365.40191650390625, | |
| "logps/rejected": -278.2363586425781, | |
| "loss": 0.1485, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.31133976578712463, | |
| "rewards/margins": 0.6167228817939758, | |
| "rewards/rejected": -0.9280625581741333, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "eval_logits/chosen": -2.513258934020996, | |
| "eval_logits/rejected": -2.4997854232788086, | |
| "eval_logps/chosen": -313.2470703125, | |
| "eval_logps/rejected": -334.7263488769531, | |
| "eval_loss": 0.18025632202625275, | |
| "eval_rewards/accuracies": 0.640625, | |
| "eval_rewards/chosen": -0.5620743632316589, | |
| "eval_rewards/margins": 0.2116563767194748, | |
| "eval_rewards/rejected": -0.7737306952476501, | |
| "eval_runtime": 53.274, | |
| "eval_samples_per_second": 37.542, | |
| "eval_steps_per_second": 0.601, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 4.991375032514749e-07, | |
| "logits/chosen": -2.4822776317596436, | |
| "logits/rejected": -2.445827007293701, | |
| "logps/chosen": -388.3408508300781, | |
| "logps/rejected": -346.55419921875, | |
| "loss": 0.1265, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.5192545652389526, | |
| "rewards/margins": 0.6625990867614746, | |
| "rewards/rejected": -1.1818536520004272, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 4.98121775121344e-07, | |
| "logits/chosen": -2.4391703605651855, | |
| "logits/rejected": -2.4168076515197754, | |
| "logps/chosen": -354.26580810546875, | |
| "logps/rejected": -301.9068298339844, | |
| "loss": 0.1549, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.4692833423614502, | |
| "rewards/margins": 0.6351093053817749, | |
| "rewards/rejected": -1.104392647743225, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.96717657955441e-07, | |
| "logits/chosen": -2.456784963607788, | |
| "logits/rejected": -2.402589797973633, | |
| "logps/chosen": -380.1900939941406, | |
| "logps/rejected": -345.0179443359375, | |
| "loss": 0.1076, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.7707170248031616, | |
| "rewards/margins": 0.8249391317367554, | |
| "rewards/rejected": -1.595656156539917, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.949273496411216e-07, | |
| "logits/chosen": -2.3544960021972656, | |
| "logits/rejected": -2.34092378616333, | |
| "logps/chosen": -446.9285583496094, | |
| "logps/rejected": -376.6002197265625, | |
| "loss": 0.0663, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -1.0127520561218262, | |
| "rewards/margins": 1.086138367652893, | |
| "rewards/rejected": -2.098890542984009, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.927536525770046e-07, | |
| "logits/chosen": -2.376620054244995, | |
| "logits/rejected": -2.3480007648468018, | |
| "logps/chosen": -434.1917419433594, | |
| "logps/rejected": -401.83587646484375, | |
| "loss": 0.067, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.034598708152771, | |
| "rewards/margins": 0.8842358589172363, | |
| "rewards/rejected": -1.9188346862792969, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.901999692863326e-07, | |
| "logits/chosen": -2.378688335418701, | |
| "logits/rejected": -2.3023369312286377, | |
| "logps/chosen": -454.9664611816406, | |
| "logps/rejected": -411.72027587890625, | |
| "loss": 0.0667, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.2886327505111694, | |
| "rewards/margins": 0.9617815017700195, | |
| "rewards/rejected": -2.2504146099090576, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.872702970909464e-07, | |
| "logits/chosen": -2.327148914337158, | |
| "logits/rejected": -2.255551338195801, | |
| "logps/chosen": -474.45538330078125, | |
| "logps/rejected": -428.8330078125, | |
| "loss": 0.0538, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.456198811531067, | |
| "rewards/margins": 0.9602320790290833, | |
| "rewards/rejected": -2.416430950164795, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.839692218542131e-07, | |
| "logits/chosen": -2.3194737434387207, | |
| "logits/rejected": -2.289604902267456, | |
| "logps/chosen": -398.30853271484375, | |
| "logps/rejected": -385.2601013183594, | |
| "loss": 0.0682, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.9385713338851929, | |
| "rewards/margins": 1.0706655979156494, | |
| "rewards/rejected": -2.009237051010132, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.803019108026997e-07, | |
| "logits/chosen": -2.3338797092437744, | |
| "logits/rejected": -2.2716145515441895, | |
| "logps/chosen": -423.6573181152344, | |
| "logps/rejected": -402.0962219238281, | |
| "loss": 0.0702, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.050208568572998, | |
| "rewards/margins": 0.8981093168258667, | |
| "rewards/rejected": -1.9483178853988647, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.7627410443782887e-07, | |
| "logits/chosen": -2.3091745376586914, | |
| "logits/rejected": -2.2324705123901367, | |
| "logps/chosen": -481.0274353027344, | |
| "logps/rejected": -428.9169921875, | |
| "loss": 0.0592, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.2154655456542969, | |
| "rewards/margins": 1.0429514646530151, | |
| "rewards/rejected": -2.2584171295166016, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "eval_logits/chosen": -2.2729034423828125, | |
| "eval_logits/rejected": -2.2396340370178223, | |
| "eval_logps/chosen": -431.05743408203125, | |
| "eval_logps/rejected": -490.1517639160156, | |
| "eval_loss": 0.06622401624917984, | |
| "eval_rewards/accuracies": 0.6796875, | |
| "eval_rewards/chosen": -1.7401777505874634, | |
| "eval_rewards/margins": 0.5878072381019592, | |
| "eval_rewards/rejected": -2.3279850482940674, | |
| "eval_runtime": 53.2522, | |
| "eval_samples_per_second": 37.557, | |
| "eval_steps_per_second": 0.601, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.7189210755018034e-07, | |
| "logits/chosen": -2.2162270545959473, | |
| "logits/rejected": -2.109973430633545, | |
| "logps/chosen": -463.5928649902344, | |
| "logps/rejected": -448.2275390625, | |
| "loss": 0.0504, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.341931939125061, | |
| "rewards/margins": 1.1762769222259521, | |
| "rewards/rejected": -2.5182089805603027, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.671627793504988e-07, | |
| "logits/chosen": -2.222045421600342, | |
| "logits/rejected": -2.127629280090332, | |
| "logps/chosen": -493.30133056640625, | |
| "logps/rejected": -467.193115234375, | |
| "loss": 0.0508, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -1.4817438125610352, | |
| "rewards/margins": 1.220350980758667, | |
| "rewards/rejected": -2.702094554901123, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.6209352273286095e-07, | |
| "logits/chosen": -2.233363628387451, | |
| "logits/rejected": -2.184459686279297, | |
| "logps/chosen": -456.435302734375, | |
| "logps/rejected": -444.67449951171875, | |
| "loss": 0.0543, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.346243143081665, | |
| "rewards/margins": 1.1588122844696045, | |
| "rewards/rejected": -2.5050556659698486, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.56692272686805e-07, | |
| "logits/chosen": -2.227086067199707, | |
| "logits/rejected": -2.1657800674438477, | |
| "logps/chosen": -463.6180114746094, | |
| "logps/rejected": -463.9268493652344, | |
| "loss": 0.0517, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -1.6306241750717163, | |
| "rewards/margins": 1.0510774850845337, | |
| "rewards/rejected": -2.68170166015625, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 4.5096748387656326e-07, | |
| "logits/chosen": -2.2646913528442383, | |
| "logits/rejected": -2.1847498416900635, | |
| "logps/chosen": -526.8988647460938, | |
| "logps/rejected": -455.5740661621094, | |
| "loss": 0.049, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.5727968215942383, | |
| "rewards/margins": 1.1092311143875122, | |
| "rewards/rejected": -2.68202805519104, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.4492811740683877e-07, | |
| "logits/chosen": -2.272037982940674, | |
| "logits/rejected": -2.1871819496154785, | |
| "logps/chosen": -510.32061767578125, | |
| "logps/rejected": -452.8402404785156, | |
| "loss": 0.0492, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.5913618803024292, | |
| "rewards/margins": 1.0165178775787354, | |
| "rewards/rejected": -2.607879400253296, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.3858362679584354e-07, | |
| "logits/chosen": -2.17218017578125, | |
| "logits/rejected": -2.114896774291992, | |
| "logps/chosen": -487.0856018066406, | |
| "logps/rejected": -454.93115234375, | |
| "loss": 0.0458, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.7943446636199951, | |
| "rewards/margins": 0.9125874638557434, | |
| "rewards/rejected": -2.706932544708252, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.3194394317755245e-07, | |
| "logits/chosen": -2.1766180992126465, | |
| "logits/rejected": -2.0868289470672607, | |
| "logps/chosen": -574.8035888671875, | |
| "logps/rejected": -496.2310485839844, | |
| "loss": 0.0411, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -2.0265350341796875, | |
| "rewards/margins": 1.1947269439697266, | |
| "rewards/rejected": -3.221261501312256, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.2501945975633914e-07, | |
| "logits/chosen": -2.130429744720459, | |
| "logits/rejected": -2.0812618732452393, | |
| "logps/chosen": -565.4262084960938, | |
| "logps/rejected": -521.2449951171875, | |
| "loss": 0.0359, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.1656954288482666, | |
| "rewards/margins": 1.03809654712677, | |
| "rewards/rejected": -3.203791856765747, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.1782101553832405e-07, | |
| "logits/chosen": -2.1119463443756104, | |
| "logits/rejected": -2.0517754554748535, | |
| "logps/chosen": -525.1202392578125, | |
| "logps/rejected": -516.9113159179688, | |
| "loss": 0.0394, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.1311917304992676, | |
| "rewards/margins": 1.1441915035247803, | |
| "rewards/rejected": -3.2753829956054688, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "eval_logits/chosen": -2.138880968093872, | |
| "eval_logits/rejected": -2.1100966930389404, | |
| "eval_logps/chosen": -494.104736328125, | |
| "eval_logps/rejected": -555.0247802734375, | |
| "eval_loss": 0.04939676821231842, | |
| "eval_rewards/accuracies": 0.6953125, | |
| "eval_rewards/chosen": -2.37065052986145, | |
| "eval_rewards/margins": 0.606063961982727, | |
| "eval_rewards/rejected": -2.976714611053467, | |
| "eval_runtime": 53.3026, | |
| "eval_samples_per_second": 37.522, | |
| "eval_steps_per_second": 0.6, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.103598783649029e-07, | |
| "logits/chosen": -2.1108057498931885, | |
| "logits/rejected": -2.0741026401519775, | |
| "logps/chosen": -521.8655395507812, | |
| "logps/rejected": -494.017578125, | |
| "loss": 0.0399, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -1.8981420993804932, | |
| "rewards/margins": 0.9213889837265015, | |
| "rewards/rejected": -2.819530963897705, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.026477272750119e-07, | |
| "logits/chosen": -2.195026159286499, | |
| "logits/rejected": -2.128283977508545, | |
| "logps/chosen": -519.8826904296875, | |
| "logps/rejected": -489.1722717285156, | |
| "loss": 0.0407, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -1.8402801752090454, | |
| "rewards/margins": 1.171037197113037, | |
| "rewards/rejected": -3.011317491531372, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 3.9469663422373864e-07, | |
| "logits/chosen": -2.1483047008514404, | |
| "logits/rejected": -2.095182418823242, | |
| "logps/chosen": -515.9031372070312, | |
| "logps/rejected": -511.58416748046875, | |
| "loss": 0.0362, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -1.987370252609253, | |
| "rewards/margins": 1.1986653804779053, | |
| "rewards/rejected": -3.186035633087158, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.865190451858954e-07, | |
| "logits/chosen": -2.1236190795898438, | |
| "logits/rejected": -2.038846492767334, | |
| "logps/chosen": -567.5887451171875, | |
| "logps/rejected": -531.2391357421875, | |
| "loss": 0.0378, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.1050643920898438, | |
| "rewards/margins": 1.3257628679275513, | |
| "rewards/rejected": -3.4308273792266846, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.781277606741327e-07, | |
| "logits/chosen": -2.073742628097534, | |
| "logits/rejected": -2.0120556354522705, | |
| "logps/chosen": -546.9391479492188, | |
| "logps/rejected": -522.168701171875, | |
| "loss": 0.0363, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.975027084350586, | |
| "rewards/margins": 1.3364970684051514, | |
| "rewards/rejected": -3.311523914337158, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 3.6953591570208996e-07, | |
| "logits/chosen": -2.0751125812530518, | |
| "logits/rejected": -1.9913240671157837, | |
| "logps/chosen": -532.4324951171875, | |
| "logps/rejected": -570.5128173828125, | |
| "loss": 0.0339, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.149209976196289, | |
| "rewards/margins": 1.5435600280761719, | |
| "rewards/rejected": -3.692770004272461, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.607569592239452e-07, | |
| "logits/chosen": -2.0640273094177246, | |
| "logits/rejected": -1.9882128238677979, | |
| "logps/chosen": -538.2806396484375, | |
| "logps/rejected": -503.5306701660156, | |
| "loss": 0.0419, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -1.8893718719482422, | |
| "rewards/margins": 1.269718050956726, | |
| "rewards/rejected": -3.159090042114258, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.518046330825494e-07, | |
| "logits/chosen": -2.1525139808654785, | |
| "logits/rejected": -2.09255051612854, | |
| "logps/chosen": -541.6168823242188, | |
| "logps/rejected": -525.7049560546875, | |
| "loss": 0.0488, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -1.759751319885254, | |
| "rewards/margins": 1.318419098854065, | |
| "rewards/rejected": -3.0781702995300293, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.4269295049909713e-07, | |
| "logits/chosen": -2.111579179763794, | |
| "logits/rejected": -2.008594512939453, | |
| "logps/chosen": -519.3077392578125, | |
| "logps/rejected": -495.54736328125, | |
| "loss": 0.0479, | |
| "rewards/accuracies": 0.71875, | |
| "rewards/chosen": -1.9454580545425415, | |
| "rewards/margins": 1.1256051063537598, | |
| "rewards/rejected": -3.0710630416870117, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.3343617413800453e-07, | |
| "logits/chosen": -2.116332530975342, | |
| "logits/rejected": -2.0108766555786133, | |
| "logps/chosen": -560.86083984375, | |
| "logps/rejected": -513.68408203125, | |
| "loss": 0.0401, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -1.8742320537567139, | |
| "rewards/margins": 1.4719246625900269, | |
| "rewards/rejected": -3.346156597137451, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "eval_logits/chosen": -2.0799078941345215, | |
| "eval_logits/rejected": -2.042891502380371, | |
| "eval_logps/chosen": -499.7916259765625, | |
| "eval_logps/rejected": -568.111572265625, | |
| "eval_loss": 0.052323117852211, | |
| "eval_rewards/accuracies": 0.703125, | |
| "eval_rewards/chosen": -2.4275197982788086, | |
| "eval_rewards/margins": 0.6800626516342163, | |
| "eval_rewards/rejected": -3.1075825691223145, | |
| "eval_runtime": 53.2431, | |
| "eval_samples_per_second": 37.564, | |
| "eval_steps_per_second": 0.601, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.2404879378132893e-07, | |
| "logits/chosen": -2.1349825859069824, | |
| "logits/rejected": -2.0403599739074707, | |
| "logps/chosen": -573.2166137695312, | |
| "logps/rejected": -546.21484375, | |
| "loss": 0.0393, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.9816631078720093, | |
| "rewards/margins": 1.3866467475891113, | |
| "rewards/rejected": -3.3683102130889893, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.1454550364767894e-07, | |
| "logits/chosen": -2.0241355895996094, | |
| "logits/rejected": -1.955384612083435, | |
| "logps/chosen": -587.6964721679688, | |
| "logps/rejected": -564.29833984375, | |
| "loss": 0.0336, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.4298739433288574, | |
| "rewards/margins": 1.2927566766738892, | |
| "rewards/rejected": -3.722630262374878, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 3.049411793911154e-07, | |
| "logits/chosen": -2.0459389686584473, | |
| "logits/rejected": -1.9516912698745728, | |
| "logps/chosen": -587.4751586914062, | |
| "logps/rejected": -571.4771728515625, | |
| "loss": 0.0377, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.2811179161071777, | |
| "rewards/margins": 1.313342571258545, | |
| "rewards/rejected": -3.5944607257843018, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.9525085481604914e-07, | |
| "logits/chosen": -2.076460361480713, | |
| "logits/rejected": -2.008845090866089, | |
| "logps/chosen": -501.68743896484375, | |
| "logps/rejected": -492.162841796875, | |
| "loss": 0.0406, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.8489151000976562, | |
| "rewards/margins": 1.1866779327392578, | |
| "rewards/rejected": -3.035592794418335, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 2.854896983445833e-07, | |
| "logits/chosen": -2.0890183448791504, | |
| "logits/rejected": -1.99982488155365, | |
| "logps/chosen": -555.748291015625, | |
| "logps/rejected": -527.487060546875, | |
| "loss": 0.0446, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -2.0640885829925537, | |
| "rewards/margins": 1.2410598993301392, | |
| "rewards/rejected": -3.3051486015319824, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.7567298927313654e-07, | |
| "logits/chosen": -2.0581395626068115, | |
| "logits/rejected": -1.9576694965362549, | |
| "logps/chosen": -517.2969970703125, | |
| "logps/rejected": -494.01580810546875, | |
| "loss": 0.0444, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.7231872081756592, | |
| "rewards/margins": 1.3869811296463013, | |
| "rewards/rejected": -3.110168695449829, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 2.658160938555123e-07, | |
| "logits/chosen": -2.0949597358703613, | |
| "logits/rejected": -1.978247046470642, | |
| "logps/chosen": -530.58642578125, | |
| "logps/rejected": -518.8834838867188, | |
| "loss": 0.0408, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.8325612545013428, | |
| "rewards/margins": 1.4025018215179443, | |
| "rewards/rejected": -3.235063076019287, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.559344412498532e-07, | |
| "logits/chosen": -2.0444254875183105, | |
| "logits/rejected": -1.958059310913086, | |
| "logps/chosen": -538.9998779296875, | |
| "logps/rejected": -531.4126586914062, | |
| "loss": 0.0282, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -1.9186712503433228, | |
| "rewards/margins": 1.5285937786102295, | |
| "rewards/rejected": -3.447265148162842, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2.460434993671294e-07, | |
| "logits/chosen": -2.060708522796631, | |
| "logits/rejected": -1.9541358947753906, | |
| "logps/chosen": -524.92431640625, | |
| "logps/rejected": -490.29669189453125, | |
| "loss": 0.0367, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -1.956221342086792, | |
| "rewards/margins": 1.1774486303329468, | |
| "rewards/rejected": -3.1336700916290283, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.361587506589672e-07, | |
| "logits/chosen": -2.0169599056243896, | |
| "logits/rejected": -1.9417072534561157, | |
| "logps/chosen": -512.1517333984375, | |
| "logps/rejected": -506.2262268066406, | |
| "loss": 0.0335, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -1.9970229864120483, | |
| "rewards/margins": 1.3175480365753174, | |
| "rewards/rejected": -3.314570665359497, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "eval_logits/chosen": -2.0455641746520996, | |
| "eval_logits/rejected": -2.005706310272217, | |
| "eval_logps/chosen": -497.6726989746094, | |
| "eval_logps/rejected": -580.1129150390625, | |
| "eval_loss": 0.04609997943043709, | |
| "eval_rewards/accuracies": 0.71484375, | |
| "eval_rewards/chosen": -2.4063305854797363, | |
| "eval_rewards/margins": 0.8212659358978271, | |
| "eval_rewards/rejected": -3.2275965213775635, | |
| "eval_runtime": 53.2473, | |
| "eval_samples_per_second": 37.561, | |
| "eval_steps_per_second": 0.601, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.2629566788271613e-07, | |
| "logits/chosen": -2.0371432304382324, | |
| "logits/rejected": -1.9557580947875977, | |
| "logps/chosen": -560.0842895507812, | |
| "logps/rejected": -532.668212890625, | |
| "loss": 0.038, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -1.9359846115112305, | |
| "rewards/margins": 1.3490617275238037, | |
| "rewards/rejected": -3.285046339035034, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.1646968988169135e-07, | |
| "logits/chosen": -2.011871814727783, | |
| "logits/rejected": -1.9335308074951172, | |
| "logps/chosen": -507.52337646484375, | |
| "logps/rejected": -516.07958984375, | |
| "loss": 0.0395, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.9623944759368896, | |
| "rewards/margins": 1.2198023796081543, | |
| "rewards/rejected": -3.182196855545044, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.0669619741850232e-07, | |
| "logits/chosen": -1.9674116373062134, | |
| "logits/rejected": -1.874136209487915, | |
| "logps/chosen": -560.131591796875, | |
| "logps/rejected": -497.582275390625, | |
| "loss": 0.0333, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.1217665672302246, | |
| "rewards/margins": 1.2892667055130005, | |
| "rewards/rejected": -3.4110336303710938, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 1.9699048909929518e-07, | |
| "logits/chosen": -1.9731553792953491, | |
| "logits/rejected": -1.9132474660873413, | |
| "logps/chosen": -545.8017578125, | |
| "logps/rejected": -487.0414123535156, | |
| "loss": 0.0348, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -2.1676135063171387, | |
| "rewards/margins": 1.0356405973434448, | |
| "rewards/rejected": -3.203253984451294, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.8736775742659732e-07, | |
| "logits/chosen": -1.9223613739013672, | |
| "logits/rejected": -1.8391857147216797, | |
| "logps/chosen": -517.1666870117188, | |
| "logps/rejected": -498.1993713378906, | |
| "loss": 0.0342, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.948054313659668, | |
| "rewards/margins": 1.2562531232833862, | |
| "rewards/rejected": -3.2043070793151855, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.7784306501824616e-07, | |
| "logits/chosen": -1.9654014110565186, | |
| "logits/rejected": -1.8999271392822266, | |
| "logps/chosen": -574.9034423828125, | |
| "logps/rejected": -553.96240234375, | |
| "loss": 0.0336, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.0827620029449463, | |
| "rewards/margins": 1.3543148040771484, | |
| "rewards/rejected": -3.4370765686035156, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.6843132102963025e-07, | |
| "logits/chosen": -1.9901962280273438, | |
| "logits/rejected": -1.8920910358428955, | |
| "logps/chosen": -571.3399658203125, | |
| "logps/rejected": -549.0781860351562, | |
| "loss": 0.0364, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.1259512901306152, | |
| "rewards/margins": 1.4393842220306396, | |
| "rewards/rejected": -3.565335512161255, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1.591472578161458e-07, | |
| "logits/chosen": -2.0036253929138184, | |
| "logits/rejected": -1.9164316654205322, | |
| "logps/chosen": -597.6680908203125, | |
| "logps/rejected": -563.1033935546875, | |
| "loss": 0.0314, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -2.2230658531188965, | |
| "rewards/margins": 1.4771394729614258, | |
| "rewards/rejected": -3.7002053260803223, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.5000540787240274e-07, | |
| "logits/chosen": -1.9651165008544922, | |
| "logits/rejected": -1.9028050899505615, | |
| "logps/chosen": -573.396728515625, | |
| "logps/rejected": -580.0013427734375, | |
| "loss": 0.0282, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.4309842586517334, | |
| "rewards/margins": 1.3270018100738525, | |
| "rewards/rejected": -3.757986068725586, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.410200810842749e-07, | |
| "logits/chosen": -1.9549171924591064, | |
| "logits/rejected": -1.8664356470108032, | |
| "logps/chosen": -546.6082153320312, | |
| "logps/rejected": -557.6231689453125, | |
| "loss": 0.0273, | |
| "rewards/accuracies": 0.731249988079071, | |
| "rewards/chosen": -2.497020959854126, | |
| "rewards/margins": 1.3337533473968506, | |
| "rewards/rejected": -3.8307743072509766, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "eval_logits/chosen": -1.9557578563690186, | |
| "eval_logits/rejected": -1.9161585569381714, | |
| "eval_logps/chosen": -541.6861572265625, | |
| "eval_logps/rejected": -628.8740844726562, | |
| "eval_loss": 0.04087229445576668, | |
| "eval_rewards/accuracies": 0.70703125, | |
| "eval_rewards/chosen": -2.8464653491973877, | |
| "eval_rewards/margins": 0.8687426447868347, | |
| "eval_rewards/rejected": -3.715208053588867, | |
| "eval_runtime": 53.2271, | |
| "eval_samples_per_second": 37.575, | |
| "eval_steps_per_second": 0.601, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.322053423294041e-07, | |
| "logits/chosen": -1.9264421463012695, | |
| "logits/rejected": -1.8596795797348022, | |
| "logps/chosen": -555.9921875, | |
| "logps/rejected": -528.2036743164062, | |
| "loss": 0.0364, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.3642823696136475, | |
| "rewards/margins": 1.1525689363479614, | |
| "rewards/rejected": -3.5168514251708984, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.2357498946121905e-07, | |
| "logits/chosen": -2.047217845916748, | |
| "logits/rejected": -1.9547252655029297, | |
| "logps/chosen": -573.0296630859375, | |
| "logps/rejected": -531.27880859375, | |
| "loss": 0.04, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.1570980548858643, | |
| "rewards/margins": 1.236627459526062, | |
| "rewards/rejected": -3.393725633621216, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.1514253171093161e-07, | |
| "logits/chosen": -2.0023763179779053, | |
| "logits/rejected": -1.9147300720214844, | |
| "logps/chosen": -558.9900512695312, | |
| "logps/rejected": -542.7515869140625, | |
| "loss": 0.0424, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.013390064239502, | |
| "rewards/margins": 1.3009288311004639, | |
| "rewards/rejected": -3.314318895339966, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1.0692116854131883e-07, | |
| "logits/chosen": -1.9406598806381226, | |
| "logits/rejected": -1.8927663564682007, | |
| "logps/chosen": -502.59344482421875, | |
| "logps/rejected": -507.50372314453125, | |
| "loss": 0.0437, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.104419231414795, | |
| "rewards/margins": 1.1473711729049683, | |
| "rewards/rejected": -3.251791000366211, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 9.89237689853889e-08, | |
| "logits/chosen": -1.9217647314071655, | |
| "logits/rejected": -1.840735673904419, | |
| "logps/chosen": -539.6429443359375, | |
| "logps/rejected": -520.7713623046875, | |
| "loss": 0.0348, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -2.0584945678710938, | |
| "rewards/margins": 1.4084278345108032, | |
| "rewards/rejected": -3.4669222831726074, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 9.11628515022765e-08, | |
| "logits/chosen": -1.936431884765625, | |
| "logits/rejected": -1.8774102926254272, | |
| "logps/chosen": -507.42413330078125, | |
| "logps/rejected": -489.027099609375, | |
| "loss": 0.0321, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.9116840362548828, | |
| "rewards/margins": 1.3113796710968018, | |
| "rewards/rejected": -3.2230639457702637, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 8.365056438189486e-08, | |
| "logits/chosen": -1.9381834268569946, | |
| "logits/rejected": -1.8834428787231445, | |
| "logps/chosen": -545.0704345703125, | |
| "logps/rejected": -511.5694885253906, | |
| "loss": 0.0361, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.27673602104187, | |
| "rewards/margins": 1.0493533611297607, | |
| "rewards/rejected": -3.326089382171631, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.639866672902101e-08, | |
| "logits/chosen": -1.9358975887298584, | |
| "logits/rejected": -1.8624534606933594, | |
| "logps/chosen": -552.2891845703125, | |
| "logps/rejected": -548.515625, | |
| "loss": 0.0358, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.3348231315612793, | |
| "rewards/margins": 1.1951119899749756, | |
| "rewards/rejected": -3.529935121536255, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 6.941851005657851e-08, | |
| "logits/chosen": -1.9736747741699219, | |
| "logits/rejected": -1.8825995922088623, | |
| "logps/chosen": -592.0359497070312, | |
| "logps/rejected": -565.7269897460938, | |
| "loss": 0.0392, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.2588744163513184, | |
| "rewards/margins": 1.3530534505844116, | |
| "rewards/rejected": -3.6119277477264404, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 6.272102051693051e-08, | |
| "logits/chosen": -1.9536447525024414, | |
| "logits/rejected": -1.8848438262939453, | |
| "logps/chosen": -544.1442260742188, | |
| "logps/rejected": -547.0964965820312, | |
| "loss": 0.0377, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -2.0580058097839355, | |
| "rewards/margins": 1.3238760232925415, | |
| "rewards/rejected": -3.3818821907043457, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_logits/chosen": -1.967289686203003, | |
| "eval_logits/rejected": -1.9273663759231567, | |
| "eval_logps/chosen": -510.2101745605469, | |
| "eval_logps/rejected": -594.1712036132812, | |
| "eval_loss": 0.0495593398809433, | |
| "eval_rewards/accuracies": 0.72265625, | |
| "eval_rewards/chosen": -2.531705617904663, | |
| "eval_rewards/margins": 0.836473286151886, | |
| "eval_rewards/rejected": -3.3681788444519043, | |
| "eval_runtime": 53.2631, | |
| "eval_samples_per_second": 37.549, | |
| "eval_steps_per_second": 0.601, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 5.6316681798995844e-08, | |
| "logits/chosen": -1.9590046405792236, | |
| "logits/rejected": -1.8676058053970337, | |
| "logps/chosen": -530.419921875, | |
| "logps/rejected": -494.63720703125, | |
| "loss": 0.0395, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.950580358505249, | |
| "rewards/margins": 1.374487280845642, | |
| "rewards/rejected": -3.3250679969787598, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 5.0215518717961256e-08, | |
| "logits/chosen": -1.9477427005767822, | |
| "logits/rejected": -1.9216333627700806, | |
| "logps/chosen": -568.46337890625, | |
| "logps/rejected": -551.7440185546875, | |
| "loss": 0.0382, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.0898067951202393, | |
| "rewards/margins": 1.2489385604858398, | |
| "rewards/rejected": -3.3387451171875, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.4427081523275925e-08, | |
| "logits/chosen": -1.9811092615127563, | |
| "logits/rejected": -1.8924198150634766, | |
| "logps/chosen": -560.4685668945312, | |
| "logps/rejected": -545.5777587890625, | |
| "loss": 0.0384, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.166611433029175, | |
| "rewards/margins": 1.3478820323944092, | |
| "rewards/rejected": -3.514493465423584, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.896043094949061e-08, | |
| "logits/chosen": -1.952548623085022, | |
| "logits/rejected": -1.9003283977508545, | |
| "logps/chosen": -536.4952392578125, | |
| "logps/rejected": -554.130859375, | |
| "loss": 0.0368, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -2.2593836784362793, | |
| "rewards/margins": 1.1702790260314941, | |
| "rewards/rejected": -3.4296627044677734, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 3.3824124033343557e-08, | |
| "logits/chosen": -1.978271245956421, | |
| "logits/rejected": -1.8869798183441162, | |
| "logps/chosen": -595.8272094726562, | |
| "logps/rejected": -596.80908203125, | |
| "loss": 0.0356, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -2.2419838905334473, | |
| "rewards/margins": 1.5834529399871826, | |
| "rewards/rejected": -3.825437068939209, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 2.9026200719291904e-08, | |
| "logits/chosen": -1.9425359964370728, | |
| "logits/rejected": -1.8644979000091553, | |
| "logps/chosen": -541.2769775390625, | |
| "logps/rejected": -546.4044799804688, | |
| "loss": 0.0394, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -2.3050169944763184, | |
| "rewards/margins": 1.2533900737762451, | |
| "rewards/rejected": -3.5584073066711426, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 2.4574171274456433e-08, | |
| "logits/chosen": -1.9859609603881836, | |
| "logits/rejected": -1.8982467651367188, | |
| "logps/chosen": -581.1209716796875, | |
| "logps/rejected": -553.73291015625, | |
| "loss": 0.036, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -1.9945024251937866, | |
| "rewards/margins": 1.5215613842010498, | |
| "rewards/rejected": -3.516064167022705, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.047500453267881e-08, | |
| "logits/chosen": -1.9515501260757446, | |
| "logits/rejected": -1.857346534729004, | |
| "logps/chosen": -565.3553466796875, | |
| "logps/rejected": -536.8668823242188, | |
| "loss": 0.0422, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -2.2416954040527344, | |
| "rewards/margins": 1.2734299898147583, | |
| "rewards/rejected": -3.515125274658203, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 1.673511698609292e-08, | |
| "logits/chosen": -1.9258711338043213, | |
| "logits/rejected": -1.8735402822494507, | |
| "logps/chosen": -532.5035400390625, | |
| "logps/rejected": -557.9683837890625, | |
| "loss": 0.0354, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.2362167835235596, | |
| "rewards/margins": 1.4313932657241821, | |
| "rewards/rejected": -3.6676101684570312, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.3360362741285769e-08, | |
| "logits/chosen": -1.9208987951278687, | |
| "logits/rejected": -1.8273556232452393, | |
| "logps/chosen": -566.8057861328125, | |
| "logps/rejected": -583.4982299804688, | |
| "loss": 0.0352, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -2.093482732772827, | |
| "rewards/margins": 1.81829035282135, | |
| "rewards/rejected": -3.911773681640625, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "eval_logits/chosen": -1.9501041173934937, | |
| "eval_logits/rejected": -1.9090631008148193, | |
| "eval_logps/chosen": -521.0439453125, | |
| "eval_logps/rejected": -606.3505249023438, | |
| "eval_loss": 0.04646110162138939, | |
| "eval_rewards/accuracies": 0.72265625, | |
| "eval_rewards/chosen": -2.640043258666992, | |
| "eval_rewards/margins": 0.8499288558959961, | |
| "eval_rewards/rejected": -3.4899723529815674, | |
| "eval_runtime": 53.2044, | |
| "eval_samples_per_second": 37.591, | |
| "eval_steps_per_second": 0.601, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 1.0356024355769433e-08, | |
| "logits/chosen": -1.950596570968628, | |
| "logits/rejected": -1.882962942123413, | |
| "logps/chosen": -582.3919067382812, | |
| "logps/rejected": -592.8079223632812, | |
| "loss": 0.0361, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.399203062057495, | |
| "rewards/margins": 1.3655624389648438, | |
| "rewards/rejected": -3.7647652626037598, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 7.726804569108597e-09, | |
| "logits/chosen": -1.9748871326446533, | |
| "logits/rejected": -1.888811707496643, | |
| "logps/chosen": -553.1395263671875, | |
| "logps/rejected": -559.5152587890625, | |
| "loss": 0.0332, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -2.228854179382324, | |
| "rewards/margins": 1.4253013134002686, | |
| "rewards/rejected": -3.654155731201172, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 5.476818941645561e-09, | |
| "logits/chosen": -1.944598913192749, | |
| "logits/rejected": -1.8884025812149048, | |
| "logps/chosen": -540.9061279296875, | |
| "logps/rejected": -574.0789794921875, | |
| "loss": 0.036, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.298133134841919, | |
| "rewards/margins": 1.5589947700500488, | |
| "rewards/rejected": -3.8571276664733887, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 3.609589412347347e-09, | |
| "logits/chosen": -1.995321273803711, | |
| "logits/rejected": -1.909641981124878, | |
| "logps/chosen": -570.28369140625, | |
| "logps/rejected": -541.8267211914062, | |
| "loss": 0.0347, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -2.329395055770874, | |
| "rewards/margins": 1.3247489929199219, | |
| "rewards/rejected": -3.654143810272217, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.1280387858572667e-09, | |
| "logits/chosen": -1.9664127826690674, | |
| "logits/rejected": -1.8978990316390991, | |
| "logps/chosen": -567.4363403320312, | |
| "logps/rejected": -551.3905639648438, | |
| "loss": 0.0347, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -2.342649459838867, | |
| "rewards/margins": 1.2894551753997803, | |
| "rewards/rejected": -3.6321043968200684, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 1.03448615738172e-09, | |
| "logits/chosen": -1.91958749294281, | |
| "logits/rejected": -1.8549039363861084, | |
| "logps/chosen": -635.5400390625, | |
| "logps/rejected": -613.4205322265625, | |
| "loss": 0.0354, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -2.2879905700683594, | |
| "rewards/margins": 1.6663434505462646, | |
| "rewards/rejected": -3.954333782196045, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 3.3064328257259575e-10, | |
| "logits/chosen": -1.9804208278656006, | |
| "logits/rejected": -1.891332983970642, | |
| "logps/chosen": -612.157958984375, | |
| "logps/rejected": -556.913818359375, | |
| "loss": 0.0366, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -2.3392751216888428, | |
| "rewards/margins": 1.2889056205749512, | |
| "rewards/rejected": -3.628180980682373, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 1.7611898088715216e-11, | |
| "logits/chosen": -1.952614426612854, | |
| "logits/rejected": -1.846652626991272, | |
| "logps/chosen": -577.5618896484375, | |
| "logps/rejected": -596.773193359375, | |
| "loss": 0.0347, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -2.3046746253967285, | |
| "rewards/margins": 1.4788968563079834, | |
| "rewards/rejected": -3.783571720123291, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 883, | |
| "total_flos": 0.0, | |
| "train_loss": 0.06580274857555349, | |
| "train_runtime": 7965.4661, | |
| "train_samples_per_second": 14.19, | |
| "train_steps_per_second": 0.111 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 883, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |