Initial upload of alquistcoder_F1_MAIN_DPO model files

model-00001-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:81ace2f1dc017ba41ec9c842af014cea1a4317a7bcbf07a3c57cbb5ff62fe02f
+size 4975120816

model-00002-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4173f1d5bc73b6d7b9291defdd8112d3526c7697c029fa3845b9bc4b742277e4
+size 4832141488

model-00003-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7c87ffc81fa842dbfcd00f2951470b38abcaca8ab55d54b62840fdaac00c126
+size 4832141512

model-00004-of-00004.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e221988b20ec2da09324c678a57f4f7b8631ca2f6348c952ebd524d5de3c5ef
+size 3163092328

trainer_state.json

@@ -2,543 +2,288 @@
   "best_global_step": null,
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 2.0,
+  "epoch": 1.0,
   "eval_steps": 500,
-  "global_step": 356,
+  "global_step": 183,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.056338028169014086,
-      "grad_norm": 574.27978515625,
-      "learning_rate": 6.25e-07,
-      "logits/chosen": 2.945640802383423,
-      "logits/rejected": 3.0515265464782715,
-      "logps/chosen": -246.39541625976562,
-      "logps/rejected": -195.46109008789062,
-      "loss": 0.9304,
-      "rewards/accuracies": 0.5218750238418579,
-      "rewards/chosen": 0.03325582295656204,
-      "rewards/margins": 0.014068150892853737,
-      "rewards/rejected": 0.019187677651643753,
+      "epoch": 0.0547945205479452,
+      "grad_norm": 513.996826171875,
+      "learning_rate": 4.090909090909091e-07,
+      "logits/chosen": 2.990995407104492,
+      "logits/rejected": 3.0081257820129395,
+      "logps/chosen": -298.52886962890625,
+      "logps/rejected": -202.96295166015625,
+      "loss": 0.8843,
+      "rewards/accuracies": 0.526562511920929,
+      "rewards/chosen": 0.010423189960420132,
+      "rewards/margins": 0.05534166842699051,
+      "rewards/rejected": -0.0449184887111187,
       "step": 10
     },
     {
-      "epoch": 0.11267605633802817,
-      "grad_norm": 267.4449462890625,
-      "learning_rate": 1.3194444444444446e-06,
-      "logits/chosen": 3.0365357398986816,
-      "logits/rejected": 3.0756676197052,
-      "logps/chosen": -231.8340606689453,
-      "logps/rejected": -196.01499938964844,
-      "loss": 0.5293,
-      "rewards/accuracies": 0.737500011920929,
-      "rewards/chosen": 1.023498773574829,
-      "rewards/margins": 1.6332218647003174,
-      "rewards/rejected": -0.6097229719161987,
+      "epoch": 0.1095890410958904,
+      "grad_norm": 461.9533996582031,
+      "learning_rate": 8.636363636363637e-07,
+      "logits/chosen": 2.934217929840088,
+      "logits/rejected": 2.919574737548828,
+      "logps/chosen": -276.70391845703125,
+      "logps/rejected": -200.52728271484375,
+      "loss": 0.6858,
+      "rewards/accuracies": 0.65625,
+      "rewards/chosen": 0.7733574509620667,
+      "rewards/margins": 0.9247980117797852,
+      "rewards/rejected": -0.15144045650959015,
       "step": 20
     },
     {
-      "epoch": 0.16901408450704225,
-      "grad_norm": 168.65689086914062,
-      "learning_rate": 2.0138888888888893e-06,
-      "logits/chosen": 3.038414478302002,
-      "logits/rejected": 3.0314807891845703,
-      "logps/chosen": -265.27825927734375,
-      "logps/rejected": -199.1730499267578,
-      "loss": 0.1929,
-      "rewards/accuracies": 0.925000011920929,
-      "rewards/chosen": 4.1361799240112305,
-      "rewards/margins": 6.971987724304199,
-      "rewards/rejected": -2.835808277130127,
+      "epoch": 0.1643835616438356,
+      "grad_norm": 150.57086181640625,
+      "learning_rate": 1.3181818181818182e-06,
+      "logits/chosen": 3.0546913146972656,
+      "logits/rejected": 3.0510308742523193,
+      "logps/chosen": -289.57977294921875,
+      "logps/rejected": -218.24765014648438,
+      "loss": 0.2839,
+      "rewards/accuracies": 0.8890625238418579,
+      "rewards/chosen": 2.8807666301727295,
+      "rewards/margins": 4.351069450378418,
+      "rewards/rejected": -1.4703023433685303,
       "step": 30
     },
     {
-      "epoch": 0.22535211267605634,
-      "grad_norm": 37.93267822265625,
-      "learning_rate": 2.7083333333333334e-06,
-      "logits/chosen": 3.0692856311798096,
-      "logits/rejected": 3.1041903495788574,
-      "logps/chosen": -245.3774871826172,
-      "logps/rejected": -207.87789916992188,
-      "loss": 0.1002,
-      "rewards/accuracies": 0.9671875238418579,
-      "rewards/chosen": 6.395417213439941,
-      "rewards/margins": 13.0038480758667,
-      "rewards/rejected": -6.608429908752441,
+      "epoch": 0.2191780821917808,
+      "grad_norm": 151.46690368652344,
+      "learning_rate": 1.7727272727272729e-06,
+      "logits/chosen": 3.088348150253296,
+      "logits/rejected": 3.1125292778015137,
+      "logps/chosen": -275.5274963378906,
+      "logps/rejected": -214.15737915039062,
+      "loss": 0.1502,
+      "rewards/accuracies": 0.9390624761581421,
+      "rewards/chosen": 5.216189384460449,
+      "rewards/margins": 8.827147483825684,
+      "rewards/rejected": -3.610957384109497,
       "step": 40
     },
     {
-      "epoch": 0.28169014084507044,
-      "grad_norm": 129.71878051757812,
-      "learning_rate": 3.4027777777777783e-06,
-      "logits/chosen": 2.943250894546509,
-      "logits/rejected": 2.9869542121887207,
-      "logps/chosen": -245.9439239501953,
-      "logps/rejected": -202.8123321533203,
-      "loss": 0.1375,
-      "rewards/accuracies": 0.973437488079071,
-      "rewards/chosen": 7.993535041809082,
-      "rewards/margins": 19.0964412689209,
-      "rewards/rejected": -11.102909088134766,
+      "epoch": 0.273972602739726,
+      "grad_norm": 130.53453063964844,
+      "learning_rate": 2.2272727272727274e-06,
+      "logits/chosen": 2.9992308616638184,
+      "logits/rejected": 3.075270175933838,
+      "logps/chosen": -259.50836181640625,
+      "logps/rejected": -203.08811950683594,
+      "loss": 0.1281,
+      "rewards/accuracies": 0.957812488079071,
+      "rewards/chosen": 6.97296667098999,
+      "rewards/margins": 12.8120698928833,
+      "rewards/rejected": -5.839103698730469,
       "step": 50
     },
     {
-      "epoch": 0.3380281690140845,
-      "grad_norm": 57.64502716064453,
-      "learning_rate": 4.097222222222222e-06,
-      "logits/chosen": 3.0049643516540527,
-      "logits/rejected": 3.0770554542541504,
-      "logps/chosen": -236.73410034179688,
-      "logps/rejected": -210.151611328125,
-      "loss": 0.1722,
-      "rewards/accuracies": 0.965624988079071,
-      "rewards/chosen": 9.21152114868164,
-      "rewards/margins": 22.748985290527344,
-      "rewards/rejected": -13.53746509552002,
+      "epoch": 0.3287671232876712,
+      "grad_norm": 110.53746032714844,
+      "learning_rate": 2.6818181818181822e-06,
+      "logits/chosen": 3.088064670562744,
+      "logits/rejected": 2.986386775970459,
+      "logps/chosen": -325.0431213378906,
+      "logps/rejected": -224.979736328125,
+      "loss": 0.1444,
+      "rewards/accuracies": 0.964062511920929,
+      "rewards/chosen": 10.117586135864258,
+      "rewards/margins": 18.425701141357422,
+      "rewards/rejected": -8.30811595916748,
       "step": 60
     },
     {
-      "epoch": 0.39436619718309857,
-      "grad_norm": 105.868408203125,
-      "learning_rate": 4.791666666666668e-06,
-      "logits/chosen": 3.0406033992767334,
-      "logits/rejected": 3.0739083290100098,
-      "logps/chosen": -256.1812744140625,
-      "logps/rejected": -204.35446166992188,
-      "loss": 0.1129,
-      "rewards/accuracies": 0.979687511920929,
-      "rewards/chosen": 8.951081275939941,
-      "rewards/margins": 21.80655860900879,
-      "rewards/rejected": -12.85547924041748,
+      "epoch": 0.3835616438356164,
+      "grad_norm": 95.7762222290039,
+      "learning_rate": 3.1363636363636367e-06,
+      "logits/chosen": 3.040494918823242,
+      "logits/rejected": 3.022307872772217,
+      "logps/chosen": -273.9972839355469,
+      "logps/rejected": -211.2712860107422,
+      "loss": 0.1443,
+      "rewards/accuracies": 0.9671875238418579,
+      "rewards/chosen": 9.627126693725586,
+      "rewards/margins": 20.590023040771484,
+      "rewards/rejected": -10.962896347045898,
       "step": 70
     },
     {
-      "epoch": 0.4507042253521127,
-      "grad_norm": 168.75357055664062,
-      "learning_rate": 4.998524282731094e-06,
-      "logits/chosen": 3.0504136085510254,
-      "logits/rejected": 3.0491514205932617,
-      "logps/chosen": -255.90316772460938,
-      "logps/rejected": -206.5497283935547,
-      "loss": 0.3368,
-      "rewards/accuracies": 0.96875,
-      "rewards/chosen": 8.498771667480469,
-      "rewards/margins": 22.880569458007812,
-      "rewards/rejected": -14.381797790527344,
+      "epoch": 0.4383561643835616,
+      "grad_norm": 179.24440002441406,
+      "learning_rate": 3.590909090909091e-06,
+      "logits/chosen": 3.0298266410827637,
+      "logits/rejected": 3.0730605125427246,
+      "logps/chosen": -280.2432861328125,
+      "logps/rejected": -214.884033203125,
+      "loss": 0.142,
+      "rewards/accuracies": 0.979687511920929,
+      "rewards/chosen": 10.719534873962402,
+      "rewards/margins": 23.82217788696289,
+      "rewards/rejected": -13.102640151977539,
       "step": 80
     },
     {
-      "epoch": 0.5070422535211268,
-      "grad_norm": 109.43980407714844,
-      "learning_rate": 4.991300473502437e-06,
-      "logits/chosen": 3.0595602989196777,
-      "logits/rejected": 3.1263351440429688,
-      "logps/chosen": -228.88070678710938,
-      "logps/rejected": -218.99783325195312,
-      "loss": 0.3032,
-      "rewards/accuracies": 0.9624999761581421,
-      "rewards/chosen": 6.379992485046387,
-      "rewards/margins": 28.722070693969727,
-      "rewards/rejected": -22.34208106994629,
+      "epoch": 0.4931506849315068,
+      "grad_norm": 2.0804860591888428,
+      "learning_rate": 4.045454545454546e-06,
+      "logits/chosen": 3.1072518825531006,
+      "logits/rejected": 3.067288875579834,
+      "logps/chosen": -294.1097106933594,
+      "logps/rejected": -219.38949584960938,
+      "loss": 0.1262,
+      "rewards/accuracies": 0.981249988079071,
+      "rewards/chosen": 12.069157600402832,
+      "rewards/margins": 29.454524993896484,
+      "rewards/rejected": -17.385366439819336,
       "step": 90
     },
     {
-      "epoch": 0.5633802816901409,
-      "grad_norm": 128.98760986328125,
-      "learning_rate": 4.978074903220964e-06,
-      "logits/chosen": 3.1774604320526123,
-      "logits/rejected": 3.181511402130127,
-      "logps/chosen": -228.3056640625,
-      "logps/rejected": -222.18038940429688,
-      "loss": 0.2029,
-      "rewards/accuracies": 0.9781249761581421,
-      "rewards/chosen": 5.856657028198242,
-      "rewards/margins": 34.18266677856445,
-      "rewards/rejected": -28.326007843017578,
+      "epoch": 0.547945205479452,
+      "grad_norm": 134.90240478515625,
+      "learning_rate": 4.5e-06,
+      "logits/chosen": 2.9695353507995605,
+      "logits/rejected": 2.9900407791137695,
+      "logps/chosen": -270.2259826660156,
+      "logps/rejected": -210.32302856445312,
+      "loss": 0.1125,
+      "rewards/accuracies": 0.984375,
+      "rewards/chosen": 8.642024040222168,
+      "rewards/margins": 26.856210708618164,
+      "rewards/rejected": -18.21418571472168,
       "step": 100
     },
     {
-      "epoch": 0.6197183098591549,
-      "grad_norm": 144.1073455810547,
-      "learning_rate": 4.958879433443904e-06,
-      "logits/chosen": 3.218785047531128,
-      "logits/rejected": 3.1809394359588623,
-      "logps/chosen": -243.4155731201172,
-      "logps/rejected": -236.49868774414062,
-      "loss": 0.2134,
+      "epoch": 0.6027397260273972,
+      "grad_norm": 237.83163452148438,
+      "learning_rate": 4.954545454545455e-06,
+      "logits/chosen": 3.001239538192749,
+      "logits/rejected": 2.9165444374084473,
+      "logps/chosen": -261.63848876953125,
+      "logps/rejected": -217.56314086914062,
+      "loss": 0.1337,
       "rewards/accuracies": 0.9765625,
-      "rewards/chosen": 6.719731330871582,
-      "rewards/margins": 35.10533905029297,
-      "rewards/rejected": -28.385604858398438,
+      "rewards/chosen": 4.362582206726074,
+      "rewards/margins": 21.842912673950195,
+      "rewards/rejected": -17.480329513549805,
       "step": 110
     },
     {
-      "epoch": 0.676056338028169,
-      "grad_norm": 149.68939208984375,
-      "learning_rate": 4.933760307739277e-06,
-      "logits/chosen": 3.1437926292419434,
-      "logits/rejected": 3.1433277130126953,
-      "logps/chosen": -255.87796020507812,
-      "logps/rejected": -231.52182006835938,
-      "loss": 0.2745,
-      "rewards/accuracies": 0.981249988079071,
-      "rewards/chosen": 6.636951446533203,
-      "rewards/margins": 37.604583740234375,
-      "rewards/rejected": -30.967632293701172,
+      "epoch": 0.6575342465753424,
+      "grad_norm": 124.83686065673828,
+      "learning_rate": 4.998976350571773e-06,
+      "logits/chosen": 3.0631394386291504,
+      "logits/rejected": 3.0034124851226807,
+      "logps/chosen": -296.0355224609375,
+      "logps/rejected": -219.4881134033203,
+      "loss": 0.176,
+      "rewards/accuracies": 0.96875,
+      "rewards/chosen": 3.6324734687805176,
+      "rewards/margins": 20.854042053222656,
+      "rewards/rejected": -17.221569061279297,
       "step": 120
     },
     {
-      "epoch": 0.7323943661971831,
-      "grad_norm": 171.0680389404297,
-      "learning_rate": 4.90277804028108e-06,
-      "logits/chosen": 3.206712245941162,
-      "logits/rejected": 3.2256839275360107,
-      "logps/chosen": -261.3406982421875,
-      "logps/rejected": -228.3511505126953,
-      "loss": 0.5352,
-      "rewards/accuracies": 0.964062511920929,
-      "rewards/chosen": 4.778416633605957,
-      "rewards/margins": 33.36115646362305,
-      "rewards/rejected": -28.58274269104004,
+      "epoch": 0.7123287671232876,
+      "grad_norm": 110.8822250366211,
+      "learning_rate": 4.995438885558294e-06,
+      "logits/chosen": 3.0476179122924805,
+      "logits/rejected": 2.9690792560577393,
+      "logps/chosen": -292.52276611328125,
+      "logps/rejected": -210.3925018310547,
+      "loss": 0.2762,
+      "rewards/accuracies": 0.9781249761581421,
+      "rewards/chosen": 5.373471736907959,
+      "rewards/margins": 26.076580047607422,
+      "rewards/rejected": -20.703105926513672,
       "step": 130
     },
     {
-      "epoch": 0.7887323943661971,
-      "grad_norm": 118.76382446289062,
-      "learning_rate": 4.866007270065345e-06,
-      "logits/chosen": 3.2129406929016113,
-      "logits/rejected": 3.227287769317627,
-      "logps/chosen": -246.7953643798828,
-      "logps/rejected": -230.45651245117188,
-      "loss": 0.2311,
-      "rewards/accuracies": 0.9765625,
-      "rewards/chosen": 0.6871539354324341,
-      "rewards/margins": 36.90045166015625,
-      "rewards/rejected": -36.213294982910156,
+      "epoch": 0.7671232876712328,
+      "grad_norm": 72.18496704101562,
+      "learning_rate": 4.989378542821969e-06,
+      "logits/chosen": 3.0710926055908203,
+      "logits/rejected": 3.0577285289764404,
+      "logps/chosen": -284.55230712890625,
+      "logps/rejected": -230.9425506591797,
+      "loss": 0.2372,
+      "rewards/accuracies": 0.973437488079071,
+      "rewards/chosen": 5.454714775085449,
+      "rewards/margins": 30.868602752685547,
+      "rewards/rejected": -25.413890838623047,
       "step": 140
     },
     {
-      "epoch": 0.8450704225352113,
-      "grad_norm": 198.268310546875,
-      "learning_rate": 4.823536581098262e-06,
-      "logits/chosen": 3.1713156700134277,
-      "logits/rejected": 3.152228355407715,
-      "logps/chosen": -238.29159545898438,
-      "logps/rejected": -238.2996826171875,
-      "loss": 0.2612,
-      "rewards/accuracies": 0.96875,
-      "rewards/chosen": 1.5464673042297363,
-      "rewards/margins": 41.4525146484375,
-      "rewards/rejected": -39.90605163574219,
+      "epoch": 0.821917808219178,
+      "grad_norm": 104.61406707763672,
+      "learning_rate": 4.9808014493426124e-06,
+      "logits/chosen": 3.053307294845581,
+      "logits/rejected": 3.0027899742126465,
+      "logps/chosen": -282.54864501953125,
+      "logps/rejected": -236.41592407226562,
+      "loss": 0.6492,
+      "rewards/accuracies": 0.9546874761581421,
+      "rewards/chosen": 0.29179587960243225,
+      "rewards/margins": 30.52816390991211,
+      "rewards/rejected": -30.23636817932129,
       "step": 150
     },
     {
-      "epoch": 0.9014084507042254,
-      "grad_norm": 191.44915771484375,
-      "learning_rate": 4.775468288989545e-06,
-      "logits/chosen": 3.2888169288635254,
-      "logits/rejected": 3.2170681953430176,
-      "logps/chosen": -251.0401153564453,
-      "logps/rejected": -236.00479125976562,
-      "loss": 0.51,
-      "rewards/accuracies": 0.971875011920929,
-      "rewards/chosen": 2.7297756671905518,
-      "rewards/margins": 42.41936492919922,
-      "rewards/rejected": -39.68959426879883,
+      "epoch": 0.8767123287671232,
+      "grad_norm": 114.0179443359375,
+      "learning_rate": 4.9697162765239595e-06,
+      "logits/chosen": 3.0813591480255127,
+      "logits/rejected": 3.093292713165283,
+      "logps/chosen": -265.5400390625,
+      "logps/rejected": -233.4171142578125,
+      "loss": 0.1703,
+      "rewards/accuracies": 0.981249988079071,
+      "rewards/chosen": 2.905339002609253,
+      "rewards/margins": 36.841583251953125,
+      "rewards/rejected": -33.936241149902344,
       "step": 160
     },
     {
-      "epoch": 0.9577464788732394,
-      "grad_norm": 73.10994720458984,
-      "learning_rate": 4.7219181944651695e-06,
-      "logits/chosen": 3.2599377632141113,
-      "logits/rejected": 3.2560737133026123,
-      "logps/chosen": -245.1952667236328,
-      "logps/rejected": -241.4735565185547,
-      "loss": 0.313,
-      "rewards/accuracies": 0.973437488079071,
-      "rewards/chosen": -0.8346878290176392,
-      "rewards/margins": 44.581295013427734,
-      "rewards/rejected": -45.415985107421875,
+      "epoch": 0.9315068493150684,
+      "grad_norm": 132.4547576904297,
+      "learning_rate": 4.9561342314269055e-06,
+      "logits/chosen": 3.124277114868164,
+      "logits/rejected": 3.048166036605835,
+      "logps/chosen": -262.30194091796875,
+      "logps/rejected": -232.39297485351562,
+      "loss": 0.4256,
+      "rewards/accuracies": 0.9515625238418579,
+      "rewards/chosen": -1.3073980808258057,
+      "rewards/margins": 30.139415740966797,
+      "rewards/rejected": -31.446813583374023,
       "step": 170
     },
     {
-      "epoch": 1.0112676056338028,
-      "grad_norm": 71.17866516113281,
-      "learning_rate": 4.663015304393279e-06,
-      "logits/chosen": 3.3155758380889893,
-      "logits/rejected": 3.3333241939544678,
-      "logps/chosen": -270.6006774902344,
-      "logps/rejected": -241.6360321044922,
-      "loss": 0.338,
-      "rewards/accuracies": 0.9736841917037964,
-      "rewards/chosen": 0.171695277094841,
-      "rewards/margins": 45.20180130004883,
-      "rewards/rejected": -45.03010177612305,
+      "epoch": 0.9863013698630136,
+      "grad_norm": 1.0581492185592651,
+      "learning_rate": 4.940069045439226e-06,
+      "logits/chosen": 3.164407968521118,
+      "logits/rejected": 3.164742946624756,
+      "logps/chosen": -274.58673095703125,
+      "logps/rejected": -232.9495849609375,
+      "loss": 0.3806,
+      "rewards/accuracies": 0.9703124761581421,
+      "rewards/chosen": 0.4807693064212799,
+      "rewards/margins": 33.424041748046875,
+      "rewards/rejected": -32.943275451660156,
       "step": 180
-    },
-    {
-      "epoch": 1.0676056338028168,
-      "grad_norm": 113.73174285888672,
-      "learning_rate": 4.59890152099534e-06,
-      "logits/chosen": 3.508065700531006,
-      "logits/rejected": 3.5083301067352295,
-      "logps/chosen": -266.78387451171875,
-      "logps/rejected": -252.5698699951172,
-      "loss": 0.8076,
-      "rewards/accuracies": 0.964062511920929,
-      "rewards/chosen": 10.96754264831543,
-      "rewards/margins": 63.106483459472656,
-      "rewards/rejected": -52.138938903808594,
-      "step": 190
-    },
-    {
-      "epoch": 1.123943661971831,
-      "grad_norm": 0.890201985836029,
-      "learning_rate": 4.5297312999912625e-06,
-      "logits/chosen": 3.5750420093536377,
-      "logits/rejected": 3.5598015785217285,
-      "logps/chosen": -244.08615112304688,
-      "logps/rejected": -249.61984252929688,
-      "loss": 0.4625,
-      "rewards/accuracies": 0.9765625,
-      "rewards/chosen": 8.597976684570312,
-      "rewards/margins": 65.22010803222656,
-      "rewards/rejected": -56.62213897705078,
-      "step": 200
-    },
-    {
-      "epoch": 1.180281690140845,
-      "grad_norm": 204.70803833007812,
-      "learning_rate": 4.455671278502042e-06,
-      "logits/chosen": 3.6519992351531982,
-      "logits/rejected": 3.6381404399871826,
-      "logps/chosen": -240.0857696533203,
-      "logps/rejected": -238.0441131591797,
-      "loss": 0.653,
-      "rewards/accuracies": 0.9546874761581421,
-      "rewards/chosen": 6.305617332458496,
-      "rewards/margins": 53.15886688232422,
-      "rewards/rejected": -46.8532600402832,
-      "step": 210
-    },
-    {
-      "epoch": 1.236619718309859,
-      "grad_norm": 139.86141967773438,
-      "learning_rate": 4.376899873606336e-06,
-      "logits/chosen": 3.622997283935547,
-      "logits/rejected": 3.6359076499938965,
-      "logps/chosen": -259.12725830078125,
-      "logps/rejected": -250.89501953125,
-      "loss": 0.9057,
-      "rewards/accuracies": 0.9453125,
-      "rewards/chosen": 4.471914291381836,
-      "rewards/margins": 51.15581130981445,
-      "rewards/rejected": -46.683895111083984,
-      "step": 220
-    },
-    {
-      "epoch": 1.2929577464788733,
-      "grad_norm": 94.85035705566406,
-      "learning_rate": 4.293606852518101e-06,
-      "logits/chosen": 3.6598830223083496,
-      "logits/rejected": 3.594897508621216,
-      "logps/chosen": -263.00079345703125,
-      "logps/rejected": -248.59524536132812,
-      "loss": 0.3918,
-      "rewards/accuracies": 0.9750000238418579,
-      "rewards/chosen": 2.2544591426849365,
-      "rewards/margins": 56.7413444519043,
-      "rewards/rejected": -54.48688507080078,
-      "step": 230
-    },
-    {
-      "epoch": 1.3492957746478873,
-      "grad_norm": 37.788360595703125,
-      "learning_rate": 4.205992875420742e-06,
-      "logits/chosen": 3.57995343208313,
-      "logits/rejected": 3.5664329528808594,
-      "logps/chosen": -243.8499755859375,
-      "logps/rejected": -252.65353393554688,
-      "loss": 0.4873,
-      "rewards/accuracies": 0.9781249761581421,
-      "rewards/chosen": 7.1952104568481445,
-      "rewards/margins": 64.06478118896484,
-      "rewards/rejected": -56.86956787109375,
-      "step": 240
-    },
-    {
-      "epoch": 1.4056338028169013,
-      "grad_norm": 159.25875854492188,
-      "learning_rate": 4.114269012059169e-06,
-      "logits/chosen": 3.5491080284118652,
-      "logits/rejected": 3.510200023651123,
-      "logps/chosen": -253.5936279296875,
-      "logps/rejected": -254.4651336669922,
-      "loss": 0.3683,
-      "rewards/accuracies": 0.9624999761581421,
-      "rewards/chosen": 4.560586452484131,
-      "rewards/margins": 59.49503707885742,
-      "rewards/rejected": -54.9344482421875,
-      "step": 250
-    },
-    {
-      "epoch": 1.4619718309859155,
-      "grad_norm": 113.43865966796875,
-      "learning_rate": 4.018656233254278e-06,
-      "logits/chosen": 3.4692349433898926,
-      "logits/rejected": 3.3723576068878174,
-      "logps/chosen": -253.46249389648438,
-      "logps/rejected": -268.2179870605469,
-      "loss": 0.3905,
-      "rewards/accuracies": 0.984375,
-      "rewards/chosen": 3.9269375801086426,
-      "rewards/margins": 74.34013366699219,
-      "rewards/rejected": -70.41320037841797,
-      "step": 260
-    },
-    {
-      "epoch": 1.5183098591549296,
-      "grad_norm": 36.17473602294922,
-      "learning_rate": 3.919384878564902e-06,
-      "logits/chosen": 3.488018035888672,
-      "logits/rejected": 3.4679126739501953,
-      "logps/chosen": -248.9741668701172,
-      "logps/rejected": -282.92840576171875,
-      "loss": 0.5658,
-      "rewards/accuracies": 0.9781249761581421,
-      "rewards/chosen": 4.083906650543213,
-      "rewards/margins": 77.73112487792969,
-      "rewards/rejected": -73.647216796875,
-      "step": 270
-    },
-    {
-      "epoch": 1.5746478873239438,
-      "grad_norm": 177.54251098632812,
-      "learning_rate": 3.816694101379631e-06,
-      "logits/chosen": 3.6568634510040283,
-      "logits/rejected": 3.6167445182800293,
-      "logps/chosen": -240.26895141601562,
-      "logps/rejected": -263.4219665527344,
-      "loss": 0.3476,
-      "rewards/accuracies": 0.979687511920929,
-      "rewards/chosen": 4.189328193664551,
-      "rewards/margins": 66.97871398925781,
-      "rewards/rejected": -62.78938674926758,
-      "step": 280
-    },
-    {
-      "epoch": 1.6309859154929578,
-      "grad_norm": 76.29850006103516,
-      "learning_rate": 3.7108312927753533e-06,
-      "logits/chosen": 3.5761330127716064,
-      "logits/rejected": 3.5349583625793457,
-      "logps/chosen": -256.83697509765625,
-      "logps/rejected": -257.34063720703125,
-      "loss": 0.2007,
-      "rewards/accuracies": 0.987500011920929,
-      "rewards/chosen": 3.5510315895080566,
-      "rewards/margins": 67.4579849243164,
-      "rewards/rejected": -63.906944274902344,
-      "step": 290
-    },
-    {
-      "epoch": 1.6873239436619718,
-      "grad_norm": 88.87206268310547,
-      "learning_rate": 3.6020514855304856e-06,
-      "logits/chosen": 3.4625442028045654,
-      "logits/rejected": 3.4353599548339844,
-      "logps/chosen": -246.8657684326172,
-      "logps/rejected": -262.365966796875,
-      "loss": 0.3359,
-      "rewards/accuracies": 0.9765625,
-      "rewards/chosen": -1.413503885269165,
-      "rewards/margins": 65.82698059082031,
-      "rewards/rejected": -67.240478515625,
-      "step": 300
-    },
-    {
-      "epoch": 1.7436619718309858,
-      "grad_norm": 109.3917007446289,
-      "learning_rate": 3.4906167397286643e-06,
-      "logits/chosen": 3.4818034172058105,
-      "logits/rejected": 3.438699722290039,
-      "logps/chosen": -248.93112182617188,
-      "logps/rejected": -257.42694091796875,
-      "loss": 0.4265,
-      "rewards/accuracies": 0.9828125238418579,
-      "rewards/chosen": 0.5807734131813049,
-      "rewards/margins": 62.88755416870117,
-      "rewards/rejected": -62.306793212890625,
-      "step": 310
-    },
-    {
-      "epoch": 1.8,
-      "grad_norm": 88.23355102539062,
-      "learning_rate": 3.3767955114330586e-06,
-      "logits/chosen": 3.3789494037628174,
-      "logits/rejected": 3.329315662384033,
-      "logps/chosen": -258.67291259765625,
-      "logps/rejected": -259.5810546875,
-      "loss": 0.4489,
-      "rewards/accuracies": 0.9624999761581421,
-      "rewards/chosen": -2.7831504344940186,
-      "rewards/margins": 59.0428466796875,
-      "rewards/rejected": -61.82598876953125,
-      "step": 320
-    },
-    {
-      "epoch": 1.856338028169014,
-      "grad_norm": 1.0168672800064087,
-      "learning_rate": 3.2608620059521935e-06,
-      "logits/chosen": 3.38688325881958,
-      "logits/rejected": 3.349194288253784,
-      "logps/chosen": -233.11666870117188,
-      "logps/rejected": -270.48590087890625,
-      "loss": 0.5576,
-      "rewards/accuracies": 0.979687511920929,
-      "rewards/chosen": -4.500494480133057,
-      "rewards/margins": 67.32182312011719,
-      "rewards/rejected": -71.82231140136719,
-      "step": 330
-    },
-    {
-      "epoch": 1.9126760563380283,
-      "grad_norm": 106.62237548828125,
-      "learning_rate": 3.14309551725535e-06,
-      "logits/chosen": 3.401437759399414,
-      "logits/rejected": 3.3626091480255127,
-      "logps/chosen": -247.9010009765625,
-      "logps/rejected": -256.19097900390625,
-      "loss": 0.4024,
-      "rewards/accuracies": 0.979687511920929,
-      "rewards/chosen": 0.5352304577827454,
-      "rewards/margins": 63.979896545410156,
-      "rewards/rejected": -63.444664001464844,
-      "step": 340
-    },
-    {
-      "epoch": 1.9690140845070423,
-      "grad_norm": 159.5343475341797,
-      "learning_rate": 3.0237797551289228e-06,
-      "logits/chosen": 3.4421494007110596,
-      "logits/rejected": 3.4059250354766846,
-      "logps/chosen": -229.91134643554688,
-      "logps/rejected": -252.29019165039062,
-      "loss": 0.6959,
-      "rewards/accuracies": 0.971875011920929,
-      "rewards/chosen": 0.6608916521072388,
-      "rewards/margins": 60.5278205871582,
-      "rewards/rejected": -59.86692428588867,
-      "step": 350
     }
   ],
   "logging_steps": 10,
-  "max_steps": 712,
+  "max_steps": 1098,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 4,
+  "num_train_epochs": 6,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7789fd19e3b22d6de6b1b2cd69ff66551be72d745bf4062a2a77caaab0386055
+size 6520