phi-2-gpo-v8-i1 / trainer_state.json
lole25's picture
Model save
94c355f verified
raw
history blame
90.6 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 1.0,
"eval_steps": 500,
"global_step": 1875,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.0,
"learning_rate": 2.6595744680851065e-08,
"logits/chosen": 0.24484609067440033,
"logits/rejected": 0.2733721137046814,
"logps/chosen": -382.60247802734375,
"logps/rejected": -260.64013671875,
"loss": 0.3272,
"rewards/accuracies": 0.0,
"rewards/chosen": 0.0,
"rewards/margins": 0.0,
"rewards/rejected": 0.0,
"step": 1
},
{
"epoch": 0.01,
"learning_rate": 2.6595744680851066e-07,
"logits/chosen": 0.04333009943366051,
"logits/rejected": 0.2742147743701935,
"logps/chosen": -312.38092041015625,
"logps/rejected": -271.8465881347656,
"loss": 0.3266,
"rewards/accuracies": 0.4305555522441864,
"rewards/chosen": -6.604947702726349e-05,
"rewards/margins": -4.0390281355939806e-05,
"rewards/rejected": -2.565919203334488e-05,
"step": 10
},
{
"epoch": 0.01,
"learning_rate": 5.319148936170213e-07,
"logits/chosen": 0.14737281203269958,
"logits/rejected": 0.26748910546302795,
"logps/chosen": -351.84014892578125,
"logps/rejected": -283.82989501953125,
"loss": 0.3538,
"rewards/accuracies": 0.4437499940395355,
"rewards/chosen": 4.673796411225339e-06,
"rewards/margins": -5.658002191921696e-05,
"rewards/rejected": 6.125382788013667e-05,
"step": 20
},
{
"epoch": 0.02,
"learning_rate": 7.97872340425532e-07,
"logits/chosen": 0.08816106617450714,
"logits/rejected": 0.21678781509399414,
"logps/chosen": -344.03814697265625,
"logps/rejected": -279.57830810546875,
"loss": 0.3558,
"rewards/accuracies": 0.40625,
"rewards/chosen": -6.16505931247957e-05,
"rewards/margins": -0.0001463174121454358,
"rewards/rejected": 8.466680446872488e-05,
"step": 30
},
{
"epoch": 0.02,
"learning_rate": 1.0638297872340427e-06,
"logits/chosen": 0.12371940910816193,
"logits/rejected": 0.18927700817584991,
"logps/chosen": -326.63323974609375,
"logps/rejected": -280.051025390625,
"loss": 0.2908,
"rewards/accuracies": 0.45625001192092896,
"rewards/chosen": 0.00011619376164162531,
"rewards/margins": -1.4456210010393988e-05,
"rewards/rejected": 0.00013064996164757758,
"step": 40
},
{
"epoch": 0.03,
"learning_rate": 1.3297872340425533e-06,
"logits/chosen": 0.03745291382074356,
"logits/rejected": 0.17850083112716675,
"logps/chosen": -333.9073486328125,
"logps/rejected": -289.7876892089844,
"loss": 0.3622,
"rewards/accuracies": 0.5687500238418579,
"rewards/chosen": 0.00023793391301296651,
"rewards/margins": 0.00018272039596922696,
"rewards/rejected": 5.521355342352763e-05,
"step": 50
},
{
"epoch": 0.03,
"learning_rate": 1.595744680851064e-06,
"logits/chosen": 0.05472108721733093,
"logits/rejected": 0.256857305765152,
"logps/chosen": -351.0667419433594,
"logps/rejected": -273.13787841796875,
"loss": 0.34,
"rewards/accuracies": 0.5562499761581421,
"rewards/chosen": 0.000246922456426546,
"rewards/margins": 0.00023249602236319333,
"rewards/rejected": 1.4426443158299662e-05,
"step": 60
},
{
"epoch": 0.04,
"learning_rate": 1.8617021276595745e-06,
"logits/chosen": 0.05988805368542671,
"logits/rejected": 0.22936634719371796,
"logps/chosen": -292.13360595703125,
"logps/rejected": -273.4328308105469,
"loss": 0.339,
"rewards/accuracies": 0.581250011920929,
"rewards/chosen": 0.0005934828659519553,
"rewards/margins": 0.000584075809456408,
"rewards/rejected": 9.407010111317504e-06,
"step": 70
},
{
"epoch": 0.04,
"learning_rate": 2.1276595744680853e-06,
"logits/chosen": 0.09555818885564804,
"logits/rejected": 0.28471559286117554,
"logps/chosen": -417.53369140625,
"logps/rejected": -316.70379638671875,
"loss": 0.3408,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": 0.0012717883801087737,
"rewards/margins": 0.0011722706258296967,
"rewards/rejected": 9.951768151950091e-05,
"step": 80
},
{
"epoch": 0.05,
"learning_rate": 2.393617021276596e-06,
"logits/chosen": 0.0800800547003746,
"logits/rejected": 0.24734680354595184,
"logps/chosen": -357.4005432128906,
"logps/rejected": -265.1945495605469,
"loss": 0.3418,
"rewards/accuracies": 0.71875,
"rewards/chosen": 0.0021034968085587025,
"rewards/margins": 0.001932466751895845,
"rewards/rejected": 0.00017103021673392504,
"step": 90
},
{
"epoch": 0.05,
"learning_rate": 2.6595744680851065e-06,
"logits/chosen": 0.07155252248048782,
"logits/rejected": 0.21769502758979797,
"logps/chosen": -328.1473693847656,
"logps/rejected": -286.50335693359375,
"loss": 0.3372,
"rewards/accuracies": 0.71875,
"rewards/chosen": 0.0025882914196699858,
"rewards/margins": 0.002097602002322674,
"rewards/rejected": 0.0004906897665932775,
"step": 100
},
{
"epoch": 0.06,
"learning_rate": 2.9255319148936174e-06,
"logits/chosen": 0.11561211198568344,
"logits/rejected": 0.235809326171875,
"logps/chosen": -371.8827209472656,
"logps/rejected": -298.6903991699219,
"loss": 0.3574,
"rewards/accuracies": 0.6625000238418579,
"rewards/chosen": 0.0038680892903357744,
"rewards/margins": 0.0038350392132997513,
"rewards/rejected": 3.3050309866666794e-05,
"step": 110
},
{
"epoch": 0.06,
"learning_rate": 3.191489361702128e-06,
"logits/chosen": 0.1021943911910057,
"logits/rejected": 0.2624170482158661,
"logps/chosen": -338.0439758300781,
"logps/rejected": -303.2659912109375,
"loss": 0.3179,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": 0.006541919894516468,
"rewards/margins": 0.007115885615348816,
"rewards/rejected": -0.000573965604417026,
"step": 120
},
{
"epoch": 0.07,
"learning_rate": 3.457446808510639e-06,
"logits/chosen": 0.1403232365846634,
"logits/rejected": 0.2289782017469406,
"logps/chosen": -322.43145751953125,
"logps/rejected": -290.4974670410156,
"loss": 0.3224,
"rewards/accuracies": 0.6937500238418579,
"rewards/chosen": 0.007107354700565338,
"rewards/margins": 0.009225147776305676,
"rewards/rejected": -0.0021177921444177628,
"step": 130
},
{
"epoch": 0.07,
"learning_rate": 3.723404255319149e-06,
"logits/chosen": 0.10777691751718521,
"logits/rejected": 0.3141722083091736,
"logps/chosen": -403.12591552734375,
"logps/rejected": -301.6325988769531,
"loss": 0.3034,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": 0.009438706561923027,
"rewards/margins": 0.01668398641049862,
"rewards/rejected": -0.007245280779898167,
"step": 140
},
{
"epoch": 0.08,
"learning_rate": 3.98936170212766e-06,
"logits/chosen": 0.144662007689476,
"logits/rejected": 0.29330867528915405,
"logps/chosen": -332.96417236328125,
"logps/rejected": -288.76763916015625,
"loss": 0.3229,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": 0.008229254744946957,
"rewards/margins": 0.020817989483475685,
"rewards/rejected": -0.012588733807206154,
"step": 150
},
{
"epoch": 0.09,
"learning_rate": 4.255319148936171e-06,
"logits/chosen": 0.13444170355796814,
"logits/rejected": 0.27335208654403687,
"logps/chosen": -375.7054443359375,
"logps/rejected": -343.5438232421875,
"loss": 0.3046,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.012767407111823559,
"rewards/margins": 0.03700178116559982,
"rewards/rejected": -0.049769192934036255,
"step": 160
},
{
"epoch": 0.09,
"learning_rate": 4.521276595744681e-06,
"logits/chosen": 0.15973123908042908,
"logits/rejected": 0.3123999238014221,
"logps/chosen": -407.84600830078125,
"logps/rejected": -383.7323913574219,
"loss": 0.2715,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.050918400287628174,
"rewards/margins": 0.05725928395986557,
"rewards/rejected": -0.10817768424749374,
"step": 170
},
{
"epoch": 0.1,
"learning_rate": 4.787234042553192e-06,
"logits/chosen": 0.17064544558525085,
"logits/rejected": 0.2910129129886627,
"logps/chosen": -472.98193359375,
"logps/rejected": -524.0833129882812,
"loss": 0.2907,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.16761399805545807,
"rewards/margins": 0.09154665470123291,
"rewards/rejected": -0.2591606676578522,
"step": 180
},
{
"epoch": 0.1,
"learning_rate": 4.999982660399688e-06,
"logits/chosen": 0.22318069636821747,
"logits/rejected": 0.33600229024887085,
"logps/chosen": -601.25390625,
"logps/rejected": -670.4763793945312,
"loss": 0.2971,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.23291774094104767,
"rewards/margins": 0.13643768429756165,
"rewards/rejected": -0.3693554401397705,
"step": 190
},
{
"epoch": 0.11,
"learning_rate": 4.99937579964398e-06,
"logits/chosen": 0.2132268249988556,
"logits/rejected": 0.3239246606826782,
"logps/chosen": -533.351806640625,
"logps/rejected": -600.036376953125,
"loss": 0.2656,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.18504436314105988,
"rewards/margins": 0.11939746141433716,
"rewards/rejected": -0.3044418692588806,
"step": 200
},
{
"epoch": 0.11,
"learning_rate": 4.9979021993870645e-06,
"logits/chosen": 0.19825667142868042,
"logits/rejected": 0.28912845253944397,
"logps/chosen": -540.4673461914062,
"logps/rejected": -592.4431762695312,
"loss": 0.2631,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.19064101576805115,
"rewards/margins": 0.1160808652639389,
"rewards/rejected": -0.30672189593315125,
"step": 210
},
{
"epoch": 0.12,
"learning_rate": 4.995562370647553e-06,
"logits/chosen": 0.20192313194274902,
"logits/rejected": 0.3630911707878113,
"logps/chosen": -584.4435424804688,
"logps/rejected": -627.2318115234375,
"loss": 0.2925,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.23791804909706116,
"rewards/margins": 0.10572359710931778,
"rewards/rejected": -0.3436416685581207,
"step": 220
},
{
"epoch": 0.12,
"learning_rate": 4.992357124836838e-06,
"logits/chosen": 0.2386132925748825,
"logits/rejected": 0.37924957275390625,
"logps/chosen": -515.5679931640625,
"logps/rejected": -620.7157592773438,
"loss": 0.2397,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.1845434010028839,
"rewards/margins": 0.14988870918750763,
"rewards/rejected": -0.33443212509155273,
"step": 230
},
{
"epoch": 0.13,
"learning_rate": 4.9882875734777044e-06,
"logits/chosen": 0.26459917426109314,
"logits/rejected": 0.4155924320220947,
"logps/chosen": -588.0615844726562,
"logps/rejected": -641.0905151367188,
"loss": 0.2725,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": -0.22051545977592468,
"rewards/margins": 0.13321301341056824,
"rewards/rejected": -0.3537284731864929,
"step": 240
},
{
"epoch": 0.13,
"learning_rate": 4.983355127818882e-06,
"logits/chosen": 0.272189199924469,
"logits/rejected": 0.39367181062698364,
"logps/chosen": -541.9008178710938,
"logps/rejected": -633.7440185546875,
"loss": 0.2557,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.212556391954422,
"rewards/margins": 0.14239820837974548,
"rewards/rejected": -0.3549546003341675,
"step": 250
},
{
"epoch": 0.14,
"learning_rate": 4.977561498345639e-06,
"logits/chosen": 0.2471778690814972,
"logits/rejected": 0.3926684260368347,
"logps/chosen": -590.729736328125,
"logps/rejected": -684.1608276367188,
"loss": 0.2675,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.25724658370018005,
"rewards/margins": 0.1302761733531952,
"rewards/rejected": -0.38752275705337524,
"step": 260
},
{
"epoch": 0.14,
"learning_rate": 4.970908694186624e-06,
"logits/chosen": 0.28343576192855835,
"logits/rejected": 0.40314382314682007,
"logps/chosen": -634.3505859375,
"logps/rejected": -726.0316772460938,
"loss": 0.2491,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.26509755849838257,
"rewards/margins": 0.15606968104839325,
"rewards/rejected": -0.4211672246456146,
"step": 270
},
{
"epoch": 0.15,
"learning_rate": 4.9633990224171305e-06,
"logits/chosen": 0.2620583474636078,
"logits/rejected": 0.44860973954200745,
"logps/chosen": -647.643798828125,
"logps/rejected": -790.6961669921875,
"loss": 0.239,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.3211327791213989,
"rewards/margins": 0.18398535251617432,
"rewards/rejected": -0.505118191242218,
"step": 280
},
{
"epoch": 0.15,
"learning_rate": 4.955035087259046e-06,
"logits/chosen": 0.2921401858329773,
"logits/rejected": 0.3832509219646454,
"logps/chosen": -621.01708984375,
"logps/rejected": -749.591552734375,
"loss": 0.2637,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.2830132246017456,
"rewards/margins": 0.17077818512916565,
"rewards/rejected": -0.45379146933555603,
"step": 290
},
{
"epoch": 0.16,
"learning_rate": 4.945819789177756e-06,
"logits/chosen": 0.25945979356765747,
"logits/rejected": 0.3613050580024719,
"logps/chosen": -625.9049072265625,
"logps/rejected": -672.7540893554688,
"loss": 0.2865,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.249430850148201,
"rewards/margins": 0.13059653341770172,
"rewards/rejected": -0.3800273835659027,
"step": 300
},
{
"epoch": 0.17,
"learning_rate": 4.935756323876306e-06,
"logits/chosen": 0.2395036220550537,
"logits/rejected": 0.403165340423584,
"logps/chosen": -589.790283203125,
"logps/rejected": -662.6749267578125,
"loss": 0.2506,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.23766255378723145,
"rewards/margins": 0.1399877816438675,
"rewards/rejected": -0.37765032052993774,
"step": 310
},
{
"epoch": 0.17,
"learning_rate": 4.924848181187199e-06,
"logits/chosen": 0.22068071365356445,
"logits/rejected": 0.3754776120185852,
"logps/chosen": -586.9141845703125,
"logps/rejected": -691.4939575195312,
"loss": 0.2237,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.2247442901134491,
"rewards/margins": 0.1634247750043869,
"rewards/rejected": -0.3881690800189972,
"step": 320
},
{
"epoch": 0.18,
"learning_rate": 4.913099143862173e-06,
"logits/chosen": 0.30473384261131287,
"logits/rejected": 0.3405100703239441,
"logps/chosen": -584.8270874023438,
"logps/rejected": -695.466796875,
"loss": 0.2435,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.24067524075508118,
"rewards/margins": 0.1444186419248581,
"rewards/rejected": -0.3850938677787781,
"step": 330
},
{
"epoch": 0.18,
"learning_rate": 4.900513286260416e-06,
"logits/chosen": 0.21341583132743835,
"logits/rejected": 0.35156476497650146,
"logps/chosen": -598.0617065429688,
"logps/rejected": -732.0949096679688,
"loss": 0.2782,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -0.25233131647109985,
"rewards/margins": 0.16543644666671753,
"rewards/rejected": -0.41776782274246216,
"step": 340
},
{
"epoch": 0.19,
"learning_rate": 4.887094972935645e-06,
"logits/chosen": 0.18200455605983734,
"logits/rejected": 0.29839521646499634,
"logps/chosen": -616.2686767578125,
"logps/rejected": -624.9378662109375,
"loss": 0.277,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.23792997002601624,
"rewards/margins": 0.10732170194387436,
"rewards/rejected": -0.3452516794204712,
"step": 350
},
{
"epoch": 0.19,
"learning_rate": 4.87284885712256e-06,
"logits/chosen": 0.1593484878540039,
"logits/rejected": 0.2911715507507324,
"logps/chosen": -610.3870239257812,
"logps/rejected": -689.4932861328125,
"loss": 0.2368,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.23754961788654327,
"rewards/margins": 0.149972602725029,
"rewards/rejected": -0.38752222061157227,
"step": 360
},
{
"epoch": 0.2,
"learning_rate": 4.857779879123181e-06,
"logits/chosen": 0.24016913771629333,
"logits/rejected": 0.27100852131843567,
"logps/chosen": -611.1376953125,
"logps/rejected": -736.8298950195312,
"loss": 0.2271,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.26364389061927795,
"rewards/margins": 0.16632166504859924,
"rewards/rejected": -0.4299655556678772,
"step": 370
},
{
"epoch": 0.2,
"learning_rate": 4.841893264593643e-06,
"logits/chosen": 0.20674113929271698,
"logits/rejected": 0.3422744572162628,
"logps/chosen": -657.8020629882812,
"logps/rejected": -761.3865966796875,
"loss": 0.2234,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.29639673233032227,
"rewards/margins": 0.1845850795507431,
"rewards/rejected": -0.48098182678222656,
"step": 380
},
{
"epoch": 0.21,
"learning_rate": 4.825194522732023e-06,
"logits/chosen": 0.2215661108493805,
"logits/rejected": 0.37389490008354187,
"logps/chosen": -650.2525634765625,
"logps/rejected": -728.6116943359375,
"loss": 0.2679,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.31919926404953003,
"rewards/margins": 0.15174493193626404,
"rewards/rejected": -0.47094422578811646,
"step": 390
},
{
"epoch": 0.21,
"learning_rate": 4.807689444367853e-06,
"logits/chosen": 0.2118716537952423,
"logits/rejected": 0.3723188638687134,
"logps/chosen": -704.6563720703125,
"logps/rejected": -808.8038330078125,
"loss": 0.2123,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -0.31687626242637634,
"rewards/margins": 0.1988348513841629,
"rewards/rejected": -0.5157111883163452,
"step": 400
},
{
"epoch": 0.22,
"learning_rate": 4.78938409995396e-06,
"logits/chosen": 0.22452159225940704,
"logits/rejected": 0.37957900762557983,
"logps/chosen": -664.1927490234375,
"logps/rejected": -804.2332153320312,
"loss": 0.2293,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.32922399044036865,
"rewards/margins": 0.18910188972949982,
"rewards/rejected": -0.5183258652687073,
"step": 410
},
{
"epoch": 0.22,
"learning_rate": 4.770284837461342e-06,
"logits/chosen": 0.2467646598815918,
"logits/rejected": 0.4089822769165039,
"logps/chosen": -735.5042114257812,
"logps/rejected": -830.4088134765625,
"loss": 0.2384,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.35978516936302185,
"rewards/margins": 0.164186030626297,
"rewards/rejected": -0.5239711999893188,
"step": 420
},
{
"epoch": 0.23,
"learning_rate": 4.7503982801778015e-06,
"logits/chosen": 0.2413889616727829,
"logits/rejected": 0.336980938911438,
"logps/chosen": -716.7408447265625,
"logps/rejected": -788.5014038085938,
"loss": 0.2541,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.32929089665412903,
"rewards/margins": 0.17145808041095734,
"rewards/rejected": -0.5007489919662476,
"step": 430
},
{
"epoch": 0.23,
"learning_rate": 4.729731324411104e-06,
"logits/chosen": 0.24484267830848694,
"logits/rejected": 0.3694288730621338,
"logps/chosen": -679.9339599609375,
"logps/rejected": -745.811279296875,
"loss": 0.2733,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.31394630670547485,
"rewards/margins": 0.154400035738945,
"rewards/rejected": -0.46834641695022583,
"step": 440
},
{
"epoch": 0.24,
"learning_rate": 4.7082911370974645e-06,
"logits/chosen": 0.21108904480934143,
"logits/rejected": 0.41315707564353943,
"logps/chosen": -622.7232666015625,
"logps/rejected": -786.5183715820312,
"loss": 0.2156,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.2859620153903961,
"rewards/margins": 0.19183282554149628,
"rewards/rejected": -0.477794885635376,
"step": 450
},
{
"epoch": 0.25,
"learning_rate": 4.68608515331618e-06,
"logits/chosen": 0.20401279628276825,
"logits/rejected": 0.33231475949287415,
"logps/chosen": -675.4547729492188,
"logps/rejected": -823.3021240234375,
"loss": 0.2434,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.3432995080947876,
"rewards/margins": 0.17998036742210388,
"rewards/rejected": -0.5232799053192139,
"step": 460
},
{
"epoch": 0.25,
"learning_rate": 4.663121073711269e-06,
"logits/chosen": 0.2568013668060303,
"logits/rejected": 0.4122098982334137,
"logps/chosen": -765.1724853515625,
"logps/rejected": -959.1907348632812,
"loss": 0.1761,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.40182775259017944,
"rewards/margins": 0.24152612686157227,
"rewards/rejected": -0.6433539390563965,
"step": 470
},
{
"epoch": 0.26,
"learning_rate": 4.63940686182103e-06,
"logits/chosen": 0.2044995129108429,
"logits/rejected": 0.43161505460739136,
"logps/chosen": -713.668701171875,
"logps/rejected": -906.75732421875,
"loss": 0.2101,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -0.3692263960838318,
"rewards/margins": 0.2429826259613037,
"rewards/rejected": -0.6122089624404907,
"step": 480
},
{
"epoch": 0.26,
"learning_rate": 4.614950741316425e-06,
"logits/chosen": 0.17913873493671417,
"logits/rejected": 0.2993609309196472,
"logps/chosen": -633.526611328125,
"logps/rejected": -757.5572509765625,
"loss": 0.2595,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.3094690442085266,
"rewards/margins": 0.1829945147037506,
"rewards/rejected": -0.4924635887145996,
"step": 490
},
{
"epoch": 0.27,
"learning_rate": 4.589761193149254e-06,
"logits/chosen": 0.15795853734016418,
"logits/rejected": 0.2824157774448395,
"logps/chosen": -617.8711547851562,
"logps/rejected": -708.7660522460938,
"loss": 0.243,
"rewards/accuracies": 0.6937500238418579,
"rewards/chosen": -0.2822186350822449,
"rewards/margins": 0.15801504254341125,
"rewards/rejected": -0.4402337074279785,
"step": 500
},
{
"epoch": 0.27,
"learning_rate": 4.563846952611112e-06,
"logits/chosen": 0.1760106384754181,
"logits/rejected": 0.37177279591560364,
"logps/chosen": -664.34765625,
"logps/rejected": -775.051513671875,
"loss": 0.2209,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.27559608221054077,
"rewards/margins": 0.18908929824829102,
"rewards/rejected": -0.4646853506565094,
"step": 510
},
{
"epoch": 0.28,
"learning_rate": 4.537217006304141e-06,
"logits/chosen": 0.14940790832042694,
"logits/rejected": 0.3582964241504669,
"logps/chosen": -674.0979614257812,
"logps/rejected": -782.9205932617188,
"loss": 0.2616,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.32881468534469604,
"rewards/margins": 0.18159161508083344,
"rewards/rejected": -0.5104063153266907,
"step": 520
},
{
"epoch": 0.28,
"learning_rate": 4.50988058902464e-06,
"logits/chosen": 0.16591671109199524,
"logits/rejected": 0.2972433567047119,
"logps/chosen": -665.7161865234375,
"logps/rejected": -733.9686889648438,
"loss": 0.227,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.3075326979160309,
"rewards/margins": 0.16415508091449738,
"rewards/rejected": -0.47168779373168945,
"step": 530
},
{
"epoch": 0.29,
"learning_rate": 4.481847180560593e-06,
"logits/chosen": 0.22560055553913116,
"logits/rejected": 0.29765042662620544,
"logps/chosen": -669.79443359375,
"logps/rejected": -775.9317626953125,
"loss": 0.2387,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3137567639350891,
"rewards/margins": 0.17480607330799103,
"rewards/rejected": -0.48856282234191895,
"step": 540
},
{
"epoch": 0.29,
"learning_rate": 4.453126502404253e-06,
"logits/chosen": 0.17669783532619476,
"logits/rejected": 0.35067084431648254,
"logps/chosen": -771.7174072265625,
"logps/rejected": -894.76416015625,
"loss": 0.2047,
"rewards/accuracies": 0.8187500238418579,
"rewards/chosen": -0.3670485317707062,
"rewards/margins": 0.2496991902589798,
"rewards/rejected": -0.6167477369308472,
"step": 550
},
{
"epoch": 0.3,
"learning_rate": 4.423728514380892e-06,
"logits/chosen": 0.18999402225017548,
"logits/rejected": 0.3391267955303192,
"logps/chosen": -684.16943359375,
"logps/rejected": -815.6094360351562,
"loss": 0.2587,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.3591926693916321,
"rewards/margins": 0.1908990740776062,
"rewards/rejected": -0.5500917434692383,
"step": 560
},
{
"epoch": 0.3,
"learning_rate": 4.393663411194918e-06,
"logits/chosen": 0.14545580744743347,
"logits/rejected": 0.4434194564819336,
"logps/chosen": -728.5840454101562,
"logps/rejected": -858.6510009765625,
"loss": 0.2285,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.343728631734848,
"rewards/margins": 0.2311449944972992,
"rewards/rejected": -0.5748736262321472,
"step": 570
},
{
"epoch": 0.31,
"learning_rate": 4.362941618894523e-06,
"logits/chosen": 0.23178081214427948,
"logits/rejected": 0.35969728231430054,
"logps/chosen": -674.2166137695312,
"logps/rejected": -835.3448486328125,
"loss": 0.1914,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -0.3182132840156555,
"rewards/margins": 0.23260562121868134,
"rewards/rejected": -0.5508188605308533,
"step": 580
},
{
"epoch": 0.31,
"learning_rate": 4.331573791256116e-06,
"logits/chosen": 0.23522010445594788,
"logits/rejected": 0.3881956934928894,
"logps/chosen": -717.13720703125,
"logps/rejected": -823.9237060546875,
"loss": 0.2441,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.3153195381164551,
"rewards/margins": 0.22255754470825195,
"rewards/rejected": -0.537877082824707,
"step": 590
},
{
"epoch": 0.32,
"learning_rate": 4.299570806089786e-06,
"logits/chosen": 0.18161171674728394,
"logits/rejected": 0.3819955289363861,
"logps/chosen": -676.7005004882812,
"logps/rejected": -885.69384765625,
"loss": 0.1855,
"rewards/accuracies": 0.8062499761581421,
"rewards/chosen": -0.3431881070137024,
"rewards/margins": 0.26160916686058044,
"rewards/rejected": -0.6047972440719604,
"step": 600
},
{
"epoch": 0.33,
"learning_rate": 4.266943761467057e-06,
"logits/chosen": 0.15835508704185486,
"logits/rejected": 0.3054640293121338,
"logps/chosen": -675.2325439453125,
"logps/rejected": -845.0226440429688,
"loss": 0.2303,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.3516426086425781,
"rewards/margins": 0.2019956409931183,
"rewards/rejected": -0.5536382794380188,
"step": 610
},
{
"epoch": 0.33,
"learning_rate": 4.233703971872287e-06,
"logits/chosen": 0.19981974363327026,
"logits/rejected": 0.31465253233909607,
"logps/chosen": -674.1165771484375,
"logps/rejected": -779.7503051757812,
"loss": 0.2182,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.30122750997543335,
"rewards/margins": 0.17384760081768036,
"rewards/rejected": -0.4750750660896301,
"step": 620
},
{
"epoch": 0.34,
"learning_rate": 4.1998629642789925e-06,
"logits/chosen": 0.13447639346122742,
"logits/rejected": 0.3254922926425934,
"logps/chosen": -601.1633911132812,
"logps/rejected": -713.8211669921875,
"loss": 0.2735,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.26304560899734497,
"rewards/margins": 0.1775732934474945,
"rewards/rejected": -0.4406188428401947,
"step": 630
},
{
"epoch": 0.34,
"learning_rate": 4.165432474152505e-06,
"logits/chosen": 0.1678907424211502,
"logits/rejected": 0.29623210430145264,
"logps/chosen": -637.0558471679688,
"logps/rejected": -760.5277099609375,
"loss": 0.2456,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.29105207324028015,
"rewards/margins": 0.1802116334438324,
"rewards/rejected": -0.4712636470794678,
"step": 640
},
{
"epoch": 0.35,
"learning_rate": 4.130424441380308e-06,
"logits/chosen": 0.18655350804328918,
"logits/rejected": 0.3515748381614685,
"logps/chosen": -621.6171875,
"logps/rejected": -770.1412963867188,
"loss": 0.2009,
"rewards/accuracies": 0.793749988079071,
"rewards/chosen": -0.2730523347854614,
"rewards/margins": 0.20337846875190735,
"rewards/rejected": -0.476430743932724,
"step": 650
},
{
"epoch": 0.35,
"learning_rate": 4.09485100613151e-06,
"logits/chosen": 0.2070898562669754,
"logits/rejected": 0.3552184998989105,
"logps/chosen": -644.9053955078125,
"logps/rejected": -800.6888427734375,
"loss": 0.2057,
"rewards/accuracies": 0.793749988079071,
"rewards/chosen": -0.2831494212150574,
"rewards/margins": 0.21793799102306366,
"rewards/rejected": -0.501087486743927,
"step": 660
},
{
"epoch": 0.36,
"learning_rate": 4.058724504646834e-06,
"logits/chosen": 0.16107818484306335,
"logits/rejected": 0.32218271493911743,
"logps/chosen": -676.8900146484375,
"logps/rejected": -834.2060546875,
"loss": 0.1891,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.3277270495891571,
"rewards/margins": 0.22579827904701233,
"rewards/rejected": -0.5535253286361694,
"step": 670
},
{
"epoch": 0.36,
"learning_rate": 4.022057464960632e-06,
"logits/chosen": 0.20739403367042542,
"logits/rejected": 0.26471713185310364,
"logps/chosen": -687.0540161132812,
"logps/rejected": -849.7962646484375,
"loss": 0.2414,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.37053388357162476,
"rewards/margins": 0.19931861758232117,
"rewards/rejected": -0.5698524713516235,
"step": 680
},
{
"epoch": 0.37,
"learning_rate": 3.984862602556383e-06,
"logits/chosen": 0.20264072716236115,
"logits/rejected": 0.2857401669025421,
"logps/chosen": -638.552734375,
"logps/rejected": -793.6614990234375,
"loss": 0.2271,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.32019636034965515,
"rewards/margins": 0.1985076367855072,
"rewards/rejected": -0.5187040567398071,
"step": 690
},
{
"epoch": 0.37,
"learning_rate": 3.947152815957187e-06,
"logits/chosen": 0.17015784978866577,
"logits/rejected": 0.25645551085472107,
"logps/chosen": -696.7833862304688,
"logps/rejected": -824.9973754882812,
"loss": 0.2486,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.36269837617874146,
"rewards/margins": 0.18247218430042267,
"rewards/rejected": -0.5451706051826477,
"step": 700
},
{
"epoch": 0.38,
"learning_rate": 3.908941182252785e-06,
"logits/chosen": 0.2156047374010086,
"logits/rejected": 0.20423674583435059,
"logps/chosen": -619.2033081054688,
"logps/rejected": -690.5067138671875,
"loss": 0.3245,
"rewards/accuracies": 0.643750011920929,
"rewards/chosen": -0.32582998275756836,
"rewards/margins": 0.12705637514591217,
"rewards/rejected": -0.45288634300231934,
"step": 710
},
{
"epoch": 0.38,
"learning_rate": 3.8702409525646535e-06,
"logits/chosen": 0.14758387207984924,
"logits/rejected": 0.3225356340408325,
"logps/chosen": -671.9241333007812,
"logps/rejected": -804.1224365234375,
"loss": 0.2129,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.30027109384536743,
"rewards/margins": 0.19518795609474182,
"rewards/rejected": -0.49545907974243164,
"step": 720
},
{
"epoch": 0.39,
"learning_rate": 3.8310655474507495e-06,
"logits/chosen": 0.1207294911146164,
"logits/rejected": 0.3056190311908722,
"logps/chosen": -655.4552001953125,
"logps/rejected": -807.0453491210938,
"loss": 0.1871,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.29568520188331604,
"rewards/margins": 0.21632209420204163,
"rewards/rejected": -0.5120072960853577,
"step": 730
},
{
"epoch": 0.39,
"learning_rate": 3.7914285522515002e-06,
"logits/chosen": 0.17697741091251373,
"logits/rejected": 0.2939094603061676,
"logps/chosen": -667.5819091796875,
"logps/rejected": -791.62353515625,
"loss": 0.2156,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.32927432656288147,
"rewards/margins": 0.17469336092472076,
"rewards/rejected": -0.5039676427841187,
"step": 740
},
{
"epoch": 0.4,
"learning_rate": 3.751343712378639e-06,
"logits/chosen": 0.15482628345489502,
"logits/rejected": 0.3409688174724579,
"logps/chosen": -716.8922119140625,
"logps/rejected": -896.0595703125,
"loss": 0.1907,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.35755792260169983,
"rewards/margins": 0.24695225059986115,
"rewards/rejected": -0.6045101881027222,
"step": 750
},
{
"epoch": 0.41,
"learning_rate": 3.710824928548546e-06,
"logits/chosen": 0.17361067235469818,
"logits/rejected": 0.3247506320476532,
"logps/chosen": -759.477294921875,
"logps/rejected": -859.95703125,
"loss": 0.2266,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.3791542947292328,
"rewards/margins": 0.19504983723163605,
"rewards/rejected": -0.5742040872573853,
"step": 760
},
{
"epoch": 0.41,
"learning_rate": 3.6698862519617225e-06,
"logits/chosen": 0.1680970937013626,
"logits/rejected": 0.3031509816646576,
"logps/chosen": -673.4447021484375,
"logps/rejected": -791.8770751953125,
"loss": 0.2111,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.33166688680648804,
"rewards/margins": 0.19004380702972412,
"rewards/rejected": -0.5217106938362122,
"step": 770
},
{
"epoch": 0.42,
"learning_rate": 3.6285418794300793e-06,
"logits/chosen": 0.1661594659090042,
"logits/rejected": 0.29215800762176514,
"logps/chosen": -656.0603637695312,
"logps/rejected": -792.6107788085938,
"loss": 0.2202,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.33995765447616577,
"rewards/margins": 0.1824071854352951,
"rewards/rejected": -0.5223647952079773,
"step": 780
},
{
"epoch": 0.42,
"learning_rate": 3.5868061484537365e-06,
"logits/chosen": 0.1200464516878128,
"logits/rejected": 0.2862261235713959,
"logps/chosen": -711.8590087890625,
"logps/rejected": -884.7125244140625,
"loss": 0.1871,
"rewards/accuracies": 0.800000011920929,
"rewards/chosen": -0.3494313359260559,
"rewards/margins": 0.2312242090702057,
"rewards/rejected": -0.5806554555892944,
"step": 790
},
{
"epoch": 0.43,
"learning_rate": 3.5446935322490285e-06,
"logits/chosen": 0.11773556470870972,
"logits/rejected": 0.29994910955429077,
"logps/chosen": -680.5064086914062,
"logps/rejected": -856.234375,
"loss": 0.1885,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.35676535964012146,
"rewards/margins": 0.22748513519763947,
"rewards/rejected": -0.5842504501342773,
"step": 800
},
{
"epoch": 0.43,
"learning_rate": 3.502218634729447e-06,
"logits/chosen": 0.09288445115089417,
"logits/rejected": 0.33626627922058105,
"logps/chosen": -774.9942626953125,
"logps/rejected": -963.0445556640625,
"loss": 0.2172,
"rewards/accuracies": 0.793749988079071,
"rewards/chosen": -0.40824300050735474,
"rewards/margins": 0.24136242270469666,
"rewards/rejected": -0.649605393409729,
"step": 810
},
{
"epoch": 0.44,
"learning_rate": 3.459396185441265e-06,
"logits/chosen": 0.1620679795742035,
"logits/rejected": 0.31993260979652405,
"logps/chosen": -728.768798828125,
"logps/rejected": -806.7156982421875,
"loss": 0.2566,
"rewards/accuracies": 0.6812499761581421,
"rewards/chosen": -0.3659815490245819,
"rewards/margins": 0.17412114143371582,
"rewards/rejected": -0.5401026606559753,
"step": 820
},
{
"epoch": 0.44,
"learning_rate": 3.4162410344555834e-06,
"logits/chosen": 0.14278681576251984,
"logits/rejected": 0.32590746879577637,
"logps/chosen": -703.8787231445312,
"logps/rejected": -857.1585083007812,
"loss": 0.2124,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.3663283884525299,
"rewards/margins": 0.21885888278484344,
"rewards/rejected": -0.5851873159408569,
"step": 830
},
{
"epoch": 0.45,
"learning_rate": 3.3727681472185937e-06,
"logits/chosen": 0.19465771317481995,
"logits/rejected": 0.28358665108680725,
"logps/chosen": -671.1470336914062,
"logps/rejected": -810.312744140625,
"loss": 0.2081,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.31882521510124207,
"rewards/margins": 0.20828664302825928,
"rewards/rejected": -0.5271118879318237,
"step": 840
},
{
"epoch": 0.45,
"learning_rate": 3.3289925993618217e-06,
"logits/chosen": 0.12119020521640778,
"logits/rejected": 0.3242005705833435,
"logps/chosen": -680.9307861328125,
"logps/rejected": -822.7205810546875,
"loss": 0.2258,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.328025221824646,
"rewards/margins": 0.21506837010383606,
"rewards/rejected": -0.5430935621261597,
"step": 850
},
{
"epoch": 0.46,
"learning_rate": 3.2849295714741643e-06,
"logits/chosen": 0.11271035671234131,
"logits/rejected": 0.2887061536312103,
"logps/chosen": -675.8961181640625,
"logps/rejected": -808.447265625,
"loss": 0.2158,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.29701152443885803,
"rewards/margins": 0.21012258529663086,
"rewards/rejected": -0.5071341395378113,
"step": 860
},
{
"epoch": 0.46,
"learning_rate": 3.2405943438375287e-06,
"logits/chosen": 0.15118266642093658,
"logits/rejected": 0.29830652475357056,
"logps/chosen": -648.36767578125,
"logps/rejected": -793.2464599609375,
"loss": 0.2488,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.30504560470581055,
"rewards/margins": 0.21465754508972168,
"rewards/rejected": -0.5197031497955322,
"step": 870
},
{
"epoch": 0.47,
"learning_rate": 3.1960022911279036e-06,
"logits/chosen": 0.10310302674770355,
"logits/rejected": 0.26414334774017334,
"logps/chosen": -628.9229736328125,
"logps/rejected": -789.9716186523438,
"loss": 0.2024,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.2890340983867645,
"rewards/margins": 0.2192925661802292,
"rewards/rejected": -0.5083267092704773,
"step": 880
},
{
"epoch": 0.47,
"learning_rate": 3.1511688770836844e-06,
"logits/chosen": 0.12954168021678925,
"logits/rejected": 0.2752276659011841,
"logps/chosen": -620.2590942382812,
"logps/rejected": -767.0093994140625,
"loss": 0.2052,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.275544673204422,
"rewards/margins": 0.21567711234092712,
"rewards/rejected": -0.4912218153476715,
"step": 890
},
{
"epoch": 0.48,
"learning_rate": 3.1061096491431307e-06,
"logits/chosen": 0.12713904678821564,
"logits/rejected": 0.2954176068305969,
"logps/chosen": -646.6031494140625,
"logps/rejected": -788.7861938476562,
"loss": 0.2272,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.2984674274921417,
"rewards/margins": 0.20731091499328613,
"rewards/rejected": -0.5057783722877502,
"step": 900
},
{
"epoch": 0.49,
"learning_rate": 3.0608402330527796e-06,
"logits/chosen": 0.11980082839727402,
"logits/rejected": 0.20131537318229675,
"logps/chosen": -652.871337890625,
"logps/rejected": -815.8560791015625,
"loss": 0.2053,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.2841240167617798,
"rewards/margins": 0.2087261974811554,
"rewards/rejected": -0.49285024404525757,
"step": 910
},
{
"epoch": 0.49,
"learning_rate": 3.0153763274487176e-06,
"logits/chosen": 0.11366409063339233,
"logits/rejected": 0.20989951491355896,
"logps/chosen": -632.2529296875,
"logps/rejected": -742.0712280273438,
"loss": 0.2521,
"rewards/accuracies": 0.637499988079071,
"rewards/chosen": -0.30224618315696716,
"rewards/margins": 0.17021192610263824,
"rewards/rejected": -0.4724581241607666,
"step": 920
},
{
"epoch": 0.5,
"learning_rate": 2.9697336984125683e-06,
"logits/chosen": 0.12950178980827332,
"logits/rejected": 0.205675408244133,
"logps/chosen": -666.3436279296875,
"logps/rejected": -803.6804809570312,
"loss": 0.2306,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.3005353808403015,
"rewards/margins": 0.2294267863035202,
"rewards/rejected": -0.5299621820449829,
"step": 930
},
{
"epoch": 0.5,
"learning_rate": 2.923928174004094e-06,
"logits/chosen": 0.09100206196308136,
"logits/rejected": 0.23861780762672424,
"logps/chosen": -660.6749267578125,
"logps/rejected": -811.4717407226562,
"loss": 0.2064,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.2903476655483246,
"rewards/margins": 0.21249572932720184,
"rewards/rejected": -0.50284343957901,
"step": 940
},
{
"epoch": 0.51,
"learning_rate": 2.8779756387723036e-06,
"logits/chosen": 0.11082760244607925,
"logits/rejected": 0.23662987351417542,
"logps/chosen": -665.4554443359375,
"logps/rejected": -868.9638671875,
"loss": 0.201,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.34308990836143494,
"rewards/margins": 0.24668097496032715,
"rewards/rejected": -0.5897709131240845,
"step": 950
},
{
"epoch": 0.51,
"learning_rate": 2.831892028246968e-06,
"logits/chosen": 0.15180227160453796,
"logits/rejected": 0.21313562989234924,
"logps/chosen": -717.8250732421875,
"logps/rejected": -806.0482177734375,
"loss": 0.2638,
"rewards/accuracies": 0.675000011920929,
"rewards/chosen": -0.3388670086860657,
"rewards/margins": 0.1811041533946991,
"rewards/rejected": -0.5199711322784424,
"step": 960
},
{
"epoch": 0.52,
"learning_rate": 2.7856933234124617e-06,
"logits/chosen": 0.07411827147006989,
"logits/rejected": 0.16980966925621033,
"logps/chosen": -684.9701538085938,
"logps/rejected": -862.4952392578125,
"loss": 0.2099,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.362778902053833,
"rewards/margins": 0.2055092304944992,
"rewards/rejected": -0.5682881474494934,
"step": 970
},
{
"epoch": 0.52,
"learning_rate": 2.7393955451658387e-06,
"logits/chosen": 0.10116670280694962,
"logits/rejected": 0.21950002014636993,
"logps/chosen": -674.0352783203125,
"logps/rejected": -800.4089965820312,
"loss": 0.2249,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.3368411660194397,
"rewards/margins": 0.1978548765182495,
"rewards/rejected": -0.5346960425376892,
"step": 980
},
{
"epoch": 0.53,
"learning_rate": 2.6930147487610667e-06,
"logits/chosen": 0.0984363779425621,
"logits/rejected": 0.2318364828824997,
"logps/chosen": -728.8323974609375,
"logps/rejected": -877.23681640625,
"loss": 0.2061,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.36335819959640503,
"rewards/margins": 0.22163307666778564,
"rewards/rejected": -0.5849912762641907,
"step": 990
},
{
"epoch": 0.53,
"learning_rate": 2.6465670182413487e-06,
"logits/chosen": 0.12672573328018188,
"logits/rejected": 0.19936290383338928,
"logps/chosen": -735.6030883789062,
"logps/rejected": -855.2264404296875,
"loss": 0.2383,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.3681355118751526,
"rewards/margins": 0.18624527752399445,
"rewards/rejected": -0.5543807744979858,
"step": 1000
},
{
"epoch": 0.54,
"learning_rate": 2.6000684608614594e-06,
"logits/chosen": 0.09366180747747421,
"logits/rejected": 0.24154536426067352,
"logps/chosen": -741.3905029296875,
"logps/rejected": -863.9640502929688,
"loss": 0.2183,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.3723335266113281,
"rewards/margins": 0.20260421931743622,
"rewards/rejected": -0.5749377608299255,
"step": 1010
},
{
"epoch": 0.54,
"learning_rate": 2.5535352015020338e-06,
"logits/chosen": 0.12707850337028503,
"logits/rejected": 0.262873113155365,
"logps/chosen": -742.9122314453125,
"logps/rejected": -812.1138916015625,
"loss": 0.2556,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.34759730100631714,
"rewards/margins": 0.17379847168922424,
"rewards/rejected": -0.5213958024978638,
"step": 1020
},
{
"epoch": 0.55,
"learning_rate": 2.506983377077741e-06,
"logits/chosen": 0.06376073509454727,
"logits/rejected": 0.2528199851512909,
"logps/chosen": -648.4464721679688,
"logps/rejected": -814.0970458984375,
"loss": 0.238,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.31961458921432495,
"rewards/margins": 0.20580220222473145,
"rewards/rejected": -0.5254167318344116,
"step": 1030
},
{
"epoch": 0.55,
"learning_rate": 2.460429130941289e-06,
"logits/chosen": 0.08957502990961075,
"logits/rejected": 0.18970297276973724,
"logps/chosen": -675.5179443359375,
"logps/rejected": -772.6138916015625,
"loss": 0.2382,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.31163010001182556,
"rewards/margins": 0.17897368967533112,
"rewards/rejected": -0.49060383439064026,
"step": 1040
},
{
"epoch": 0.56,
"learning_rate": 2.413888607285192e-06,
"logits/chosen": 0.12193132936954498,
"logits/rejected": 0.2238079011440277,
"logps/chosen": -623.6652221679688,
"logps/rejected": -791.8870849609375,
"loss": 0.1977,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.3007582128047943,
"rewards/margins": 0.20639415085315704,
"rewards/rejected": -0.5071523785591125,
"step": 1050
},
{
"epoch": 0.57,
"learning_rate": 2.367377945543249e-06,
"logits/chosen": 0.09605869650840759,
"logits/rejected": 0.18777312338352203,
"logps/chosen": -642.5596313476562,
"logps/rejected": -781.3046875,
"loss": 0.1997,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.30638837814331055,
"rewards/margins": 0.18436425924301147,
"rewards/rejected": -0.490752637386322,
"step": 1060
},
{
"epoch": 0.57,
"learning_rate": 2.320913274793676e-06,
"logits/chosen": 0.09697475284337997,
"logits/rejected": 0.21658281981945038,
"logps/chosen": -663.3118286132812,
"logps/rejected": -803.4898681640625,
"loss": 0.1913,
"rewards/accuracies": 0.793749988079071,
"rewards/chosen": -0.2993699908256531,
"rewards/margins": 0.22615595161914825,
"rewards/rejected": -0.5255259275436401,
"step": 1070
},
{
"epoch": 0.58,
"learning_rate": 2.27451070816582e-06,
"logits/chosen": 0.1070629209280014,
"logits/rejected": 0.25807613134384155,
"logps/chosen": -694.64111328125,
"logps/rejected": -816.5810546875,
"loss": 0.2196,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.32364705204963684,
"rewards/margins": 0.22359292209148407,
"rewards/rejected": -0.5472399592399597,
"step": 1080
},
{
"epoch": 0.58,
"learning_rate": 2.228186337252414e-06,
"logits/chosen": 0.11755505949258804,
"logits/rejected": 0.33424052596092224,
"logps/chosen": -726.0146484375,
"logps/rejected": -885.8994140625,
"loss": 0.2243,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.37689560651779175,
"rewards/margins": 0.21640503406524658,
"rewards/rejected": -0.5933005809783936,
"step": 1090
},
{
"epoch": 0.59,
"learning_rate": 2.1819562265292946e-06,
"logits/chosen": 0.08579425513744354,
"logits/rejected": 0.2378162145614624,
"logps/chosen": -713.3672485351562,
"logps/rejected": -939.802734375,
"loss": 0.1748,
"rewards/accuracies": 0.8062499761581421,
"rewards/chosen": -0.37114056944847107,
"rewards/margins": 0.28332680463790894,
"rewards/rejected": -0.6544674038887024,
"step": 1100
},
{
"epoch": 0.59,
"learning_rate": 2.1358364077845236e-06,
"logits/chosen": 0.06883491575717926,
"logits/rejected": 0.21484391391277313,
"logps/chosen": -713.2088623046875,
"logps/rejected": -950.1226806640625,
"loss": 0.2204,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.38237085938453674,
"rewards/margins": 0.26614242792129517,
"rewards/rejected": -0.6485133171081543,
"step": 1110
},
{
"epoch": 0.6,
"learning_rate": 2.089842874558849e-06,
"logits/chosen": 0.12803800404071808,
"logits/rejected": 0.1985175758600235,
"logps/chosen": -675.037841796875,
"logps/rejected": -824.63427734375,
"loss": 0.2147,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.3487189710140228,
"rewards/margins": 0.19292229413986206,
"rewards/rejected": -0.5416412949562073,
"step": 1120
},
{
"epoch": 0.6,
"learning_rate": 2.0439915765994242e-06,
"logits/chosen": 0.08092626929283142,
"logits/rejected": 0.22723694145679474,
"logps/chosen": -757.0364990234375,
"logps/rejected": -870.3096923828125,
"loss": 0.1899,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.3505310118198395,
"rewards/margins": 0.22947946190834045,
"rewards/rejected": -0.5800104141235352,
"step": 1130
},
{
"epoch": 0.61,
"learning_rate": 1.9982984143287186e-06,
"logits/chosen": 0.10787869989871979,
"logits/rejected": 0.1968112587928772,
"logps/chosen": -630.917236328125,
"logps/rejected": -771.1486206054688,
"loss": 0.2133,
"rewards/accuracies": 0.6812499761581421,
"rewards/chosen": -0.3110920190811157,
"rewards/margins": 0.18016596138477325,
"rewards/rejected": -0.4912579655647278,
"step": 1140
},
{
"epoch": 0.61,
"learning_rate": 1.95277923333053e-06,
"logits/chosen": 0.1406387984752655,
"logits/rejected": 0.22435446083545685,
"logps/chosen": -709.8244018554688,
"logps/rejected": -844.6785278320312,
"loss": 0.2164,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.34992507100105286,
"rewards/margins": 0.2139689177274704,
"rewards/rejected": -0.5638940334320068,
"step": 1150
},
{
"epoch": 0.62,
"learning_rate": 1.9074498188550156e-06,
"logits/chosen": 0.13347899913787842,
"logits/rejected": 0.2880741059780121,
"logps/chosen": -704.9302978515625,
"logps/rejected": -873.6126708984375,
"loss": 0.2,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3522919714450836,
"rewards/margins": 0.2265801876783371,
"rewards/rejected": -0.5788721442222595,
"step": 1160
},
{
"epoch": 0.62,
"learning_rate": 1.862325890344643e-06,
"logits/chosen": 0.08221012353897095,
"logits/rejected": 0.25464189052581787,
"logps/chosen": -757.3726806640625,
"logps/rejected": -926.8704223632812,
"loss": 0.2066,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.37381109595298767,
"rewards/margins": 0.23461690545082092,
"rewards/rejected": -0.6084280014038086,
"step": 1170
},
{
"epoch": 0.63,
"learning_rate": 1.817423095982972e-06,
"logits/chosen": 0.13979586958885193,
"logits/rejected": 0.24614198505878448,
"logps/chosen": -731.94140625,
"logps/rejected": -828.9541015625,
"loss": 0.2687,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.3592011630535126,
"rewards/margins": 0.1861731857061386,
"rewards/rejected": -0.5453742742538452,
"step": 1180
},
{
"epoch": 0.63,
"learning_rate": 1.7727570072681293e-06,
"logits/chosen": 0.09579949080944061,
"logits/rejected": 0.22970926761627197,
"logps/chosen": -669.8802490234375,
"logps/rejected": -843.9857177734375,
"loss": 0.2002,
"rewards/accuracies": 0.8062499761581421,
"rewards/chosen": -0.3365887999534607,
"rewards/margins": 0.22111590206623077,
"rewards/rejected": -0.5577046275138855,
"step": 1190
},
{
"epoch": 0.64,
"learning_rate": 1.7283431136128961e-06,
"logits/chosen": 0.1350405365228653,
"logits/rejected": 0.2594318985939026,
"logps/chosen": -712.8129272460938,
"logps/rejected": -832.94677734375,
"loss": 0.2323,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.34674012660980225,
"rewards/margins": 0.1996394544839859,
"rewards/rejected": -0.5463796257972717,
"step": 1200
},
{
"epoch": 0.65,
"learning_rate": 1.6841968169732478e-06,
"logits/chosen": 0.16419488191604614,
"logits/rejected": 0.25972408056259155,
"logps/chosen": -692.2588500976562,
"logps/rejected": -844.2239990234375,
"loss": 0.2063,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.3172731101512909,
"rewards/margins": 0.21843883395195007,
"rewards/rejected": -0.5357118844985962,
"step": 1210
},
{
"epoch": 0.65,
"learning_rate": 1.6403334265072284e-06,
"logits/chosen": 0.12711533904075623,
"logits/rejected": 0.24427728354930878,
"logps/chosen": -723.783935546875,
"logps/rejected": -863.6396484375,
"loss": 0.2058,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.33928173780441284,
"rewards/margins": 0.23463325202465057,
"rewards/rejected": -0.5739150047302246,
"step": 1220
},
{
"epoch": 0.66,
"learning_rate": 1.5967681532660066e-06,
"logits/chosen": 0.10758714377880096,
"logits/rejected": 0.2805503308773041,
"logps/chosen": -701.0147705078125,
"logps/rejected": -815.3477783203125,
"loss": 0.2015,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.3337828814983368,
"rewards/margins": 0.2107526808977127,
"rewards/rejected": -0.5445355176925659,
"step": 1230
},
{
"epoch": 0.66,
"learning_rate": 1.5535161049189463e-06,
"logits/chosen": 0.12466589361429214,
"logits/rejected": 0.26773038506507874,
"logps/chosen": -640.2819213867188,
"logps/rejected": -786.9939575195312,
"loss": 0.2169,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.33266982436180115,
"rewards/margins": 0.1839940994977951,
"rewards/rejected": -0.5166639089584351,
"step": 1240
},
{
"epoch": 0.67,
"learning_rate": 1.5105922805145356e-06,
"logits/chosen": 0.12364304065704346,
"logits/rejected": 0.25227856636047363,
"logps/chosen": -671.4842529296875,
"logps/rejected": -803.79345703125,
"loss": 0.2247,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.3221370577812195,
"rewards/margins": 0.21037554740905762,
"rewards/rejected": -0.5325126051902771,
"step": 1250
},
{
"epoch": 0.67,
"learning_rate": 1.4680115652789823e-06,
"logits/chosen": 0.13939639925956726,
"logits/rejected": 0.23471331596374512,
"logps/chosen": -714.9220581054688,
"logps/rejected": -864.4716796875,
"loss": 0.2297,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.3482537567615509,
"rewards/margins": 0.22038355469703674,
"rewards/rejected": -0.5686372518539429,
"step": 1260
},
{
"epoch": 0.68,
"learning_rate": 1.4257887254542767e-06,
"logits/chosen": 0.16542884707450867,
"logits/rejected": 0.24698495864868164,
"logps/chosen": -668.9215698242188,
"logps/rejected": -835.5013427734375,
"loss": 0.191,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.3274608254432678,
"rewards/margins": 0.2046995609998703,
"rewards/rejected": -0.5321604013442993,
"step": 1270
},
{
"epoch": 0.68,
"learning_rate": 1.3839384031775227e-06,
"logits/chosen": 0.15713492035865784,
"logits/rejected": 0.2516704201698303,
"logps/chosen": -739.1085815429688,
"logps/rejected": -858.5968627929688,
"loss": 0.2318,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.36160725355148315,
"rewards/margins": 0.20901015400886536,
"rewards/rejected": -0.5706173777580261,
"step": 1280
},
{
"epoch": 0.69,
"learning_rate": 1.342475111403298e-06,
"logits/chosen": 0.10080404579639435,
"logits/rejected": 0.24557176232337952,
"logps/chosen": -697.7256469726562,
"logps/rejected": -914.9107666015625,
"loss": 0.1877,
"rewards/accuracies": 0.8062499761581421,
"rewards/chosen": -0.3470718264579773,
"rewards/margins": 0.25500741600990295,
"rewards/rejected": -0.6020792722702026,
"step": 1290
},
{
"epoch": 0.69,
"learning_rate": 1.3014132288708209e-06,
"logits/chosen": 0.12770727276802063,
"logits/rejected": 0.19649390876293182,
"logps/chosen": -698.4237060546875,
"logps/rejected": -831.1112060546875,
"loss": 0.2174,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -0.36567771434783936,
"rewards/margins": 0.18466141819953918,
"rewards/rejected": -0.5503391623497009,
"step": 1300
},
{
"epoch": 0.7,
"learning_rate": 1.2607669951176549e-06,
"logits/chosen": 0.1088666170835495,
"logits/rejected": 0.26579782366752625,
"logps/chosen": -697.934326171875,
"logps/rejected": -863.8990478515625,
"loss": 0.184,
"rewards/accuracies": 0.793749988079071,
"rewards/chosen": -0.32992303371429443,
"rewards/margins": 0.2491859644651413,
"rewards/rejected": -0.5791090130805969,
"step": 1310
},
{
"epoch": 0.7,
"learning_rate": 1.2205505055416891e-06,
"logits/chosen": 0.1179809719324112,
"logits/rejected": 0.21980592608451843,
"logps/chosen": -690.4783935546875,
"logps/rejected": -883.32861328125,
"loss": 0.1656,
"rewards/accuracies": 0.793749988079071,
"rewards/chosen": -0.32375839352607727,
"rewards/margins": 0.27088266611099243,
"rewards/rejected": -0.5946410298347473,
"step": 1320
},
{
"epoch": 0.71,
"learning_rate": 1.1807777065131002e-06,
"logits/chosen": 0.10531783103942871,
"logits/rejected": 0.2298382818698883,
"logps/chosen": -704.1119384765625,
"logps/rejected": -891.2111206054688,
"loss": 0.2144,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.3464652895927429,
"rewards/margins": 0.23750798404216766,
"rewards/rejected": -0.5839732885360718,
"step": 1330
},
{
"epoch": 0.71,
"learning_rate": 1.1414623905380012e-06,
"logits/chosen": 0.06631898880004883,
"logits/rejected": 0.25904709100723267,
"logps/chosen": -726.3780517578125,
"logps/rejected": -875.1380615234375,
"loss": 0.2532,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.3661043047904968,
"rewards/margins": 0.18917004764080048,
"rewards/rejected": -0.5552743673324585,
"step": 1340
},
{
"epoch": 0.72,
"learning_rate": 1.1026181914754388e-06,
"logits/chosen": 0.1238614097237587,
"logits/rejected": 0.25796738266944885,
"logps/chosen": -661.4393310546875,
"logps/rejected": -795.7490844726562,
"loss": 0.1855,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.29050904512405396,
"rewards/margins": 0.2389696091413498,
"rewards/rejected": -0.5294786691665649,
"step": 1350
},
{
"epoch": 0.73,
"learning_rate": 1.0642585798094136e-06,
"logits/chosen": 0.09917294979095459,
"logits/rejected": 0.21815705299377441,
"logps/chosen": -659.0029296875,
"logps/rejected": -792.0519409179688,
"loss": 0.224,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.3168398439884186,
"rewards/margins": 0.19915297627449036,
"rewards/rejected": -0.5159928202629089,
"step": 1360
},
{
"epoch": 0.73,
"learning_rate": 1.0263968579775522e-06,
"logits/chosen": 0.09248078614473343,
"logits/rejected": 0.21158233284950256,
"logps/chosen": -678.2177734375,
"logps/rejected": -867.001953125,
"loss": 0.1935,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.31002795696258545,
"rewards/margins": 0.25347810983657837,
"rewards/rejected": -0.5635061264038086,
"step": 1370
},
{
"epoch": 0.74,
"learning_rate": 9.89046155758058e-07,
"logits/chosen": 0.1292383223772049,
"logits/rejected": 0.24721387028694153,
"logps/chosen": -677.2322998046875,
"logps/rejected": -853.12939453125,
"loss": 0.2384,
"rewards/accuracies": 0.7124999761581421,
"rewards/chosen": -0.33947324752807617,
"rewards/margins": 0.2104976624250412,
"rewards/rejected": -0.5499709248542786,
"step": 1380
},
{
"epoch": 0.74,
"learning_rate": 9.52219425716534e-07,
"logits/chosen": 0.13778840005397797,
"logits/rejected": 0.22406592965126038,
"logps/chosen": -707.752197265625,
"logps/rejected": -878.6907958984375,
"loss": 0.2185,
"rewards/accuracies": 0.793749988079071,
"rewards/chosen": -0.3595397472381592,
"rewards/margins": 0.2256437987089157,
"rewards/rejected": -0.5851835012435913,
"step": 1390
},
{
"epoch": 0.75,
"learning_rate": 9.15929438714262e-07,
"logits/chosen": 0.14451025426387787,
"logits/rejected": 0.23547664284706116,
"logps/chosen": -704.7767333984375,
"logps/rejected": -874.5423583984375,
"loss": 0.2143,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.36400097608566284,
"rewards/margins": 0.21412873268127441,
"rewards/rejected": -0.5781297087669373,
"step": 1400
},
{
"epoch": 0.75,
"learning_rate": 8.801887794794911e-07,
"logits/chosen": 0.057539235800504684,
"logits/rejected": 0.250136137008667,
"logps/chosen": -740.0396118164062,
"logps/rejected": -905.1911010742188,
"loss": 0.169,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.3726292550563812,
"rewards/margins": 0.24727007746696472,
"rewards/rejected": -0.619899332523346,
"step": 1410
},
{
"epoch": 0.76,
"learning_rate": 8.450098422432787e-07,
"logits/chosen": 0.061049748212099075,
"logits/rejected": 0.25947609543800354,
"logps/chosen": -725.9439697265625,
"logps/rejected": -906.8638916015625,
"loss": 0.234,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.38042646646499634,
"rewards/margins": 0.21715033054351807,
"rewards/rejected": -0.5975767970085144,
"step": 1420
},
{
"epoch": 0.76,
"learning_rate": 8.104048264413858e-07,
"logits/chosen": 0.10633182525634766,
"logits/rejected": 0.19451837241649628,
"logps/chosen": -708.9187622070312,
"logps/rejected": -866.9814453125,
"loss": 0.2272,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.3733717203140259,
"rewards/margins": 0.21014182269573212,
"rewards/rejected": -0.5835135579109192,
"step": 1430
},
{
"epoch": 0.77,
"learning_rate": 7.763857324837321e-07,
"logits/chosen": 0.12362603843212128,
"logits/rejected": 0.2242850810289383,
"logps/chosen": -738.5260009765625,
"logps/rejected": -900.5517578125,
"loss": 0.1976,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.39003950357437134,
"rewards/margins": 0.22742080688476562,
"rewards/rejected": -0.6174603700637817,
"step": 1440
},
{
"epoch": 0.77,
"learning_rate": 7.429643575928605e-07,
"logits/chosen": 0.10756851732730865,
"logits/rejected": 0.2637716233730316,
"logps/chosen": -668.5317993164062,
"logps/rejected": -864.6610107421875,
"loss": 0.2104,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3528197407722473,
"rewards/margins": 0.22202415764331818,
"rewards/rejected": -0.5748439431190491,
"step": 1450
},
{
"epoch": 0.78,
"learning_rate": 7.101522917128709e-07,
"logits/chosen": 0.13962891697883606,
"logits/rejected": 0.20390410721302032,
"logps/chosen": -674.7015380859375,
"logps/rejected": -883.2214965820312,
"loss": 0.182,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3491703271865845,
"rewards/margins": 0.23920337855815887,
"rewards/rejected": -0.5883737206459045,
"step": 1460
},
{
"epoch": 0.78,
"learning_rate": 6.779609134902312e-07,
"logits/chosen": 0.11834552139043808,
"logits/rejected": 0.21453364193439484,
"logps/chosen": -725.1397705078125,
"logps/rejected": -918.6031494140625,
"loss": 0.2117,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.38278132677078247,
"rewards/margins": 0.2413547784090042,
"rewards/rejected": -0.6241360902786255,
"step": 1470
},
{
"epoch": 0.79,
"learning_rate": 6.464013863278629e-07,
"logits/chosen": 0.11449946463108063,
"logits/rejected": 0.25861063599586487,
"logps/chosen": -732.5802612304688,
"logps/rejected": -897.5850830078125,
"loss": 0.1954,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.3760877251625061,
"rewards/margins": 0.24298934638500214,
"rewards/rejected": -0.6190770864486694,
"step": 1480
},
{
"epoch": 0.79,
"learning_rate": 6.154846545138696e-07,
"logits/chosen": 0.12956863641738892,
"logits/rejected": 0.2529379427433014,
"logps/chosen": -747.3843994140625,
"logps/rejected": -893.8402099609375,
"loss": 0.2207,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.38915324211120605,
"rewards/margins": 0.20487980544567108,
"rewards/rejected": -0.5940330624580383,
"step": 1490
},
{
"epoch": 0.8,
"learning_rate": 5.852214394262515e-07,
"logits/chosen": 0.1093892902135849,
"logits/rejected": 0.25721386075019836,
"logps/chosen": -754.2517700195312,
"logps/rejected": -897.6614990234375,
"loss": 0.1903,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.3746017813682556,
"rewards/margins": 0.22877414524555206,
"rewards/rejected": -0.6033759117126465,
"step": 1500
},
{
"epoch": 0.81,
"learning_rate": 5.556222358149191e-07,
"logits/chosen": 0.16093885898590088,
"logits/rejected": 0.24726077914237976,
"logps/chosen": -697.4678955078125,
"logps/rejected": -852.6463623046875,
"loss": 0.2077,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.343566358089447,
"rewards/margins": 0.22954010963439941,
"rewards/rejected": -0.5731065273284912,
"step": 1510
},
{
"epoch": 0.81,
"learning_rate": 5.266973081622992e-07,
"logits/chosen": 0.105752132833004,
"logits/rejected": 0.14859572052955627,
"logps/chosen": -721.93701171875,
"logps/rejected": -894.5123291015625,
"loss": 0.2127,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.38259005546569824,
"rewards/margins": 0.21411553025245667,
"rewards/rejected": -0.5967055559158325,
"step": 1520
},
{
"epoch": 0.82,
"learning_rate": 4.984566871237942e-07,
"logits/chosen": 0.10524594783782959,
"logits/rejected": 0.2024340182542801,
"logps/chosen": -694.057373046875,
"logps/rejected": -898.6950073242188,
"loss": 0.2056,
"rewards/accuracies": 0.8062499761581421,
"rewards/chosen": -0.3508402407169342,
"rewards/margins": 0.23787379264831543,
"rewards/rejected": -0.5887140035629272,
"step": 1530
},
{
"epoch": 0.82,
"learning_rate": 4.709101660493251e-07,
"logits/chosen": 0.13965895771980286,
"logits/rejected": 0.2609595060348511,
"logps/chosen": -718.2183837890625,
"logps/rejected": -927.0887451171875,
"loss": 0.1829,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -0.3423174321651459,
"rewards/margins": 0.2680014967918396,
"rewards/rejected": -0.6103189587593079,
"step": 1540
},
{
"epoch": 0.83,
"learning_rate": 4.440672975871743e-07,
"logits/chosen": 0.08242682367563248,
"logits/rejected": 0.258888840675354,
"logps/chosen": -728.785400390625,
"logps/rejected": -884.2569580078125,
"loss": 0.1945,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3698558211326599,
"rewards/margins": 0.2148553878068924,
"rewards/rejected": -0.5847111940383911,
"step": 1550
},
{
"epoch": 0.83,
"learning_rate": 4.1793739037129134e-07,
"logits/chosen": 0.06701900064945221,
"logits/rejected": 0.24965231120586395,
"logps/chosen": -727.2434692382812,
"logps/rejected": -851.6761474609375,
"loss": 0.2054,
"rewards/accuracies": 0.6937500238418579,
"rewards/chosen": -0.36560553312301636,
"rewards/margins": 0.19481703639030457,
"rewards/rejected": -0.5604225397109985,
"step": 1560
},
{
"epoch": 0.84,
"learning_rate": 3.9252950579322405e-07,
"logits/chosen": 0.10284195840358734,
"logits/rejected": 0.2730451226234436,
"logps/chosen": -668.3700561523438,
"logps/rejected": -831.0675659179688,
"loss": 0.23,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.3593829274177551,
"rewards/margins": 0.2011059820652008,
"rewards/rejected": -0.5604888796806335,
"step": 1570
},
{
"epoch": 0.84,
"learning_rate": 3.6785245485978864e-07,
"logits/chosen": 0.13599783182144165,
"logits/rejected": 0.35692065954208374,
"logps/chosen": -757.5804443359375,
"logps/rejected": -899.2823486328125,
"loss": 0.199,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.37098321318626404,
"rewards/margins": 0.23468928039073944,
"rewards/rejected": -0.6056725978851318,
"step": 1580
},
{
"epoch": 0.85,
"learning_rate": 3.43914795137566e-07,
"logits/chosen": 0.0868467465043068,
"logits/rejected": 0.2526485323905945,
"logps/chosen": -700.9779052734375,
"logps/rejected": -871.28759765625,
"loss": 0.2033,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.3747710585594177,
"rewards/margins": 0.22533288598060608,
"rewards/rejected": -0.6001039743423462,
"step": 1590
},
{
"epoch": 0.85,
"learning_rate": 3.207248277852901e-07,
"logits/chosen": 0.1354844868183136,
"logits/rejected": 0.23974338173866272,
"logps/chosen": -731.3267822265625,
"logps/rejected": -881.2257080078125,
"loss": 0.2081,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3920326828956604,
"rewards/margins": 0.2153809517621994,
"rewards/rejected": -0.607413649559021,
"step": 1600
},
{
"epoch": 0.86,
"learning_rate": 2.9829059467515074e-07,
"logits/chosen": 0.09572525322437286,
"logits/rejected": 0.2735249698162079,
"logps/chosen": -741.9659423828125,
"logps/rejected": -868.5421752929688,
"loss": 0.1862,
"rewards/accuracies": 0.7562500238418579,
"rewards/chosen": -0.36106202006340027,
"rewards/margins": 0.24586737155914307,
"rewards/rejected": -0.6069294214248657,
"step": 1610
},
{
"epoch": 0.86,
"learning_rate": 2.766198756040153e-07,
"logits/chosen": 0.11158449947834015,
"logits/rejected": 0.2957077920436859,
"logps/chosen": -739.4963989257812,
"logps/rejected": -893.3660278320312,
"loss": 0.2249,
"rewards/accuracies": 0.71875,
"rewards/chosen": -0.34823960065841675,
"rewards/margins": 0.24774250388145447,
"rewards/rejected": -0.5959821343421936,
"step": 1620
},
{
"epoch": 0.87,
"learning_rate": 2.5572018559553155e-07,
"logits/chosen": 0.08035390824079514,
"logits/rejected": 0.30342569947242737,
"logps/chosen": -726.1354370117188,
"logps/rejected": -874.1571044921875,
"loss": 0.2268,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.3561965823173523,
"rewards/margins": 0.23391160368919373,
"rewards/rejected": -0.5901081562042236,
"step": 1630
},
{
"epoch": 0.87,
"learning_rate": 2.3559877229404864e-07,
"logits/chosen": 0.11884896457195282,
"logits/rejected": 0.2436933070421219,
"logps/chosen": -684.6835327148438,
"logps/rejected": -855.82080078125,
"loss": 0.1687,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.35456180572509766,
"rewards/margins": 0.2304537296295166,
"rewards/rejected": -0.585015594959259,
"step": 1640
},
{
"epoch": 0.88,
"learning_rate": 2.1626261345126576e-07,
"logits/chosen": 0.11550426483154297,
"logits/rejected": 0.23713016510009766,
"logps/chosen": -683.0443115234375,
"logps/rejected": -845.1898193359375,
"loss": 0.2047,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.3718727231025696,
"rewards/margins": 0.1886688768863678,
"rewards/rejected": -0.560541570186615,
"step": 1650
},
{
"epoch": 0.89,
"learning_rate": 1.9771841450646505e-07,
"logits/chosen": 0.0976637750864029,
"logits/rejected": 0.2651471495628357,
"logps/chosen": -690.2272338867188,
"logps/rejected": -833.02490234375,
"loss": 0.2112,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.34249281883239746,
"rewards/margins": 0.2113989293575287,
"rewards/rejected": -0.5538917779922485,
"step": 1660
},
{
"epoch": 0.89,
"learning_rate": 1.7997260626118758e-07,
"logits/chosen": 0.16379910707473755,
"logits/rejected": 0.21980710327625275,
"logps/chosen": -689.8270263671875,
"logps/rejected": -867.986328125,
"loss": 0.1975,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3630715310573578,
"rewards/margins": 0.21778926253318787,
"rewards/rejected": -0.5808607935905457,
"step": 1670
},
{
"epoch": 0.9,
"learning_rate": 1.6303134264914365e-07,
"logits/chosen": 0.1122732162475586,
"logits/rejected": 0.25381821393966675,
"logps/chosen": -688.6129150390625,
"logps/rejected": -869.9622802734375,
"loss": 0.1925,
"rewards/accuracies": 0.7875000238418579,
"rewards/chosen": -0.3432803153991699,
"rewards/margins": 0.24642963707447052,
"rewards/rejected": -0.5897098779678345,
"step": 1680
},
{
"epoch": 0.9,
"learning_rate": 1.469004986021355e-07,
"logits/chosen": 0.09772084653377533,
"logits/rejected": 0.2000305950641632,
"logps/chosen": -650.7125854492188,
"logps/rejected": -843.9298706054688,
"loss": 0.2155,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.3303535580635071,
"rewards/margins": 0.22402143478393555,
"rewards/rejected": -0.5543749928474426,
"step": 1690
},
{
"epoch": 0.91,
"learning_rate": 1.315856680127367e-07,
"logits/chosen": 0.09905121475458145,
"logits/rejected": 0.244928240776062,
"logps/chosen": -736.1495971679688,
"logps/rejected": -871.73876953125,
"loss": 0.2122,
"rewards/accuracies": 0.668749988079071,
"rewards/chosen": -0.37346401810646057,
"rewards/margins": 0.20723024010658264,
"rewards/rejected": -0.5806942582130432,
"step": 1700
},
{
"epoch": 0.91,
"learning_rate": 1.1709216179442817e-07,
"logits/chosen": 0.12118308246135712,
"logits/rejected": 0.2801864743232727,
"logps/chosen": -676.4124755859375,
"logps/rejected": -808.5712890625,
"loss": 0.2199,
"rewards/accuracies": 0.699999988079071,
"rewards/chosen": -0.34813016653060913,
"rewards/margins": 0.19372887909412384,
"rewards/rejected": -0.5418590307235718,
"step": 1710
},
{
"epoch": 0.92,
"learning_rate": 1.0342500603986421e-07,
"logits/chosen": 0.12361860275268555,
"logits/rejected": 0.21096520125865936,
"logps/chosen": -694.0318603515625,
"logps/rejected": -827.3196411132812,
"loss": 0.2125,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.31590598821640015,
"rewards/margins": 0.2430771291255951,
"rewards/rejected": -0.5589831471443176,
"step": 1720
},
{
"epoch": 0.92,
"learning_rate": 9.058894027791643e-08,
"logits/chosen": 0.10681068897247314,
"logits/rejected": 0.29420608282089233,
"logps/chosen": -736.9918823242188,
"logps/rejected": -838.1085205078125,
"loss": 0.2178,
"rewards/accuracies": 0.7250000238418579,
"rewards/chosen": -0.3317691385746002,
"rewards/margins": 0.2118922919034958,
"rewards/rejected": -0.5436614155769348,
"step": 1730
},
{
"epoch": 0.93,
"learning_rate": 7.858841583008592e-08,
"logits/chosen": 0.11358167231082916,
"logits/rejected": 0.2305765599012375,
"logps/chosen": -730.17236328125,
"logps/rejected": -913.6432495117188,
"loss": 0.2143,
"rewards/accuracies": 0.762499988079071,
"rewards/chosen": -0.38126105070114136,
"rewards/margins": 0.22439336776733398,
"rewards/rejected": -0.6056543588638306,
"step": 1740
},
{
"epoch": 0.93,
"learning_rate": 6.742759426686313e-08,
"logits/chosen": 0.13343390822410583,
"logits/rejected": 0.19125545024871826,
"logps/chosen": -663.9968872070312,
"logps/rejected": -848.60205078125,
"loss": 0.2123,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.33876559138298035,
"rewards/margins": 0.2210654318332672,
"rewards/rejected": -0.5598310232162476,
"step": 1750
},
{
"epoch": 0.94,
"learning_rate": 5.7110345964571104e-08,
"logits/chosen": 0.11130712181329727,
"logits/rejected": 0.20762452483177185,
"logps/chosen": -697.3875122070312,
"logps/rejected": -888.5732421875,
"loss": 0.1992,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.36774367094039917,
"rewards/margins": 0.21223720908164978,
"rewards/rejected": -0.5799809098243713,
"step": 1760
},
{
"epoch": 0.94,
"learning_rate": 4.764024876318357e-08,
"logits/chosen": 0.16033565998077393,
"logits/rejected": 0.24279725551605225,
"logps/chosen": -707.9478149414062,
"logps/rejected": -846.6570434570312,
"loss": 0.2023,
"rewards/accuracies": 0.7437499761581421,
"rewards/chosen": -0.34674227237701416,
"rewards/margins": 0.22830457985401154,
"rewards/rejected": -0.5750468969345093,
"step": 1770
},
{
"epoch": 0.95,
"learning_rate": 3.902058672559633e-08,
"logits/chosen": 0.08853977173566818,
"logits/rejected": 0.2430170327425003,
"logps/chosen": -619.2088623046875,
"logps/rejected": -804.2175903320312,
"loss": 0.2185,
"rewards/accuracies": 0.731249988079071,
"rewards/chosen": -0.3369048237800598,
"rewards/margins": 0.20054855942726135,
"rewards/rejected": -0.5374533534049988,
"step": 1780
},
{
"epoch": 0.95,
"learning_rate": 3.125434899876933e-08,
"logits/chosen": 0.12984086573123932,
"logits/rejected": 0.2853730022907257,
"logps/chosen": -716.5003662109375,
"logps/rejected": -883.9786987304688,
"loss": 0.1965,
"rewards/accuracies": 0.8062499761581421,
"rewards/chosen": -0.36456966400146484,
"rewards/margins": 0.23666837811470032,
"rewards/rejected": -0.6012380123138428,
"step": 1790
},
{
"epoch": 0.96,
"learning_rate": 2.4344228777145873e-08,
"logits/chosen": 0.14794301986694336,
"logits/rejected": 0.24739143252372742,
"logps/chosen": -699.23828125,
"logps/rejected": -848.26953125,
"loss": 0.1938,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3432579040527344,
"rewards/margins": 0.21511439979076385,
"rewards/rejected": -0.5583722591400146,
"step": 1800
},
{
"epoch": 0.97,
"learning_rate": 1.829262236869772e-08,
"logits/chosen": 0.06064503267407417,
"logits/rejected": 0.2626185119152069,
"logps/chosen": -677.8721923828125,
"logps/rejected": -822.1427612304688,
"loss": 0.2081,
"rewards/accuracies": 0.768750011920929,
"rewards/chosen": -0.3421019911766052,
"rewards/margins": 0.22748549282550812,
"rewards/rejected": -0.5695875287055969,
"step": 1810
},
{
"epoch": 0.97,
"learning_rate": 1.3101628363929586e-08,
"logits/chosen": 0.10713140666484833,
"logits/rejected": 0.28815895318984985,
"logps/chosen": -721.2071533203125,
"logps/rejected": -887.7745361328125,
"loss": 0.2041,
"rewards/accuracies": 0.793749988079071,
"rewards/chosen": -0.33774304389953613,
"rewards/margins": 0.25622567534446716,
"rewards/rejected": -0.5939687490463257,
"step": 1820
},
{
"epoch": 0.98,
"learning_rate": 8.773046908123195e-09,
"logits/chosen": 0.1304168999195099,
"logits/rejected": 0.24140071868896484,
"logps/chosen": -741.6827392578125,
"logps/rejected": -876.1754150390625,
"loss": 0.2191,
"rewards/accuracies": 0.793749988079071,
"rewards/chosen": -0.36193975806236267,
"rewards/margins": 0.23083019256591797,
"rewards/rejected": -0.592769980430603,
"step": 1830
},
{
"epoch": 0.98,
"learning_rate": 5.308379077080817e-09,
"logits/chosen": 0.11551783233880997,
"logits/rejected": 0.2661217153072357,
"logps/chosen": -677.058837890625,
"logps/rejected": -856.2086181640625,
"loss": 0.2126,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.3589734733104706,
"rewards/margins": 0.21206960082054138,
"rewards/rejected": -0.571043074131012,
"step": 1840
},
{
"epoch": 0.99,
"learning_rate": 2.7088263565760996e-09,
"logits/chosen": 0.11670766770839691,
"logits/rejected": 0.2479906976222992,
"logps/chosen": -728.128662109375,
"logps/rejected": -841.5104370117188,
"loss": 0.2705,
"rewards/accuracies": 0.706250011920929,
"rewards/chosen": -0.3727620840072632,
"rewards/margins": 0.18247419595718384,
"rewards/rejected": -0.555236279964447,
"step": 1850
},
{
"epoch": 0.99,
"learning_rate": 9.752902257023633e-10,
"logits/chosen": 0.054696790874004364,
"logits/rejected": 0.19346895813941956,
"logps/chosen": -699.7024536132812,
"logps/rejected": -866.5377197265625,
"loss": 0.2164,
"rewards/accuracies": 0.7749999761581421,
"rewards/chosen": -0.36183029413223267,
"rewards/margins": 0.21843652427196503,
"rewards/rejected": -0.5802668333053589,
"step": 1860
},
{
"epoch": 1.0,
"learning_rate": 1.083718442532189e-10,
"logits/chosen": 0.07530347257852554,
"logits/rejected": 0.30514389276504517,
"logps/chosen": -705.9031982421875,
"logps/rejected": -897.6741943359375,
"loss": 0.2217,
"rewards/accuracies": 0.737500011920929,
"rewards/chosen": -0.3515733778476715,
"rewards/margins": 0.2382936030626297,
"rewards/rejected": -0.5898669958114624,
"step": 1870
},
{
"epoch": 1.0,
"step": 1875,
"total_flos": 0.0,
"train_loss": 0.23145956670443216,
"train_runtime": 17376.0848,
"train_samples_per_second": 1.727,
"train_steps_per_second": 0.108
}
],
"logging_steps": 10,
"max_steps": 1875,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 100,
"total_flos": 0.0,
"train_batch_size": 4,
"trial_name": null,
"trial_params": null
}