k-r-l's picture
Training in progress, step 36, checkpoint
4a88de7 verified
raw
history blame
24.1 kB
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.733991717107359,
"eval_steps": 500,
"global_step": 36,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02038865880853775,
"grad_norm": 10.073711395263672,
"learning_rate": 1e-05,
"log_odds_chosen": -0.03190982714295387,
"log_odds_ratio": -0.7741295099258423,
"logits/chosen": -2.249119758605957,
"logits/rejected": -2.2597196102142334,
"logps/chosen": -5.181360721588135,
"logps/rejected": -5.149355888366699,
"loss": 5.4553,
"nll_loss": 5.37792444229126,
"rewards/accuracies": 0.4375,
"rewards/chosen": -0.5181360840797424,
"rewards/margins": -0.0032004904933273792,
"rewards/rejected": -0.5149356126785278,
"step": 1
},
{
"epoch": 0.0407773176170755,
"grad_norm": 9.501357078552246,
"learning_rate": 2e-05,
"log_odds_chosen": -0.044936653226614,
"log_odds_ratio": -0.7695960402488708,
"logits/chosen": -2.2498621940612793,
"logits/rejected": -2.260127067565918,
"logps/chosen": -5.01981258392334,
"logps/rejected": -4.974053382873535,
"loss": 5.3213,
"nll_loss": 5.2443060874938965,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.501981258392334,
"rewards/margins": -0.0045759049244225025,
"rewards/rejected": -0.49740538001060486,
"step": 2
},
{
"epoch": 0.06116597642561325,
"grad_norm": 9.111624717712402,
"learning_rate": 3e-05,
"log_odds_chosen": -0.008361914195120335,
"log_odds_ratio": -0.7590417265892029,
"logits/chosen": -2.287618398666382,
"logits/rejected": -2.3050174713134766,
"logps/chosen": -4.931052207946777,
"logps/rejected": -4.922361850738525,
"loss": 5.195,
"nll_loss": 5.119076251983643,
"rewards/accuracies": 0.453125,
"rewards/chosen": -0.4931052029132843,
"rewards/margins": -0.0008690543472766876,
"rewards/rejected": -0.4922361373901367,
"step": 3
},
{
"epoch": 0.081554635234151,
"grad_norm": 9.194869995117188,
"learning_rate": 4e-05,
"log_odds_chosen": -0.06829527020454407,
"log_odds_ratio": -0.7628177404403687,
"logits/chosen": -2.2833240032196045,
"logits/rejected": -2.2870724201202393,
"logps/chosen": -4.916200160980225,
"logps/rejected": -4.847894668579102,
"loss": 5.1872,
"nll_loss": 5.110929489135742,
"rewards/accuracies": 0.421875,
"rewards/chosen": -0.4916200041770935,
"rewards/margins": -0.006830527447164059,
"rewards/rejected": -0.48478949069976807,
"step": 4
},
{
"epoch": 0.10194329404268876,
"grad_norm": 9.331730842590332,
"learning_rate": 5e-05,
"log_odds_chosen": -0.07571306079626083,
"log_odds_ratio": -0.7841904163360596,
"logits/chosen": -2.3229925632476807,
"logits/rejected": -2.337273120880127,
"logps/chosen": -4.687668323516846,
"logps/rejected": -4.612335205078125,
"loss": 4.9652,
"nll_loss": 4.886806011199951,
"rewards/accuracies": 0.5,
"rewards/chosen": -0.46876680850982666,
"rewards/margins": -0.0075332350097596645,
"rewards/rejected": -0.46123358607292175,
"step": 5
},
{
"epoch": 0.1223319528512265,
"grad_norm": 9.857681274414062,
"learning_rate": 6e-05,
"log_odds_chosen": -0.19471649825572968,
"log_odds_ratio": -0.8327277302742004,
"logits/chosen": -2.3558108806610107,
"logits/rejected": -2.3881752490997314,
"logps/chosen": -4.344555377960205,
"logps/rejected": -4.152055263519287,
"loss": 4.6379,
"nll_loss": 4.554634094238281,
"rewards/accuracies": 0.40625,
"rewards/chosen": -0.4344555735588074,
"rewards/margins": -0.019250018522143364,
"rewards/rejected": -0.41520553827285767,
"step": 6
},
{
"epoch": 0.14272061165976427,
"grad_norm": 7.811158657073975,
"learning_rate": 7e-05,
"log_odds_chosen": -0.0166391022503376,
"log_odds_ratio": -0.7380757331848145,
"logits/chosen": -2.388141632080078,
"logits/rejected": -2.4195711612701416,
"logps/chosen": -3.969034194946289,
"logps/rejected": -3.951852798461914,
"loss": 4.2265,
"nll_loss": 4.152710914611816,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.39690345525741577,
"rewards/margins": -0.0017181318253278732,
"rewards/rejected": -0.39518529176712036,
"step": 7
},
{
"epoch": 0.163109270468302,
"grad_norm": 7.66775369644165,
"learning_rate": 8e-05,
"log_odds_chosen": 0.02697194740176201,
"log_odds_ratio": -0.7153270840644836,
"logits/chosen": -2.477844476699829,
"logits/rejected": -2.474121570587158,
"logps/chosen": -3.4541428089141846,
"logps/rejected": -3.4769277572631836,
"loss": 3.6939,
"nll_loss": 3.6224091053009033,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.34541425108909607,
"rewards/margins": 0.0022785186301916838,
"rewards/rejected": -0.3476927876472473,
"step": 8
},
{
"epoch": 0.18349792927683975,
"grad_norm": 6.6126298904418945,
"learning_rate": 9e-05,
"log_odds_chosen": 0.16790322959423065,
"log_odds_ratio": -0.6347489356994629,
"logits/chosen": -2.5679516792297363,
"logits/rejected": -2.5715692043304443,
"logps/chosen": -2.9835710525512695,
"logps/rejected": -3.140634059906006,
"loss": 3.2274,
"nll_loss": 3.163942813873291,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.2983570992946625,
"rewards/margins": 0.015706289559602737,
"rewards/rejected": -0.3140634000301361,
"step": 9
},
{
"epoch": 0.20388658808537752,
"grad_norm": 6.9558281898498535,
"learning_rate": 0.0001,
"log_odds_chosen": 0.17972886562347412,
"log_odds_ratio": -0.6277650594711304,
"logits/chosen": -2.6125497817993164,
"logits/rejected": -2.6118133068084717,
"logps/chosen": -2.6501834392547607,
"logps/rejected": -2.8179259300231934,
"loss": 2.8684,
"nll_loss": 2.8056435585021973,
"rewards/accuracies": 0.703125,
"rewards/chosen": -0.26501837372779846,
"rewards/margins": 0.01677425391972065,
"rewards/rejected": -0.2817925810813904,
"step": 10
},
{
"epoch": 0.22427524689391526,
"grad_norm": 7.142885684967041,
"learning_rate": 9.743589743589744e-05,
"log_odds_chosen": 0.11056404560804367,
"log_odds_ratio": -0.662803053855896,
"logits/chosen": -2.665982723236084,
"logits/rejected": -2.6722326278686523,
"logps/chosen": -2.1527411937713623,
"logps/rejected": -2.2486300468444824,
"loss": 2.383,
"nll_loss": 2.316676378250122,
"rewards/accuracies": 0.625,
"rewards/chosen": -0.2152741551399231,
"rewards/margins": 0.009588859975337982,
"rewards/rejected": -0.2248629927635193,
"step": 11
},
{
"epoch": 0.244663905702453,
"grad_norm": 4.808487415313721,
"learning_rate": 9.487179487179487e-05,
"log_odds_chosen": 0.09381386637687683,
"log_odds_ratio": -0.6690701842308044,
"logits/chosen": -2.6697304248809814,
"logits/rejected": -2.684809446334839,
"logps/chosen": -1.6216576099395752,
"logps/rejected": -1.695598840713501,
"loss": 1.9113,
"nll_loss": 1.8443692922592163,
"rewards/accuracies": 0.53125,
"rewards/chosen": -0.16216576099395752,
"rewards/margins": 0.007394128944724798,
"rewards/rejected": -0.16955989599227905,
"step": 12
},
{
"epoch": 0.26505256451099074,
"grad_norm": 2.420715093612671,
"learning_rate": 9.230769230769232e-05,
"log_odds_chosen": 0.10657332092523575,
"log_odds_ratio": -0.6556077599525452,
"logits/chosen": -2.683443784713745,
"logits/rejected": -2.6935503482818604,
"logps/chosen": -1.4284594058990479,
"logps/rejected": -1.508366584777832,
"loss": 1.7205,
"nll_loss": 1.654909372329712,
"rewards/accuracies": 0.65625,
"rewards/chosen": -0.14284594357013702,
"rewards/margins": 0.007990704849362373,
"rewards/rejected": -0.15083666145801544,
"step": 13
},
{
"epoch": 0.28544122331952854,
"grad_norm": 2.851985216140747,
"learning_rate": 8.974358974358975e-05,
"log_odds_chosen": 0.156551793217659,
"log_odds_ratio": -0.6308416128158569,
"logits/chosen": -2.6921679973602295,
"logits/rejected": -2.6880078315734863,
"logps/chosen": -1.332141637802124,
"logps/rejected": -1.4488908052444458,
"loss": 1.5952,
"nll_loss": 1.5321555137634277,
"rewards/accuracies": 0.671875,
"rewards/chosen": -0.13321417570114136,
"rewards/margins": 0.01167491264641285,
"rewards/rejected": -0.14488908648490906,
"step": 14
},
{
"epoch": 0.3058298821280663,
"grad_norm": 2.4120404720306396,
"learning_rate": 8.717948717948718e-05,
"log_odds_chosen": 0.0856461226940155,
"log_odds_ratio": -0.6610275506973267,
"logits/chosen": -2.6284494400024414,
"logits/rejected": -2.6575706005096436,
"logps/chosen": -1.342667579650879,
"logps/rejected": -1.4053808450698853,
"loss": 1.5791,
"nll_loss": 1.5130078792572021,
"rewards/accuracies": 0.671875,
"rewards/chosen": -0.13426676392555237,
"rewards/margins": 0.006271325517445803,
"rewards/rejected": -0.1405380815267563,
"step": 15
},
{
"epoch": 0.326218540936604,
"grad_norm": 2.2445013523101807,
"learning_rate": 8.461538461538461e-05,
"log_odds_chosen": 0.18759144842624664,
"log_odds_ratio": -0.617063045501709,
"logits/chosen": -2.574859857559204,
"logits/rejected": -2.5844004154205322,
"logps/chosen": -1.24087393283844,
"logps/rejected": -1.3730320930480957,
"loss": 1.5016,
"nll_loss": 1.439911127090454,
"rewards/accuracies": 0.703125,
"rewards/chosen": -0.1240873858332634,
"rewards/margins": 0.013215810991823673,
"rewards/rejected": -0.1373032033443451,
"step": 16
},
{
"epoch": 0.34660719974514176,
"grad_norm": 2.1423990726470947,
"learning_rate": 8.205128205128205e-05,
"log_odds_chosen": 0.2600902020931244,
"log_odds_ratio": -0.5869597792625427,
"logits/chosen": -2.4825172424316406,
"logits/rejected": -2.4918212890625,
"logps/chosen": -1.1209421157836914,
"logps/rejected": -1.2962286472320557,
"loss": 1.4016,
"nll_loss": 1.3428654670715332,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.1120942085981369,
"rewards/margins": 0.01752866432070732,
"rewards/rejected": -0.12962287664413452,
"step": 17
},
{
"epoch": 0.3669958585536795,
"grad_norm": 2.0624334812164307,
"learning_rate": 7.948717948717948e-05,
"log_odds_chosen": 0.19398407638072968,
"log_odds_ratio": -0.6082965731620789,
"logits/chosen": -2.403656482696533,
"logits/rejected": -2.4159321784973145,
"logps/chosen": -1.0768545866012573,
"logps/rejected": -1.204436182975769,
"loss": 1.3427,
"nll_loss": 1.2818515300750732,
"rewards/accuracies": 0.796875,
"rewards/chosen": -0.10768546164035797,
"rewards/margins": 0.012758150696754456,
"rewards/rejected": -0.12044361233711243,
"step": 18
},
{
"epoch": 0.38738451736221724,
"grad_norm": 2.0077810287475586,
"learning_rate": 7.692307692307693e-05,
"log_odds_chosen": 0.2774355113506317,
"log_odds_ratio": -0.5771675109863281,
"logits/chosen": -2.32011079788208,
"logits/rejected": -2.309196949005127,
"logps/chosen": -0.982606828212738,
"logps/rejected": -1.1639189720153809,
"loss": 1.272,
"nll_loss": 1.2142971754074097,
"rewards/accuracies": 0.828125,
"rewards/chosen": -0.09826068580150604,
"rewards/margins": 0.01813122071325779,
"rewards/rejected": -0.11639191210269928,
"step": 19
},
{
"epoch": 0.40777317617075504,
"grad_norm": 1.5790979862213135,
"learning_rate": 7.435897435897436e-05,
"log_odds_chosen": 0.3015543520450592,
"log_odds_ratio": -0.571921706199646,
"logits/chosen": -2.204582691192627,
"logits/rejected": -2.2353363037109375,
"logps/chosen": -0.8750602006912231,
"logps/rejected": -1.055103063583374,
"loss": 1.1722,
"nll_loss": 1.1150364875793457,
"rewards/accuracies": 0.828125,
"rewards/chosen": -0.08750602602958679,
"rewards/margins": 0.018004287034273148,
"rewards/rejected": -0.10551030933856964,
"step": 20
},
{
"epoch": 0.4281618349792928,
"grad_norm": 1.4550942182540894,
"learning_rate": 7.17948717948718e-05,
"log_odds_chosen": 0.2803484797477722,
"log_odds_ratio": -0.5891110897064209,
"logits/chosen": -2.177445888519287,
"logits/rejected": -2.1862730979919434,
"logps/chosen": -0.8740922808647156,
"logps/rejected": -1.0376415252685547,
"loss": 1.184,
"nll_loss": 1.125113606452942,
"rewards/accuracies": 0.6875,
"rewards/chosen": -0.08740923553705215,
"rewards/margins": 0.01635492593050003,
"rewards/rejected": -0.10376415401697159,
"step": 21
},
{
"epoch": 0.4485504937878305,
"grad_norm": 1.5131646394729614,
"learning_rate": 6.923076923076924e-05,
"log_odds_chosen": 0.3196752965450287,
"log_odds_ratio": -0.5673432350158691,
"logits/chosen": -2.139277458190918,
"logits/rejected": -2.1643970012664795,
"logps/chosen": -0.8622347116470337,
"logps/rejected": -1.060903549194336,
"loss": 1.1375,
"nll_loss": 1.0807565450668335,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.08622346818447113,
"rewards/margins": 0.019866881892085075,
"rewards/rejected": -0.10609035938978195,
"step": 22
},
{
"epoch": 0.46893915259636826,
"grad_norm": 1.7129428386688232,
"learning_rate": 6.666666666666667e-05,
"log_odds_chosen": 0.3558296263217926,
"log_odds_ratio": -0.551045298576355,
"logits/chosen": -2.1384575366973877,
"logits/rejected": -2.1461870670318604,
"logps/chosen": -0.8587465286254883,
"logps/rejected": -1.0661779642105103,
"loss": 1.1327,
"nll_loss": 1.0775768756866455,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.08587465435266495,
"rewards/margins": 0.02074313722550869,
"rewards/rejected": -0.10661779344081879,
"step": 23
},
{
"epoch": 0.489327811404906,
"grad_norm": 1.7440029382705688,
"learning_rate": 6.410256410256412e-05,
"log_odds_chosen": 0.32858026027679443,
"log_odds_ratio": -0.5619024038314819,
"logits/chosen": -2.1546478271484375,
"logits/rejected": -2.1749908924102783,
"logps/chosen": -0.835049033164978,
"logps/rejected": -1.0198912620544434,
"loss": 1.1057,
"nll_loss": 1.0495383739471436,
"rewards/accuracies": 0.78125,
"rewards/chosen": -0.08350490033626556,
"rewards/margins": 0.01848422922194004,
"rewards/rejected": -0.10198913514614105,
"step": 24
},
{
"epoch": 0.5097164702134438,
"grad_norm": 1.9139939546585083,
"learning_rate": 6.153846153846155e-05,
"log_odds_chosen": 0.32062438130378723,
"log_odds_ratio": -0.5715374946594238,
"logits/chosen": -2.172311305999756,
"logits/rejected": -2.1929304599761963,
"logps/chosen": -0.8502154350280762,
"logps/rejected": -1.0373834371566772,
"loss": 1.1209,
"nll_loss": 1.0637274980545044,
"rewards/accuracies": 0.75,
"rewards/chosen": -0.08502154797315598,
"rewards/margins": 0.018716804683208466,
"rewards/rejected": -0.10373835265636444,
"step": 25
},
{
"epoch": 0.5301051290219815,
"grad_norm": 2.14544415473938,
"learning_rate": 5.897435897435898e-05,
"log_odds_chosen": 0.38874953985214233,
"log_odds_ratio": -0.5423570275306702,
"logits/chosen": -2.2168030738830566,
"logits/rejected": -2.2049760818481445,
"logps/chosen": -0.8020574450492859,
"logps/rejected": -1.0160338878631592,
"loss": 1.068,
"nll_loss": 1.013757348060608,
"rewards/accuracies": 0.84375,
"rewards/chosen": -0.08020574599504471,
"rewards/margins": 0.02139764279127121,
"rewards/rejected": -0.10160338878631592,
"step": 26
},
{
"epoch": 0.5504937878305193,
"grad_norm": 2.258091449737549,
"learning_rate": 5.6410256410256414e-05,
"log_odds_chosen": 0.44487276673316956,
"log_odds_ratio": -0.5278628468513489,
"logits/chosen": -2.2525055408477783,
"logits/rejected": -2.233915328979492,
"logps/chosen": -0.7687065005302429,
"logps/rejected": -1.0176327228546143,
"loss": 1.0399,
"nll_loss": 0.9870870113372803,
"rewards/accuracies": 0.796875,
"rewards/chosen": -0.07687065005302429,
"rewards/margins": 0.02489262819290161,
"rewards/rejected": -0.1017632856965065,
"step": 27
},
{
"epoch": 0.5708824466390571,
"grad_norm": 2.6239125728607178,
"learning_rate": 5.384615384615385e-05,
"log_odds_chosen": 0.38128751516342163,
"log_odds_ratio": -0.55021071434021,
"logits/chosen": -2.2505478858947754,
"logits/rejected": -2.2465157508850098,
"logps/chosen": -0.7138630151748657,
"logps/rejected": -0.9125551581382751,
"loss": 1.0329,
"nll_loss": 0.977867603302002,
"rewards/accuracies": 0.765625,
"rewards/chosen": -0.07138630747795105,
"rewards/margins": 0.019869212061166763,
"rewards/rejected": -0.09125552326440811,
"step": 28
},
{
"epoch": 0.5912711054475948,
"grad_norm": 1.6056082248687744,
"learning_rate": 5.128205128205128e-05,
"log_odds_chosen": 0.5487304925918579,
"log_odds_ratio": -0.478320449590683,
"logits/chosen": -2.2418763637542725,
"logits/rejected": -2.2638566493988037,
"logps/chosen": -0.7001346349716187,
"logps/rejected": -1.0050172805786133,
"loss": 0.9538,
"nll_loss": 0.9059388041496277,
"rewards/accuracies": 0.921875,
"rewards/chosen": -0.07001346349716187,
"rewards/margins": 0.030488261952996254,
"rewards/rejected": -0.10050173103809357,
"step": 29
},
{
"epoch": 0.6116597642561326,
"grad_norm": 2.707125663757324,
"learning_rate": 4.871794871794872e-05,
"log_odds_chosen": 0.433153361082077,
"log_odds_ratio": -0.5276967883110046,
"logits/chosen": -2.2189157009124756,
"logits/rejected": -2.228149890899658,
"logps/chosen": -0.6423375606536865,
"logps/rejected": -0.8684220910072327,
"loss": 0.943,
"nll_loss": 0.89023357629776,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.06423375755548477,
"rewards/margins": 0.022608455270528793,
"rewards/rejected": -0.08684220910072327,
"step": 30
},
{
"epoch": 0.6320484230646702,
"grad_norm": 3.1514434814453125,
"learning_rate": 4.615384615384616e-05,
"log_odds_chosen": 0.3669428527355194,
"log_odds_ratio": -0.5465207695960999,
"logits/chosen": -2.243276357650757,
"logits/rejected": -2.231745719909668,
"logps/chosen": -0.7233006954193115,
"logps/rejected": -0.9113064408302307,
"loss": 1.0016,
"nll_loss": 0.9469515681266785,
"rewards/accuracies": 0.8125,
"rewards/chosen": -0.07233007252216339,
"rewards/margins": 0.018800577148795128,
"rewards/rejected": -0.09113065153360367,
"step": 31
},
{
"epoch": 0.652437081873208,
"grad_norm": 2.5206167697906494,
"learning_rate": 4.358974358974359e-05,
"log_odds_chosen": 0.5872430205345154,
"log_odds_ratio": -0.46654394268989563,
"logits/chosen": -2.2469942569732666,
"logits/rejected": -2.269362211227417,
"logps/chosen": -0.6239035725593567,
"logps/rejected": -0.9376953840255737,
"loss": 0.9061,
"nll_loss": 0.8594872951507568,
"rewards/accuracies": 0.875,
"rewards/chosen": -0.06239035725593567,
"rewards/margins": 0.03137918934226036,
"rewards/rejected": -0.09376954287290573,
"step": 32
},
{
"epoch": 0.6728257406817458,
"grad_norm": 1.3284610509872437,
"learning_rate": 4.1025641025641023e-05,
"log_odds_chosen": 0.5279171466827393,
"log_odds_ratio": -0.49398207664489746,
"logits/chosen": -2.263728618621826,
"logits/rejected": -2.288696050643921,
"logps/chosen": -0.6505022644996643,
"logps/rejected": -0.9282124042510986,
"loss": 0.9288,
"nll_loss": 0.8793907761573792,
"rewards/accuracies": 0.890625,
"rewards/chosen": -0.06505022943019867,
"rewards/margins": 0.027771014720201492,
"rewards/rejected": -0.09282123297452927,
"step": 33
},
{
"epoch": 0.6932143994902835,
"grad_norm": 1.2634799480438232,
"learning_rate": 3.846153846153846e-05,
"log_odds_chosen": 0.5196709036827087,
"log_odds_ratio": -0.5028817653656006,
"logits/chosen": -2.3093760013580322,
"logits/rejected": -2.3152754306793213,
"logps/chosen": -0.6717790365219116,
"logps/rejected": -0.9456109404563904,
"loss": 0.9413,
"nll_loss": 0.8909698724746704,
"rewards/accuracies": 0.796875,
"rewards/chosen": -0.0671778991818428,
"rewards/margins": 0.027383197098970413,
"rewards/rejected": -0.09456109255552292,
"step": 34
},
{
"epoch": 0.7136030582988213,
"grad_norm": 1.2822017669677734,
"learning_rate": 3.58974358974359e-05,
"log_odds_chosen": 0.44004446268081665,
"log_odds_ratio": -0.53327876329422,
"logits/chosen": -2.3234097957611084,
"logits/rejected": -2.319546937942505,
"logps/chosen": -0.7039386630058289,
"logps/rejected": -0.9365054965019226,
"loss": 0.9674,
"nll_loss": 0.9140487909317017,
"rewards/accuracies": 0.796875,
"rewards/chosen": -0.070393867790699,
"rewards/margins": 0.023256685584783554,
"rewards/rejected": -0.09365054965019226,
"step": 35
},
{
"epoch": 0.733991717107359,
"grad_norm": 1.3868314027786255,
"learning_rate": 3.3333333333333335e-05,
"log_odds_chosen": 0.40167126059532166,
"log_odds_ratio": -0.5467191338539124,
"logits/chosen": -2.3325843811035156,
"logits/rejected": -2.3239898681640625,
"logps/chosen": -0.7655161619186401,
"logps/rejected": -0.9702532291412354,
"loss": 0.9815,
"nll_loss": 0.9267975091934204,
"rewards/accuracies": 0.765625,
"rewards/chosen": -0.07655161619186401,
"rewards/margins": 0.02047371119260788,
"rewards/rejected": -0.0970253199338913,
"step": 36
}
],
"logging_steps": 1,
"max_steps": 49,
"num_input_tokens_seen": 0,
"num_train_epochs": 1,
"save_steps": 4,
"stateful_callbacks": {
"TrainerControl": {
"args": {
"should_epoch_stop": false,
"should_evaluate": false,
"should_log": false,
"should_save": true,
"should_training_stop": false
},
"attributes": {}
}
},
"total_flos": 0.0,
"train_batch_size": 1,
"trial_name": null,
"trial_params": null
}