{
  "best_metric": 6.68360710144043,
  "best_model_checkpoint": "./qwen2.5-0.5b/qwen2.5-0.5b-expo-DPO-ES-1/checkpoint-700",
  "epoch": 2.69248937175248,
  "eval_steps": 50,
  "global_step": 950,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "dpo_loss": 0.6931471824645996,
      "epoch": 0.002834199338686821,
      "grad_norm": 184.42504161174296,
      "learning_rate": 2.840909090909091e-08,
      "logits": -1.359458565711975,
      "logps": -84.69721221923828,
      "loss": 0.6931,
      "objective": 0.6931471824645996,
      "ranking_idealized": 0.5833333134651184,
      "ranking_idealized_expo": 0.5833333134651184,
      "ranking_simple": 0.5833333134651184,
      "regularize": 0.6931471824645996,
      "step": 1,
      "wo_beta": 5.271125316619873
    },
    {
      "dpo_loss": 0.7105370759963989,
      "epoch": 0.14170996693434104,
      "grad_norm": 184.8690438155141,
      "learning_rate": 1.4204545454545458e-06,
      "logits": -1.464050531387329,
      "logps": -84.1650161743164,
      "loss": 0.7017,
      "objective": 0.7105370759963989,
      "ranking_idealized": 0.5289115905761719,
      "ranking_idealized_expo": 0.5221088528633118,
      "ranking_simple": 0.5225340127944946,
      "regularize": 0.7105370759963989,
      "step": 50,
      "wo_beta": 7.069356918334961
    },
    {
      "epoch": 0.14170996693434104,
      "eval_dpo_loss": 0.856997013092041,
      "eval_logits": -1.458188533782959,
      "eval_logps": -93.02446746826172,
      "eval_loss": 0.8469988107681274,
      "eval_objective": 0.856997013092041,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.523809552192688,
      "eval_regularize": 0.856997013092041,
      "eval_runtime": 312.0194,
      "eval_samples_per_second": 18.557,
      "eval_steps_per_second": 1.548,
      "eval_wo_beta": 7.850645065307617,
      "step": 50
    },
    {
      "dpo_loss": 0.8048794865608215,
      "epoch": 0.2834199338686821,
      "grad_norm": 165.00318532005144,
      "learning_rate": 2.8409090909090916e-06,
      "logits": -1.4549062252044678,
      "logps": -84.21845245361328,
      "loss": 0.8112,
      "objective": 0.8048794865608215,
      "ranking_idealized": 0.5241666436195374,
      "ranking_idealized_expo": 0.5137500166893005,
      "ranking_simple": 0.543749988079071,
      "regularize": 0.8048794865608215,
      "step": 100,
      "wo_beta": 6.3764801025390625
    },
    {
      "epoch": 0.2834199338686821,
      "eval_dpo_loss": 1.0273518562316895,
      "eval_logits": -1.4382514953613281,
      "eval_logps": -86.68037414550781,
      "eval_loss": 1.0529025793075562,
      "eval_objective": 1.0273518562316895,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5284678936004639,
      "eval_regularize": 1.0273518562316895,
      "eval_runtime": 308.916,
      "eval_samples_per_second": 18.743,
      "eval_steps_per_second": 1.564,
      "eval_wo_beta": 7.498167991638184,
      "step": 100
    },
    {
      "dpo_loss": 1.0717324018478394,
      "epoch": 0.42512990080302315,
      "grad_norm": 132.13237608472397,
      "learning_rate": 4.2613636363636365e-06,
      "logits": -1.3764770030975342,
      "logps": -75.389892578125,
      "loss": 1.0895,
      "objective": 1.0717324018478394,
      "ranking_idealized": 0.5333333611488342,
      "ranking_idealized_expo": 0.527916669845581,
      "ranking_simple": 0.5645833611488342,
      "regularize": 1.0717324018478394,
      "step": 150,
      "wo_beta": 6.200186729431152
    },
    {
      "epoch": 0.42512990080302315,
      "eval_dpo_loss": 1.4009839296340942,
      "eval_logits": -1.2964812517166138,
      "eval_logps": -84.43347930908203,
      "eval_loss": 1.4497517347335815,
      "eval_objective": 1.4009839296340942,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5320910811424255,
      "eval_regularize": 1.4009839296340942,
      "eval_runtime": 308.861,
      "eval_samples_per_second": 18.746,
      "eval_steps_per_second": 1.564,
      "eval_wo_beta": 7.269246578216553,
      "step": 150
    },
    {
      "dpo_loss": 1.2378751039505005,
      "epoch": 0.5668398677373642,
      "grad_norm": 120.88819473175299,
      "learning_rate": 4.997168347957521e-06,
      "logits": -1.3378355503082275,
      "logps": -75.85352325439453,
      "loss": 1.2362,
      "objective": 1.2378751039505005,
      "ranking_idealized": 0.5204166769981384,
      "ranking_idealized_expo": 0.51541668176651,
      "ranking_simple": 0.5699999928474426,
      "regularize": 1.2378751039505005,
      "step": 200,
      "wo_beta": 6.1314239501953125
    },
    {
      "epoch": 0.5668398677373642,
      "eval_dpo_loss": 1.611555576324463,
      "eval_logits": -1.2954967021942139,
      "eval_logps": -77.71936798095703,
      "eval_loss": 1.7034811973571777,
      "eval_objective": 1.611555576324463,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5320910811424255,
      "eval_regularize": 1.611555576324463,
      "eval_runtime": 309.5132,
      "eval_samples_per_second": 18.707,
      "eval_steps_per_second": 1.561,
      "eval_wo_beta": 7.226413726806641,
      "step": 200
    },
    {
      "dpo_loss": 1.2734363079071045,
      "epoch": 0.7085498346717053,
      "grad_norm": 126.97448798013814,
      "learning_rate": 4.973122855144066e-06,
      "logits": -1.1745147705078125,
      "logps": -80.53407287597656,
      "loss": 1.3151,
      "objective": 1.2734363079071045,
      "ranking_idealized": 0.5249999761581421,
      "ranking_idealized_expo": 0.5162500143051147,
      "ranking_simple": 0.5929166674613953,
      "regularize": 1.2734363079071045,
      "step": 250,
      "wo_beta": 6.07757568359375
    },
    {
      "epoch": 0.7085498346717053,
      "eval_dpo_loss": 1.831916093826294,
      "eval_logits": -1.2565279006958008,
      "eval_logps": -92.72237396240234,
      "eval_loss": 1.9222025871276855,
      "eval_objective": 1.831916093826294,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5310559272766113,
      "eval_regularize": 1.831916093826294,
      "eval_runtime": 308.4533,
      "eval_samples_per_second": 18.771,
      "eval_steps_per_second": 1.566,
      "eval_wo_beta": 7.185470104217529,
      "step": 250
    },
    {
      "dpo_loss": 1.1058416366577148,
      "epoch": 0.8502598016060463,
      "grad_norm": 93.62402691602496,
      "learning_rate": 4.924776641419513e-06,
      "logits": -1.027516484260559,
      "logps": -86.21515655517578,
      "loss": 1.1899,
      "objective": 1.1058416366577148,
      "ranking_idealized": 0.5062500238418579,
      "ranking_idealized_expo": 0.4950000047683716,
      "ranking_simple": 0.5833333134651184,
      "regularize": 1.1058416366577148,
      "step": 300,
      "wo_beta": 5.9321818351745605
    },
    {
      "epoch": 0.8502598016060463,
      "eval_dpo_loss": 1.9587332010269165,
      "eval_logits": -0.9785559177398682,
      "eval_logps": -90.93506622314453,
      "eval_loss": 2.0297279357910156,
      "eval_objective": 1.9587332010269165,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5367494821548462,
      "eval_regularize": 1.9587332010269165,
      "eval_runtime": 308.5179,
      "eval_samples_per_second": 18.767,
      "eval_steps_per_second": 1.566,
      "eval_wo_beta": 6.933654308319092,
      "step": 300
    },
    {
      "dpo_loss": 1.1039565801620483,
      "epoch": 0.9919697685403873,
      "grad_norm": 81.61244111449733,
      "learning_rate": 4.8526047530778175e-06,
      "logits": -0.8597685694694519,
      "logps": -83.09705352783203,
      "loss": 1.1441,
      "objective": 1.1039565801620483,
      "ranking_idealized": 0.5354166626930237,
      "ranking_idealized_expo": 0.5254166722297668,
      "ranking_simple": 0.6141666769981384,
      "regularize": 1.1039565801620483,
      "step": 350,
      "wo_beta": 5.234945297241211
    },
    {
      "epoch": 0.9919697685403873,
      "eval_dpo_loss": 2.054466485977173,
      "eval_logits": -1.0211108922958374,
      "eval_logps": -82.12905883789062,
      "eval_loss": 2.1653144359588623,
      "eval_objective": 2.054466485977173,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.542443037033081,
      "eval_regularize": 2.054466485977173,
      "eval_runtime": 308.5833,
      "eval_samples_per_second": 18.763,
      "eval_steps_per_second": 1.565,
      "eval_wo_beta": 7.001650810241699,
      "step": 350
    },
    {
      "dpo_loss": 0.7012167572975159,
      "epoch": 1.1336797354747283,
      "grad_norm": 72.06842141832568,
      "learning_rate": 4.757316345716554e-06,
      "logits": -0.8224175572395325,
      "logps": -77.92687225341797,
      "loss": 0.725,
      "objective": 0.7012167572975159,
      "ranking_idealized": 0.5412499904632568,
      "ranking_idealized_expo": 0.5320833325386047,
      "ranking_simple": 0.6483333110809326,
      "regularize": 0.7012167572975159,
      "step": 400,
      "wo_beta": 4.492737770080566
    },
    {
      "epoch": 1.1336797354747283,
      "eval_dpo_loss": 2.2359888553619385,
      "eval_logits": -0.7528972029685974,
      "eval_logps": -84.34579467773438,
      "eval_loss": 2.288560152053833,
      "eval_objective": 2.2359888553619385,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5331262946128845,
      "eval_regularize": 2.2359888553619385,
      "eval_runtime": 308.7822,
      "eval_samples_per_second": 18.751,
      "eval_steps_per_second": 1.564,
      "eval_wo_beta": 7.154054164886475,
      "step": 400
    },
    {
      "dpo_loss": 0.78798508644104,
      "epoch": 1.2753897024090695,
      "grad_norm": 67.40926541067125,
      "learning_rate": 4.639847716126855e-06,
      "logits": -0.7312514185905457,
      "logps": -80.1772232055664,
      "loss": 0.7626,
      "objective": 0.78798508644104,
      "ranking_idealized": 0.5245833396911621,
      "ranking_idealized_expo": 0.5191666483879089,
      "ranking_simple": 0.6483333110809326,
      "regularize": 0.78798508644104,
      "step": 450,
      "wo_beta": 5.0193963050842285
    },
    {
      "epoch": 1.2753897024090695,
      "eval_dpo_loss": 2.0657246112823486,
      "eval_logits": -0.8863243460655212,
      "eval_logps": -80.59545135498047,
      "eval_loss": 2.1594510078430176,
      "eval_objective": 2.0657246112823486,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.532608687877655,
      "eval_regularize": 2.0657246112823486,
      "eval_runtime": 314.8953,
      "eval_samples_per_second": 18.387,
      "eval_steps_per_second": 1.534,
      "eval_wo_beta": 6.7939372062683105,
      "step": 450
    },
    {
      "dpo_loss": 0.7955495119094849,
      "epoch": 1.4170996693434104,
      "grad_norm": 80.21112193307313,
      "learning_rate": 4.501353102310901e-06,
      "logits": -0.717734694480896,
      "logps": -76.27667999267578,
      "loss": 0.8048,
      "objective": 0.7955495119094849,
      "ranking_idealized": 0.5054166913032532,
      "ranking_idealized_expo": 0.4970833361148834,
      "ranking_simple": 0.6424999833106995,
      "regularize": 0.7955495119094849,
      "step": 500,
      "wo_beta": 4.868187427520752
    },
    {
      "epoch": 1.4170996693434104,
      "eval_dpo_loss": 2.0975162982940674,
      "eval_logits": -0.7432325482368469,
      "eval_logps": -82.348876953125,
      "eval_loss": 2.2134060859680176,
      "eval_objective": 2.0975162982940674,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5341615080833435,
      "eval_regularize": 2.0975162982940674,
      "eval_runtime": 308.5741,
      "eval_samples_per_second": 18.764,
      "eval_steps_per_second": 1.565,
      "eval_wo_beta": 6.79838752746582,
      "step": 500
    },
    {
      "dpo_loss": 0.733691394329071,
      "epoch": 1.5588096362777515,
      "grad_norm": 73.44558837421432,
      "learning_rate": 4.34319334202531e-06,
      "logits": -0.6085548996925354,
      "logps": -79.48100280761719,
      "loss": 0.7106,
      "objective": 0.733691394329071,
      "ranking_idealized": 0.5199999809265137,
      "ranking_idealized_expo": 0.5108333230018616,
      "ranking_simple": 0.6579166650772095,
      "regularize": 0.733691394329071,
      "step": 550,
      "wo_beta": 4.594529151916504
    },
    {
      "epoch": 1.5588096362777515,
      "eval_dpo_loss": 2.069565534591675,
      "eval_logits": -0.6665427088737488,
      "eval_logps": -85.06731414794922,
      "eval_loss": 2.1704676151275635,
      "eval_objective": 2.069565534591675,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5320910811424255,
      "eval_regularize": 2.069565534591675,
      "eval_runtime": 308.5,
      "eval_samples_per_second": 18.768,
      "eval_steps_per_second": 1.566,
      "eval_wo_beta": 6.861428260803223,
      "step": 550
    },
    {
      "dpo_loss": 0.6737102270126343,
      "epoch": 1.7005196032120926,
      "grad_norm": 58.923373544475695,
      "learning_rate": 4.16692250129073e-06,
      "logits": -0.5712339282035828,
      "logps": -79.50873565673828,
      "loss": 0.6934,
      "objective": 0.6737102270126343,
      "ranking_idealized": 0.5220833420753479,
      "ranking_idealized_expo": 0.5149999856948853,
      "ranking_simple": 0.6537500023841858,
      "regularize": 0.6737102270126343,
      "step": 600,
      "wo_beta": 4.632404804229736
    },
    {
      "epoch": 1.7005196032120926,
      "eval_dpo_loss": 2.0693418979644775,
      "eval_logits": -0.7358241081237793,
      "eval_logps": -81.67726135253906,
      "eval_loss": 2.2126805782318115,
      "eval_objective": 2.0693418979644775,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5362318754196167,
      "eval_regularize": 2.0693418979644775,
      "eval_runtime": 308.5887,
      "eval_samples_per_second": 18.763,
      "eval_steps_per_second": 1.565,
      "eval_wo_beta": 6.726527690887451,
      "step": 600
    },
    {
      "dpo_loss": 0.6718880534172058,
      "epoch": 1.8422295701464337,
      "grad_norm": 71.22185869553977,
      "learning_rate": 3.974272604254906e-06,
      "logits": -0.6514729261398315,
      "logps": -79.11984252929688,
      "loss": 0.6885,
      "objective": 0.6718880534172058,
      "ranking_idealized": 0.5408333539962769,
      "ranking_idealized_expo": 0.527916669845581,
      "ranking_simple": 0.6583333611488342,
      "regularize": 0.6718880534172058,
      "step": 650,
      "wo_beta": 4.779858112335205
    },
    {
      "epoch": 1.8422295701464337,
      "eval_dpo_loss": 2.1432294845581055,
      "eval_logits": -0.6786962747573853,
      "eval_logps": -82.88699340820312,
      "eval_loss": 2.2198147773742676,
      "eval_objective": 2.1432294845581055,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5362318754196167,
      "eval_regularize": 2.1432294845581055,
      "eval_runtime": 308.8348,
      "eval_samples_per_second": 18.748,
      "eval_steps_per_second": 1.564,
      "eval_wo_beta": 6.820231914520264,
      "step": 650
    },
    {
      "dpo_loss": 0.6824830770492554,
      "epoch": 1.9839395370807746,
      "grad_norm": 69.2974675476568,
      "learning_rate": 3.767136614452458e-06,
      "logits": -0.5030444860458374,
      "logps": -78.93128967285156,
      "loss": 0.6477,
      "objective": 0.6824830770492554,
      "ranking_idealized": 0.5170833468437195,
      "ranking_idealized_expo": 0.5112500190734863,
      "ranking_simple": 0.6433333158493042,
      "regularize": 0.6824830770492554,
      "step": 700,
      "wo_beta": 4.807421684265137
    },
    {
      "epoch": 1.9839395370807746,
      "eval_dpo_loss": 2.2591824531555176,
      "eval_logits": -0.659694254398346,
      "eval_logps": -83.41045379638672,
      "eval_loss": 2.3420462608337402,
      "eval_objective": 2.2591824531555176,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5403726696968079,
      "eval_regularize": 2.2591824531555176,
      "eval_runtime": 308.6048,
      "eval_samples_per_second": 18.762,
      "eval_steps_per_second": 1.565,
      "eval_wo_beta": 6.68360710144043,
      "step": 700
    },
    {
      "dpo_loss": 0.38009390234947205,
      "epoch": 2.1256495040151155,
      "grad_norm": 45.305368238738886,
      "learning_rate": 3.547549834686222e-06,
      "logits": -0.5769148468971252,
      "logps": -79.54972076416016,
      "loss": 0.3785,
      "objective": 0.38009390234947205,
      "ranking_idealized": 0.5216666460037231,
      "ranking_idealized_expo": 0.5112500190734863,
      "ranking_simple": 0.6787499785423279,
      "regularize": 0.38009390234947205,
      "step": 750,
      "wo_beta": 4.319232940673828
    },
    {
      "epoch": 2.1256495040151155,
      "eval_dpo_loss": 2.2005364894866943,
      "eval_logits": -0.7841140627861023,
      "eval_logps": -84.03689575195312,
      "eval_loss": 2.2919445037841797,
      "eval_objective": 2.2005364894866943,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.54347825050354,
      "eval_regularize": 2.2005364894866943,
      "eval_runtime": 311.4559,
      "eval_samples_per_second": 18.59,
      "eval_steps_per_second": 1.551,
      "eval_wo_beta": 6.8514323234558105,
      "step": 750
    },
    {
      "dpo_loss": 0.33919599652290344,
      "epoch": 2.2673594709494567,
      "grad_norm": 40.098481963513144,
      "learning_rate": 3.3176699082935546e-06,
      "logits": -0.6165332198143005,
      "logps": -80.33509826660156,
      "loss": 0.3316,
      "objective": 0.33919599652290344,
      "ranking_idealized": 0.5195833444595337,
      "ranking_idealized_expo": 0.512499988079071,
      "ranking_simple": 0.6795833110809326,
      "regularize": 0.33919599652290344,
      "step": 800,
      "wo_beta": 3.9735937118530273
    },
    {
      "epoch": 2.2673594709494567,
      "eval_dpo_loss": 2.1123316287994385,
      "eval_logits": -0.6766893267631531,
      "eval_logps": -84.29901123046875,
      "eval_loss": 2.2220041751861572,
      "eval_objective": 2.1123316287994385,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5408902764320374,
      "eval_regularize": 2.1123316287994385,
      "eval_runtime": 309.0664,
      "eval_samples_per_second": 18.734,
      "eval_steps_per_second": 1.563,
      "eval_wo_beta": 6.766322612762451,
      "step": 800
    },
    {
      "dpo_loss": 0.34950682520866394,
      "epoch": 2.409069437883798,
      "grad_norm": 35.45933765377846,
      "learning_rate": 3.0797556183036582e-06,
      "logits": -0.6031672954559326,
      "logps": -80.82646942138672,
      "loss": 0.3283,
      "objective": 0.34950682520866394,
      "ranking_idealized": 0.5254166722297668,
      "ranking_idealized_expo": 0.5141666531562805,
      "ranking_simple": 0.6779166460037231,
      "regularize": 0.34950682520866394,
      "step": 850,
      "wo_beta": 4.280696392059326
    },
    {
      "epoch": 2.409069437883798,
      "eval_dpo_loss": 2.221233606338501,
      "eval_logits": -0.6537899374961853,
      "eval_logps": -85.08342742919922,
      "eval_loss": 2.3019895553588867,
      "eval_objective": 2.221233606338501,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5408902764320374,
      "eval_regularize": 2.221233606338501,
      "eval_runtime": 308.3167,
      "eval_samples_per_second": 18.779,
      "eval_steps_per_second": 1.567,
      "eval_wo_beta": 6.777287006378174,
      "step": 850
    },
    {
      "dpo_loss": 0.35370326042175293,
      "epoch": 2.550779404818139,
      "grad_norm": 63.704496862923826,
      "learning_rate": 2.8361446928038298e-06,
      "logits": -0.600394606590271,
      "logps": -79.94818115234375,
      "loss": 0.3516,
      "objective": 0.35370326042175293,
      "ranking_idealized": 0.5249999761581421,
      "ranking_idealized_expo": 0.5183333158493042,
      "ranking_simple": 0.6729166507720947,
      "regularize": 0.35370326042175293,
      "step": 900,
      "wo_beta": 4.2681732177734375
    },
    {
      "epoch": 2.550779404818139,
      "eval_dpo_loss": 2.1911425590515137,
      "eval_logits": -0.6225402355194092,
      "eval_logps": -84.7564468383789,
      "eval_loss": 2.272311210632324,
      "eval_objective": 2.1911425590515137,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.5362318754196167,
      "eval_regularize": 2.1911425590515137,
      "eval_runtime": 308.2356,
      "eval_samples_per_second": 18.784,
      "eval_steps_per_second": 1.567,
      "eval_wo_beta": 6.816178798675537,
      "step": 900
    },
    {
      "dpo_loss": 0.38168859481811523,
      "epoch": 2.69248937175248,
      "grad_norm": 42.960212326170364,
      "learning_rate": 2.5892308345974517e-06,
      "logits": -0.5876100659370422,
      "logps": -80.23540496826172,
      "loss": 0.3245,
      "objective": 0.38168859481811523,
      "ranking_idealized": 0.5162500143051147,
      "ranking_idealized_expo": 0.5058333277702332,
      "ranking_simple": 0.6866666674613953,
      "regularize": 0.38168859481811523,
      "step": 950,
      "wo_beta": 4.207254409790039
    },
    {
      "epoch": 2.69248937175248,
      "eval_dpo_loss": 2.2523224353790283,
      "eval_logits": -0.7128884196281433,
      "eval_logps": -83.6421127319336,
      "eval_loss": 2.3303723335266113,
      "eval_objective": 2.2523224353790283,
      "eval_ranking_idealized": 0.5295031070709229,
      "eval_ranking_idealized_expo": 0.5212215185165405,
      "eval_ranking_simple": 0.533643901348114,
      "eval_regularize": 2.2523224353790283,
      "eval_runtime": 308.3878,
      "eval_samples_per_second": 18.775,
      "eval_steps_per_second": 1.566,
      "eval_wo_beta": 6.89421272277832,
      "step": 950
    },
    {
      "epoch": 2.69248937175248,
      "step": 950,
      "total_flos": 0.0,
      "train_loss": 0.7491915085441188,
      "train_runtime": 26152.6935,
      "train_samples_per_second": 9.713,
      "train_steps_per_second": 0.067
    }
  ],
  "logging_steps": 50,
  "max_steps": 1760,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 50,
  "stateful_callbacks": {
    "EarlyStoppingCallback": {
      "args": {
        "early_stopping_patience": 5,
        "early_stopping_threshold": 0.0
      },
      "attributes": {
        "early_stopping_patience_counter": 0
      }
    },
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}