{
  "best_metric": 7.86433744430542,
  "best_model_checkpoint": "./qwen2.5-0.5b/qwen2.5-0.5b-expo-DPO-ES2-0.1/checkpoint-100",
  "epoch": 0.9919697685403873,
  "eval_steps": 50,
  "global_step": 350,
  "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": 18.442528136913708,
      "learning_rate": 5.681818181818181e-09,
      "logits": -1.359458565711975,
      "logps": -84.69721221923828,
      "loss": 0.6931,
      "objective": 0.6931471824645996,
      "ranking_idealized": 0.6458333134651184,
      "ranking_idealized_expo": 0.5833333134651184,
      "ranking_simple": 0.5833333134651184,
      "regularize": 0.6931471824645996,
      "step": 1,
      "wo_beta": 5.271125316619873
    },
    {
      "dpo_loss": 0.6897825002670288,
      "epoch": 0.14170996693434104,
      "grad_norm": 17.687331232419115,
      "learning_rate": 2.840909090909091e-07,
      "logits": -1.467690110206604,
      "logps": -83.78176879882812,
      "loss": 0.689,
      "objective": 0.6897825002670288,
      "ranking_idealized": 0.608418345451355,
      "ranking_idealized_expo": 0.5229591727256775,
      "ranking_simple": 0.5259353518486023,
      "regularize": 0.6897825002670288,
      "step": 50,
      "wo_beta": 7.163306713104248
    },
    {
      "epoch": 0.14170996693434104,
      "eval_dpo_loss": 0.6892488598823547,
      "eval_logits": -1.486922264099121,
      "eval_logps": -90.08146667480469,
      "eval_loss": 0.6874589920043945,
      "eval_objective": 0.6892488598823547,
      "eval_ranking_idealized": 0.6030020713806152,
      "eval_ranking_idealized_expo": 0.5222567319869995,
      "eval_ranking_simple": 0.5258799195289612,
      "eval_regularize": 0.6892488598823547,
      "eval_runtime": 308.1362,
      "eval_samples_per_second": 18.79,
      "eval_steps_per_second": 1.567,
      "eval_wo_beta": 7.8856682777404785,
      "step": 50
    },
    {
      "dpo_loss": 0.6650223731994629,
      "epoch": 0.2834199338686821,
      "grad_norm": 19.38239421391755,
      "learning_rate": 5.681818181818182e-07,
      "logits": -1.5395574569702148,
      "logps": -83.69255828857422,
      "loss": 0.6673,
      "objective": 0.6650223731994629,
      "ranking_idealized": 0.6016666889190674,
      "ranking_idealized_expo": 0.5141666531562805,
      "ranking_simple": 0.527916669845581,
      "regularize": 0.6650223731994629,
      "step": 100,
      "wo_beta": 6.79261589050293
    },
    {
      "epoch": 0.2834199338686821,
      "eval_dpo_loss": 0.6835533380508423,
      "eval_logits": -1.6164112091064453,
      "eval_logps": -90.96741485595703,
      "eval_loss": 0.6808017492294312,
      "eval_objective": 0.6835533380508423,
      "eval_ranking_idealized": 0.6030020713806152,
      "eval_ranking_idealized_expo": 0.5222567319869995,
      "eval_ranking_simple": 0.5331262946128845,
      "eval_regularize": 0.6835533380508423,
      "eval_runtime": 307.7224,
      "eval_samples_per_second": 18.816,
      "eval_steps_per_second": 1.57,
      "eval_wo_beta": 7.86433744430542,
      "step": 100
    },
    {
      "dpo_loss": 0.6406400799751282,
      "epoch": 0.42512990080302315,
      "grad_norm": 18.95367759941799,
      "learning_rate": 8.522727272727273e-07,
      "logits": -1.6548144817352295,
      "logps": -86.31576538085938,
      "loss": 0.6376,
      "objective": 0.6406400799751282,
      "ranking_idealized": 0.6066666841506958,
      "ranking_idealized_expo": 0.5287500023841858,
      "ranking_simple": 0.5554166436195374,
      "regularize": 0.6406400799751282,
      "step": 150,
      "wo_beta": 6.808902740478516
    },
    {
      "epoch": 0.42512990080302315,
      "eval_dpo_loss": 0.683335542678833,
      "eval_logits": -1.6873152256011963,
      "eval_logps": -94.63862609863281,
      "eval_loss": 0.6785484552383423,
      "eval_objective": 0.683335542678833,
      "eval_ranking_idealized": 0.6030020713806152,
      "eval_ranking_idealized_expo": 0.5222567319869995,
      "eval_ranking_simple": 0.5341615080833435,
      "eval_regularize": 0.683335542678833,
      "eval_runtime": 308.25,
      "eval_samples_per_second": 18.783,
      "eval_steps_per_second": 1.567,
      "eval_wo_beta": 8.17447566986084,
      "step": 150
    },
    {
      "dpo_loss": 0.595325231552124,
      "epoch": 0.5668398677373642,
      "grad_norm": 18.26661888886481,
      "learning_rate": 9.99433669591504e-07,
      "logits": -1.7552146911621094,
      "logps": -90.12830352783203,
      "loss": 0.5955,
      "objective": 0.595325231552124,
      "ranking_idealized": 0.5924999713897705,
      "ranking_idealized_expo": 0.5166666507720947,
      "ranking_simple": 0.5649999976158142,
      "regularize": 0.595325231552124,
      "step": 200,
      "wo_beta": 6.764244556427002
    },
    {
      "epoch": 0.5668398677373642,
      "eval_dpo_loss": 0.6818291544914246,
      "eval_logits": -1.8582650423049927,
      "eval_logps": -100.27864837646484,
      "eval_loss": 0.6808217167854309,
      "eval_objective": 0.6818291544914246,
      "eval_ranking_idealized": 0.6030020713806152,
      "eval_ranking_idealized_expo": 0.5222567319869995,
      "eval_ranking_simple": 0.5341615080833435,
      "eval_regularize": 0.6818291544914246,
      "eval_runtime": 307.6797,
      "eval_samples_per_second": 18.818,
      "eval_steps_per_second": 1.57,
      "eval_wo_beta": 7.903720855712891,
      "step": 200
    },
    {
      "dpo_loss": 0.5614926218986511,
      "epoch": 0.7085498346717053,
      "grad_norm": 17.975386844775777,
      "learning_rate": 9.94624571028813e-07,
      "logits": -1.847279667854309,
      "logps": -92.22396850585938,
      "loss": 0.5623,
      "objective": 0.5614926218986511,
      "ranking_idealized": 0.5991666913032532,
      "ranking_idealized_expo": 0.5170833468437195,
      "ranking_simple": 0.5924999713897705,
      "regularize": 0.5614926218986511,
      "step": 250,
      "wo_beta": 6.590356349945068
    },
    {
      "epoch": 0.7085498346717053,
      "eval_dpo_loss": 0.675693154335022,
      "eval_logits": -1.9406613111495972,
      "eval_logps": -97.30335998535156,
      "eval_loss": 0.6757029294967651,
      "eval_objective": 0.675693154335022,
      "eval_ranking_idealized": 0.6030020713806152,
      "eval_ranking_idealized_expo": 0.5222567319869995,
      "eval_ranking_simple": 0.5362318754196167,
      "eval_regularize": 0.675693154335022,
      "eval_runtime": 307.504,
      "eval_samples_per_second": 18.829,
      "eval_steps_per_second": 1.571,
      "eval_wo_beta": 7.916144847869873,
      "step": 250
    },
    {
      "dpo_loss": 0.5231651067733765,
      "epoch": 0.8502598016060463,
      "grad_norm": 21.98858332084632,
      "learning_rate": 9.849553282839024e-07,
      "logits": -1.9162583351135254,
      "logps": -92.107177734375,
      "loss": 0.5255,
      "objective": 0.5231651067733765,
      "ranking_idealized": 0.5799999833106995,
      "ranking_idealized_expo": 0.4970833361148834,
      "ranking_simple": 0.5849999785423279,
      "regularize": 0.5231651067733765,
      "step": 300,
      "wo_beta": 6.37797212600708
    },
    {
      "epoch": 0.8502598016060463,
      "eval_dpo_loss": 0.7118534445762634,
      "eval_logits": -2.031287431716919,
      "eval_logps": -102.48200988769531,
      "eval_loss": 0.7037488222122192,
      "eval_objective": 0.7118534445762634,
      "eval_ranking_idealized": 0.6030020713806152,
      "eval_ranking_idealized_expo": 0.5222567319869995,
      "eval_ranking_simple": 0.5351966619491577,
      "eval_regularize": 0.7118534445762634,
      "eval_runtime": 307.7923,
      "eval_samples_per_second": 18.811,
      "eval_steps_per_second": 1.569,
      "eval_wo_beta": 8.7955904006958,
      "step": 300
    },
    {
      "dpo_loss": 0.4828701913356781,
      "epoch": 0.9919697685403873,
      "grad_norm": 21.148105751270688,
      "learning_rate": 9.705209506155634e-07,
      "logits": -1.863725185394287,
      "logps": -97.61394500732422,
      "loss": 0.4939,
      "objective": 0.4828701913356781,
      "ranking_idealized": 0.60916668176651,
      "ranking_idealized_expo": 0.5270833373069763,
      "ranking_simple": 0.6449999809265137,
      "regularize": 0.4828701913356781,
      "step": 350,
      "wo_beta": 5.530836582183838
    },
    {
      "epoch": 0.9919697685403873,
      "eval_dpo_loss": 0.6916147470474243,
      "eval_logits": -1.9357593059539795,
      "eval_logps": -102.14349365234375,
      "eval_loss": 0.6897423267364502,
      "eval_objective": 0.6916147470474243,
      "eval_ranking_idealized": 0.6030020713806152,
      "eval_ranking_idealized_expo": 0.5222567319869995,
      "eval_ranking_simple": 0.5419254899024963,
      "eval_regularize": 0.6916147470474243,
      "eval_runtime": 309.7171,
      "eval_samples_per_second": 18.694,
      "eval_steps_per_second": 1.559,
      "eval_wo_beta": 8.39614486694336,
      "step": 350
    },
    {
      "epoch": 0.9919697685403873,
      "step": 350,
      "total_flos": 0.0,
      "train_loss": 0.5958910301753453,
      "train_runtime": 9674.8125,
      "train_samples_per_second": 26.255,
      "train_steps_per_second": 0.182
    }
  ],
  "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
}