{
  "best_metric": 0.3286580741405487,
  "best_model_checkpoint": "experts/mistralic-expert-17/checkpoint-600",
  "epoch": 0.24870466321243523,
  "eval_steps": 200,
  "global_step": 600,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 0.0002,
      "loss": 0.4039,
      "step": 10
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.4071,
      "step": 20
    },
    {
      "epoch": 0.01,
      "learning_rate": 0.0002,
      "loss": 0.3793,
      "step": 30
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3936,
      "step": 40
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3589,
      "step": 50
    },
    {
      "epoch": 0.02,
      "learning_rate": 0.0002,
      "loss": 0.3789,
      "step": 60
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3648,
      "step": 70
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.0002,
      "loss": 0.3755,
      "step": 80
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.3395,
      "step": 90
    },
    {
      "epoch": 0.04,
      "learning_rate": 0.0002,
      "loss": 0.366,
      "step": 100
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3333,
      "step": 110
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3233,
      "step": 120
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.0002,
      "loss": 0.3602,
      "step": 130
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3533,
      "step": 140
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0002,
      "loss": 0.3759,
      "step": 150
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3524,
      "step": 160
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3734,
      "step": 170
    },
    {
      "epoch": 0.07,
      "learning_rate": 0.0002,
      "loss": 0.3495,
      "step": 180
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3227,
      "step": 190
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.0002,
      "loss": 0.3588,
      "step": 200
    },
    {
      "epoch": 0.08,
      "eval_loss": 0.3498784005641937,
      "eval_runtime": 158.451,
      "eval_samples_per_second": 6.311,
      "eval_steps_per_second": 3.156,
      "step": 200
    },
    {
      "epoch": 0.08,
      "mmlu_eval_accuracy": 0.5977172275916227,
      "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
      "mmlu_eval_accuracy_anatomy": 0.42857142857142855,
      "mmlu_eval_accuracy_astronomy": 0.8125,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6206896551724138,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
      "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.5454545454545454,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4878048780487805,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.625,
      "mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6046511627906976,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.17647058823529413,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.7307692307692307,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454,
      "mmlu_eval_accuracy_logical_fallacies": 0.6666666666666666,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.37,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7352941176470589,
      "mmlu_eval_accuracy_prehistory": 0.5142857142857142,
      "mmlu_eval_accuracy_professional_accounting": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_law": 0.4117647058823529,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.5942028985507246,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6666666666666666,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.8421052631578947,
      "mmlu_loss": 1.397062867918774,
      "step": 200
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3607,
      "step": 210
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.0002,
      "loss": 0.3351,
      "step": 220
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3459,
      "step": 230
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3426,
      "step": 240
    },
    {
      "epoch": 0.1,
      "learning_rate": 0.0002,
      "loss": 0.3387,
      "step": 250
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3391,
      "step": 260
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0002,
      "loss": 0.3518,
      "step": 270
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3405,
      "step": 280
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3259,
      "step": 290
    },
    {
      "epoch": 0.12,
      "learning_rate": 0.0002,
      "loss": 0.3648,
      "step": 300
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3459,
      "step": 310
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.0002,
      "loss": 0.3539,
      "step": 320
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.339,
      "step": 330
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.0002,
      "loss": 0.3132,
      "step": 340
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.334,
      "step": 350
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3466,
      "step": 360
    },
    {
      "epoch": 0.15,
      "learning_rate": 0.0002,
      "loss": 0.3353,
      "step": 370
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3327,
      "step": 380
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.0002,
      "loss": 0.3083,
      "step": 390
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3095,
      "step": 400
    },
    {
      "epoch": 0.17,
      "eval_loss": 0.33575424551963806,
      "eval_runtime": 158.1528,
      "eval_samples_per_second": 6.323,
      "eval_steps_per_second": 3.162,
      "step": 400
    },
    {
      "epoch": 0.17,
      "mmlu_eval_accuracy": 0.5990684549053618,
      "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365,
      "mmlu_eval_accuracy_anatomy": 0.5,
      "mmlu_eval_accuracy_astronomy": 0.6875,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.5862068965517241,
      "mmlu_eval_accuracy_college_biology": 0.625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
      "mmlu_eval_accuracy_college_mathematics": 0.5454545454545454,
      "mmlu_eval_accuracy_college_medicine": 0.5909090909090909,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.5454545454545454,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.6666666666666666,
      "mmlu_eval_accuracy_electrical_engineering": 0.6875,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4634146341463415,
      "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
      "mmlu_eval_accuracy_global_facts": 0.4,
      "mmlu_eval_accuracy_high_school_biology": 0.59375,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.7777777777777778,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7619047619047619,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.5813953488372093,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.5769230769230769,
      "mmlu_eval_accuracy_high_school_physics": 0.058823529411764705,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
      "mmlu_eval_accuracy_high_school_world_history": 0.6923076923076923,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.8461538461538461,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.45454545454545453,
      "mmlu_eval_accuracy_management": 0.8181818181818182,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7558139534883721,
      "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842,
      "mmlu_eval_accuracy_moral_scenarios": 0.34,
      "mmlu_eval_accuracy_nutrition": 0.7272727272727273,
      "mmlu_eval_accuracy_philosophy": 0.7647058823529411,
      "mmlu_eval_accuracy_prehistory": 0.45714285714285713,
      "mmlu_eval_accuracy_professional_accounting": 0.5806451612903226,
      "mmlu_eval_accuracy_professional_law": 0.4,
      "mmlu_eval_accuracy_professional_medicine": 0.6451612903225806,
      "mmlu_eval_accuracy_professional_psychology": 0.6086956521739131,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.6296296296296297,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 0.8181818181818182,
      "mmlu_eval_accuracy_virology": 0.6111111111111112,
      "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
      "mmlu_loss": 1.5875697279879069,
      "step": 400
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3371,
      "step": 410
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0002,
      "loss": 0.3148,
      "step": 420
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3234,
      "step": 430
    },
    {
      "epoch": 0.18,
      "learning_rate": 0.0002,
      "loss": 0.3495,
      "step": 440
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.302,
      "step": 450
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.337,
      "step": 460
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.0002,
      "loss": 0.3219,
      "step": 470
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.3364,
      "step": 480
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.0002,
      "loss": 0.2821,
      "step": 490
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3121,
      "step": 500
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.0002,
      "loss": 0.3349,
      "step": 510
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3251,
      "step": 520
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3294,
      "step": 530
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0002,
      "loss": 0.3251,
      "step": 540
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3288,
      "step": 550
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.0002,
      "loss": 0.3255,
      "step": 560
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3144,
      "step": 570
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3193,
      "step": 580
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.0002,
      "loss": 0.3127,
      "step": 590
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0002,
      "loss": 0.3119,
      "step": 600
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.3286580741405487,
      "eval_runtime": 158.1596,
      "eval_samples_per_second": 6.323,
      "eval_steps_per_second": 3.161,
      "step": 600
    },
    {
      "epoch": 0.25,
      "mmlu_eval_accuracy": 0.6005865646844798,
      "mmlu_eval_accuracy_abstract_algebra": 0.45454545454545453,
      "mmlu_eval_accuracy_anatomy": 0.35714285714285715,
      "mmlu_eval_accuracy_astronomy": 0.8125,
      "mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
      "mmlu_eval_accuracy_clinical_knowledge": 0.6551724137931034,
      "mmlu_eval_accuracy_college_biology": 0.5625,
      "mmlu_eval_accuracy_college_chemistry": 0.375,
      "mmlu_eval_accuracy_college_computer_science": 0.2727272727272727,
      "mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
      "mmlu_eval_accuracy_college_medicine": 0.6363636363636364,
      "mmlu_eval_accuracy_college_physics": 0.45454545454545453,
      "mmlu_eval_accuracy_computer_security": 0.7272727272727273,
      "mmlu_eval_accuracy_conceptual_physics": 0.5,
      "mmlu_eval_accuracy_econometrics": 0.5833333333333334,
      "mmlu_eval_accuracy_electrical_engineering": 0.625,
      "mmlu_eval_accuracy_elementary_mathematics": 0.4634146341463415,
      "mmlu_eval_accuracy_formal_logic": 0.14285714285714285,
      "mmlu_eval_accuracy_global_facts": 0.3,
      "mmlu_eval_accuracy_high_school_biology": 0.6875,
      "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
      "mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
      "mmlu_eval_accuracy_high_school_european_history": 0.8333333333333334,
      "mmlu_eval_accuracy_high_school_geography": 0.8636363636363636,
      "mmlu_eval_accuracy_high_school_government_and_politics": 0.7142857142857143,
      "mmlu_eval_accuracy_high_school_macroeconomics": 0.6046511627906976,
      "mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
      "mmlu_eval_accuracy_high_school_microeconomics": 0.6538461538461539,
      "mmlu_eval_accuracy_high_school_physics": 0.11764705882352941,
      "mmlu_eval_accuracy_high_school_psychology": 0.8666666666666667,
      "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173,
      "mmlu_eval_accuracy_high_school_us_history": 0.8181818181818182,
      "mmlu_eval_accuracy_high_school_world_history": 0.6538461538461539,
      "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
      "mmlu_eval_accuracy_human_sexuality": 0.5,
      "mmlu_eval_accuracy_international_law": 0.9230769230769231,
      "mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
      "mmlu_eval_accuracy_logical_fallacies": 0.6111111111111112,
      "mmlu_eval_accuracy_machine_learning": 0.36363636363636365,
      "mmlu_eval_accuracy_management": 0.9090909090909091,
      "mmlu_eval_accuracy_marketing": 0.92,
      "mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
      "mmlu_eval_accuracy_miscellaneous": 0.7441860465116279,
      "mmlu_eval_accuracy_moral_disputes": 0.5526315789473685,
      "mmlu_eval_accuracy_moral_scenarios": 0.35,
      "mmlu_eval_accuracy_nutrition": 0.6666666666666666,
      "mmlu_eval_accuracy_philosophy": 0.7647058823529411,
      "mmlu_eval_accuracy_prehistory": 0.6,
      "mmlu_eval_accuracy_professional_accounting": 0.6129032258064516,
      "mmlu_eval_accuracy_professional_law": 0.38235294117647056,
      "mmlu_eval_accuracy_professional_medicine": 0.6774193548387096,
      "mmlu_eval_accuracy_professional_psychology": 0.5797101449275363,
      "mmlu_eval_accuracy_public_relations": 0.5,
      "mmlu_eval_accuracy_security_studies": 0.7037037037037037,
      "mmlu_eval_accuracy_sociology": 0.9545454545454546,
      "mmlu_eval_accuracy_us_foreign_policy": 0.9090909090909091,
      "mmlu_eval_accuracy_virology": 0.5555555555555556,
      "mmlu_eval_accuracy_world_religions": 0.7894736842105263,
      "mmlu_loss": 1.7061042173257384,
      "step": 600
    }
  ],
  "logging_steps": 10,
  "max_steps": 7236,
  "num_train_epochs": 3,
  "save_steps": 200,
  "total_flos": 3.6867039328272384e+17,
  "trial_name": null,
  "trial_params": null
}