prateeky2806's picture
Training in progress, step 800
956b92d
raw
history blame
24.8 kB
{
"best_metric": 0.545791745185852,
"best_model_checkpoint": "./output_v2/7b_cluster016_Nous-Hermes-llama-2-7b_partitioned_v3_standardized_016/checkpoint-800",
"epoch": 1.3366750208855471,
"global_step": 800,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02,
"learning_rate": 0.0002,
"loss": 0.6709,
"step": 10
},
{
"epoch": 0.03,
"learning_rate": 0.0002,
"loss": 0.6086,
"step": 20
},
{
"epoch": 0.05,
"learning_rate": 0.0002,
"loss": 0.6457,
"step": 30
},
{
"epoch": 0.07,
"learning_rate": 0.0002,
"loss": 0.5722,
"step": 40
},
{
"epoch": 0.08,
"learning_rate": 0.0002,
"loss": 0.6039,
"step": 50
},
{
"epoch": 0.1,
"learning_rate": 0.0002,
"loss": 0.6079,
"step": 60
},
{
"epoch": 0.12,
"learning_rate": 0.0002,
"loss": 0.6261,
"step": 70
},
{
"epoch": 0.13,
"learning_rate": 0.0002,
"loss": 0.5876,
"step": 80
},
{
"epoch": 0.15,
"learning_rate": 0.0002,
"loss": 0.5645,
"step": 90
},
{
"epoch": 0.17,
"learning_rate": 0.0002,
"loss": 0.5902,
"step": 100
},
{
"epoch": 0.18,
"learning_rate": 0.0002,
"loss": 0.6008,
"step": 110
},
{
"epoch": 0.2,
"learning_rate": 0.0002,
"loss": 0.588,
"step": 120
},
{
"epoch": 0.22,
"learning_rate": 0.0002,
"loss": 0.6001,
"step": 130
},
{
"epoch": 0.23,
"learning_rate": 0.0002,
"loss": 0.5684,
"step": 140
},
{
"epoch": 0.25,
"learning_rate": 0.0002,
"loss": 0.5776,
"step": 150
},
{
"epoch": 0.27,
"learning_rate": 0.0002,
"loss": 0.5482,
"step": 160
},
{
"epoch": 0.28,
"learning_rate": 0.0002,
"loss": 0.5759,
"step": 170
},
{
"epoch": 0.3,
"learning_rate": 0.0002,
"loss": 0.5736,
"step": 180
},
{
"epoch": 0.32,
"learning_rate": 0.0002,
"loss": 0.5417,
"step": 190
},
{
"epoch": 0.33,
"learning_rate": 0.0002,
"loss": 0.5581,
"step": 200
},
{
"epoch": 0.33,
"eval_loss": 0.5835912227630615,
"eval_runtime": 210.8375,
"eval_samples_per_second": 4.743,
"eval_steps_per_second": 2.371,
"step": 200
},
{
"epoch": 0.33,
"mmlu_eval_accuracy": 0.46615455952712886,
"mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
"mmlu_eval_accuracy_astronomy": 0.4375,
"mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
"mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
"mmlu_eval_accuracy_college_biology": 0.5,
"mmlu_eval_accuracy_college_chemistry": 0.125,
"mmlu_eval_accuracy_college_computer_science": 0.36363636363636365,
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
"mmlu_eval_accuracy_college_medicine": 0.2727272727272727,
"mmlu_eval_accuracy_college_physics": 0.45454545454545453,
"mmlu_eval_accuracy_computer_security": 0.2727272727272727,
"mmlu_eval_accuracy_conceptual_physics": 0.34615384615384615,
"mmlu_eval_accuracy_econometrics": 0.16666666666666666,
"mmlu_eval_accuracy_electrical_engineering": 0.4375,
"mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
"mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
"mmlu_eval_accuracy_global_facts": 0.6,
"mmlu_eval_accuracy_high_school_biology": 0.34375,
"mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
"mmlu_eval_accuracy_high_school_computer_science": 0.7777777777777778,
"mmlu_eval_accuracy_high_school_european_history": 0.5,
"mmlu_eval_accuracy_high_school_geography": 0.7727272727272727,
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
"mmlu_eval_accuracy_high_school_macroeconomics": 0.32558139534883723,
"mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
"mmlu_eval_accuracy_high_school_microeconomics": 0.4230769230769231,
"mmlu_eval_accuracy_high_school_physics": 0.35294117647058826,
"mmlu_eval_accuracy_high_school_psychology": 0.7,
"mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
"mmlu_eval_accuracy_high_school_us_history": 0.6363636363636364,
"mmlu_eval_accuracy_high_school_world_history": 0.5,
"mmlu_eval_accuracy_human_aging": 0.6521739130434783,
"mmlu_eval_accuracy_human_sexuality": 0.5,
"mmlu_eval_accuracy_international_law": 0.7692307692307693,
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
"mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
"mmlu_eval_accuracy_management": 0.45454545454545453,
"mmlu_eval_accuracy_marketing": 0.76,
"mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
"mmlu_eval_accuracy_miscellaneous": 0.686046511627907,
"mmlu_eval_accuracy_moral_disputes": 0.47368421052631576,
"mmlu_eval_accuracy_moral_scenarios": 0.23,
"mmlu_eval_accuracy_nutrition": 0.6060606060606061,
"mmlu_eval_accuracy_philosophy": 0.4411764705882353,
"mmlu_eval_accuracy_prehistory": 0.4857142857142857,
"mmlu_eval_accuracy_professional_accounting": 0.2903225806451613,
"mmlu_eval_accuracy_professional_law": 0.3352941176470588,
"mmlu_eval_accuracy_professional_medicine": 0.3870967741935484,
"mmlu_eval_accuracy_professional_psychology": 0.36231884057971014,
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
"mmlu_eval_accuracy_security_studies": 0.5925925925925926,
"mmlu_eval_accuracy_sociology": 0.6818181818181818,
"mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
"mmlu_eval_accuracy_virology": 0.4444444444444444,
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
"mmlu_loss": 1.1704803334048148,
"step": 200
},
{
"epoch": 0.35,
"learning_rate": 0.0002,
"loss": 0.5413,
"step": 210
},
{
"epoch": 0.37,
"learning_rate": 0.0002,
"loss": 0.5632,
"step": 220
},
{
"epoch": 0.38,
"learning_rate": 0.0002,
"loss": 0.5752,
"step": 230
},
{
"epoch": 0.4,
"learning_rate": 0.0002,
"loss": 0.5583,
"step": 240
},
{
"epoch": 0.42,
"learning_rate": 0.0002,
"loss": 0.5274,
"step": 250
},
{
"epoch": 0.43,
"learning_rate": 0.0002,
"loss": 0.6163,
"step": 260
},
{
"epoch": 0.45,
"learning_rate": 0.0002,
"loss": 0.5673,
"step": 270
},
{
"epoch": 0.47,
"learning_rate": 0.0002,
"loss": 0.5241,
"step": 280
},
{
"epoch": 0.48,
"learning_rate": 0.0002,
"loss": 0.5564,
"step": 290
},
{
"epoch": 0.5,
"learning_rate": 0.0002,
"loss": 0.5451,
"step": 300
},
{
"epoch": 0.52,
"learning_rate": 0.0002,
"loss": 0.5661,
"step": 310
},
{
"epoch": 0.53,
"learning_rate": 0.0002,
"loss": 0.5467,
"step": 320
},
{
"epoch": 0.55,
"learning_rate": 0.0002,
"loss": 0.5513,
"step": 330
},
{
"epoch": 0.57,
"learning_rate": 0.0002,
"loss": 0.5999,
"step": 340
},
{
"epoch": 0.58,
"learning_rate": 0.0002,
"loss": 0.5518,
"step": 350
},
{
"epoch": 0.6,
"learning_rate": 0.0002,
"loss": 0.4896,
"step": 360
},
{
"epoch": 0.62,
"learning_rate": 0.0002,
"loss": 0.5314,
"step": 370
},
{
"epoch": 0.63,
"learning_rate": 0.0002,
"loss": 0.5495,
"step": 380
},
{
"epoch": 0.65,
"learning_rate": 0.0002,
"loss": 0.553,
"step": 390
},
{
"epoch": 0.67,
"learning_rate": 0.0002,
"loss": 0.5324,
"step": 400
},
{
"epoch": 0.67,
"eval_loss": 0.5613595843315125,
"eval_runtime": 210.7546,
"eval_samples_per_second": 4.745,
"eval_steps_per_second": 2.372,
"step": 400
},
{
"epoch": 0.67,
"mmlu_eval_accuracy": 0.47295524277138523,
"mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
"mmlu_eval_accuracy_anatomy": 0.5714285714285714,
"mmlu_eval_accuracy_astronomy": 0.4375,
"mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
"mmlu_eval_accuracy_clinical_knowledge": 0.4482758620689655,
"mmlu_eval_accuracy_college_biology": 0.375,
"mmlu_eval_accuracy_college_chemistry": 0.125,
"mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
"mmlu_eval_accuracy_college_mathematics": 0.2727272727272727,
"mmlu_eval_accuracy_college_medicine": 0.3181818181818182,
"mmlu_eval_accuracy_college_physics": 0.45454545454545453,
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
"mmlu_eval_accuracy_conceptual_physics": 0.34615384615384615,
"mmlu_eval_accuracy_econometrics": 0.16666666666666666,
"mmlu_eval_accuracy_electrical_engineering": 0.4375,
"mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
"mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
"mmlu_eval_accuracy_global_facts": 0.5,
"mmlu_eval_accuracy_high_school_biology": 0.34375,
"mmlu_eval_accuracy_high_school_chemistry": 0.36363636363636365,
"mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
"mmlu_eval_accuracy_high_school_european_history": 0.5555555555555556,
"mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
"mmlu_eval_accuracy_high_school_macroeconomics": 0.3488372093023256,
"mmlu_eval_accuracy_high_school_mathematics": 0.3448275862068966,
"mmlu_eval_accuracy_high_school_microeconomics": 0.38461538461538464,
"mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
"mmlu_eval_accuracy_high_school_psychology": 0.7166666666666667,
"mmlu_eval_accuracy_high_school_statistics": 0.391304347826087,
"mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
"mmlu_eval_accuracy_high_school_world_history": 0.5,
"mmlu_eval_accuracy_human_aging": 0.6956521739130435,
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
"mmlu_eval_accuracy_jurisprudence": 0.36363636363636365,
"mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
"mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
"mmlu_eval_accuracy_management": 0.6363636363636364,
"mmlu_eval_accuracy_marketing": 0.76,
"mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
"mmlu_eval_accuracy_miscellaneous": 0.6744186046511628,
"mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
"mmlu_eval_accuracy_moral_scenarios": 0.23,
"mmlu_eval_accuracy_nutrition": 0.6060606060606061,
"mmlu_eval_accuracy_philosophy": 0.47058823529411764,
"mmlu_eval_accuracy_prehistory": 0.4857142857142857,
"mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
"mmlu_eval_accuracy_professional_law": 0.35294117647058826,
"mmlu_eval_accuracy_professional_medicine": 0.41935483870967744,
"mmlu_eval_accuracy_professional_psychology": 0.37681159420289856,
"mmlu_eval_accuracy_public_relations": 0.5,
"mmlu_eval_accuracy_security_studies": 0.48148148148148145,
"mmlu_eval_accuracy_sociology": 0.6363636363636364,
"mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
"mmlu_eval_accuracy_virology": 0.4444444444444444,
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
"mmlu_loss": 1.1208655159366037,
"step": 400
},
{
"epoch": 0.69,
"learning_rate": 0.0002,
"loss": 0.5643,
"step": 410
},
{
"epoch": 0.7,
"learning_rate": 0.0002,
"loss": 0.5243,
"step": 420
},
{
"epoch": 0.72,
"learning_rate": 0.0002,
"loss": 0.5433,
"step": 430
},
{
"epoch": 0.74,
"learning_rate": 0.0002,
"loss": 0.5584,
"step": 440
},
{
"epoch": 0.75,
"learning_rate": 0.0002,
"loss": 0.5438,
"step": 450
},
{
"epoch": 0.77,
"learning_rate": 0.0002,
"loss": 0.5203,
"step": 460
},
{
"epoch": 0.79,
"learning_rate": 0.0002,
"loss": 0.5328,
"step": 470
},
{
"epoch": 0.8,
"learning_rate": 0.0002,
"loss": 0.5204,
"step": 480
},
{
"epoch": 0.82,
"learning_rate": 0.0002,
"loss": 0.5194,
"step": 490
},
{
"epoch": 0.84,
"learning_rate": 0.0002,
"loss": 0.5325,
"step": 500
},
{
"epoch": 0.85,
"learning_rate": 0.0002,
"loss": 0.5407,
"step": 510
},
{
"epoch": 0.87,
"learning_rate": 0.0002,
"loss": 0.4747,
"step": 520
},
{
"epoch": 0.89,
"learning_rate": 0.0002,
"loss": 0.5048,
"step": 530
},
{
"epoch": 0.9,
"learning_rate": 0.0002,
"loss": 0.5533,
"step": 540
},
{
"epoch": 0.92,
"learning_rate": 0.0002,
"loss": 0.5219,
"step": 550
},
{
"epoch": 0.94,
"learning_rate": 0.0002,
"loss": 0.5357,
"step": 560
},
{
"epoch": 0.95,
"learning_rate": 0.0002,
"loss": 0.5327,
"step": 570
},
{
"epoch": 0.97,
"learning_rate": 0.0002,
"loss": 0.492,
"step": 580
},
{
"epoch": 0.99,
"learning_rate": 0.0002,
"loss": 0.5312,
"step": 590
},
{
"epoch": 1.0,
"learning_rate": 0.0002,
"loss": 0.5238,
"step": 600
},
{
"epoch": 1.0,
"eval_loss": 0.5487588047981262,
"eval_runtime": 211.0529,
"eval_samples_per_second": 4.738,
"eval_steps_per_second": 2.369,
"step": 600
},
{
"epoch": 1.0,
"mmlu_eval_accuracy": 0.4657057894438228,
"mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
"mmlu_eval_accuracy_anatomy": 0.5,
"mmlu_eval_accuracy_astronomy": 0.4375,
"mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
"mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
"mmlu_eval_accuracy_college_biology": 0.375,
"mmlu_eval_accuracy_college_chemistry": 0.125,
"mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
"mmlu_eval_accuracy_college_medicine": 0.3181818181818182,
"mmlu_eval_accuracy_college_physics": 0.45454545454545453,
"mmlu_eval_accuracy_computer_security": 0.36363636363636365,
"mmlu_eval_accuracy_conceptual_physics": 0.3076923076923077,
"mmlu_eval_accuracy_econometrics": 0.16666666666666666,
"mmlu_eval_accuracy_electrical_engineering": 0.4375,
"mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
"mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
"mmlu_eval_accuracy_global_facts": 0.5,
"mmlu_eval_accuracy_high_school_biology": 0.34375,
"mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
"mmlu_eval_accuracy_high_school_computer_science": 0.6666666666666666,
"mmlu_eval_accuracy_high_school_european_history": 0.6666666666666666,
"mmlu_eval_accuracy_high_school_geography": 0.7727272727272727,
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
"mmlu_eval_accuracy_high_school_macroeconomics": 0.32558139534883723,
"mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276,
"mmlu_eval_accuracy_high_school_microeconomics": 0.38461538461538464,
"mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
"mmlu_eval_accuracy_high_school_psychology": 0.6666666666666666,
"mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
"mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
"mmlu_eval_accuracy_high_school_world_history": 0.46153846153846156,
"mmlu_eval_accuracy_human_aging": 0.6521739130434783,
"mmlu_eval_accuracy_human_sexuality": 0.4166666666666667,
"mmlu_eval_accuracy_international_law": 0.8461538461538461,
"mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
"mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
"mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
"mmlu_eval_accuracy_management": 0.7272727272727273,
"mmlu_eval_accuracy_marketing": 0.76,
"mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
"mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
"mmlu_eval_accuracy_moral_disputes": 0.39473684210526316,
"mmlu_eval_accuracy_moral_scenarios": 0.24,
"mmlu_eval_accuracy_nutrition": 0.5757575757575758,
"mmlu_eval_accuracy_philosophy": 0.5,
"mmlu_eval_accuracy_prehistory": 0.45714285714285713,
"mmlu_eval_accuracy_professional_accounting": 0.25806451612903225,
"mmlu_eval_accuracy_professional_law": 0.3352941176470588,
"mmlu_eval_accuracy_professional_medicine": 0.3870967741935484,
"mmlu_eval_accuracy_professional_psychology": 0.391304347826087,
"mmlu_eval_accuracy_public_relations": 0.5833333333333334,
"mmlu_eval_accuracy_security_studies": 0.5185185185185185,
"mmlu_eval_accuracy_sociology": 0.6818181818181818,
"mmlu_eval_accuracy_us_foreign_policy": 0.5454545454545454,
"mmlu_eval_accuracy_virology": 0.5,
"mmlu_eval_accuracy_world_religions": 0.7368421052631579,
"mmlu_loss": 1.2064378902744064,
"step": 600
},
{
"epoch": 1.02,
"learning_rate": 0.0002,
"loss": 0.4304,
"step": 610
},
{
"epoch": 1.04,
"learning_rate": 0.0002,
"loss": 0.466,
"step": 620
},
{
"epoch": 1.05,
"learning_rate": 0.0002,
"loss": 0.4934,
"step": 630
},
{
"epoch": 1.07,
"learning_rate": 0.0002,
"loss": 0.4605,
"step": 640
},
{
"epoch": 1.09,
"learning_rate": 0.0002,
"loss": 0.4799,
"step": 650
},
{
"epoch": 1.1,
"learning_rate": 0.0002,
"loss": 0.4572,
"step": 660
},
{
"epoch": 1.12,
"learning_rate": 0.0002,
"loss": 0.4398,
"step": 670
},
{
"epoch": 1.14,
"learning_rate": 0.0002,
"loss": 0.4673,
"step": 680
},
{
"epoch": 1.15,
"learning_rate": 0.0002,
"loss": 0.5008,
"step": 690
},
{
"epoch": 1.17,
"learning_rate": 0.0002,
"loss": 0.4832,
"step": 700
},
{
"epoch": 1.19,
"learning_rate": 0.0002,
"loss": 0.4558,
"step": 710
},
{
"epoch": 1.2,
"learning_rate": 0.0002,
"loss": 0.4661,
"step": 720
},
{
"epoch": 1.22,
"learning_rate": 0.0002,
"loss": 0.4483,
"step": 730
},
{
"epoch": 1.24,
"learning_rate": 0.0002,
"loss": 0.5047,
"step": 740
},
{
"epoch": 1.25,
"learning_rate": 0.0002,
"loss": 0.479,
"step": 750
},
{
"epoch": 1.27,
"learning_rate": 0.0002,
"loss": 0.4975,
"step": 760
},
{
"epoch": 1.29,
"learning_rate": 0.0002,
"loss": 0.4797,
"step": 770
},
{
"epoch": 1.3,
"learning_rate": 0.0002,
"loss": 0.4884,
"step": 780
},
{
"epoch": 1.32,
"learning_rate": 0.0002,
"loss": 0.4631,
"step": 790
},
{
"epoch": 1.34,
"learning_rate": 0.0002,
"loss": 0.4541,
"step": 800
},
{
"epoch": 1.34,
"eval_loss": 0.545791745185852,
"eval_runtime": 211.0641,
"eval_samples_per_second": 4.738,
"eval_steps_per_second": 2.369,
"step": 800
},
{
"epoch": 1.34,
"mmlu_eval_accuracy": 0.46218071540689526,
"mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
"mmlu_eval_accuracy_anatomy": 0.5,
"mmlu_eval_accuracy_astronomy": 0.375,
"mmlu_eval_accuracy_business_ethics": 0.6363636363636364,
"mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
"mmlu_eval_accuracy_college_biology": 0.4375,
"mmlu_eval_accuracy_college_chemistry": 0.125,
"mmlu_eval_accuracy_college_computer_science": 0.5454545454545454,
"mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
"mmlu_eval_accuracy_college_medicine": 0.3181818181818182,
"mmlu_eval_accuracy_college_physics": 0.45454545454545453,
"mmlu_eval_accuracy_computer_security": 0.2727272727272727,
"mmlu_eval_accuracy_conceptual_physics": 0.34615384615384615,
"mmlu_eval_accuracy_econometrics": 0.16666666666666666,
"mmlu_eval_accuracy_electrical_engineering": 0.375,
"mmlu_eval_accuracy_elementary_mathematics": 0.34146341463414637,
"mmlu_eval_accuracy_formal_logic": 0.2857142857142857,
"mmlu_eval_accuracy_global_facts": 0.5,
"mmlu_eval_accuracy_high_school_biology": 0.40625,
"mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
"mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
"mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112,
"mmlu_eval_accuracy_high_school_geography": 0.7272727272727273,
"mmlu_eval_accuracy_high_school_government_and_politics": 0.6190476190476191,
"mmlu_eval_accuracy_high_school_macroeconomics": 0.3488372093023256,
"mmlu_eval_accuracy_high_school_mathematics": 0.3103448275862069,
"mmlu_eval_accuracy_high_school_microeconomics": 0.38461538461538464,
"mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
"mmlu_eval_accuracy_high_school_psychology": 0.6833333333333333,
"mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
"mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818,
"mmlu_eval_accuracy_high_school_world_history": 0.5384615384615384,
"mmlu_eval_accuracy_human_aging": 0.6956521739130435,
"mmlu_eval_accuracy_human_sexuality": 0.5833333333333334,
"mmlu_eval_accuracy_international_law": 0.7692307692307693,
"mmlu_eval_accuracy_jurisprudence": 0.45454545454545453,
"mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
"mmlu_eval_accuracy_machine_learning": 0.18181818181818182,
"mmlu_eval_accuracy_management": 0.6363636363636364,
"mmlu_eval_accuracy_marketing": 0.72,
"mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
"mmlu_eval_accuracy_miscellaneous": 0.6627906976744186,
"mmlu_eval_accuracy_moral_disputes": 0.39473684210526316,
"mmlu_eval_accuracy_moral_scenarios": 0.23,
"mmlu_eval_accuracy_nutrition": 0.6363636363636364,
"mmlu_eval_accuracy_philosophy": 0.5294117647058824,
"mmlu_eval_accuracy_prehistory": 0.45714285714285713,
"mmlu_eval_accuracy_professional_accounting": 0.22580645161290322,
"mmlu_eval_accuracy_professional_law": 0.3588235294117647,
"mmlu_eval_accuracy_professional_medicine": 0.3548387096774194,
"mmlu_eval_accuracy_professional_psychology": 0.4057971014492754,
"mmlu_eval_accuracy_public_relations": 0.4166666666666667,
"mmlu_eval_accuracy_security_studies": 0.5555555555555556,
"mmlu_eval_accuracy_sociology": 0.5909090909090909,
"mmlu_eval_accuracy_us_foreign_policy": 0.5454545454545454,
"mmlu_eval_accuracy_virology": 0.4444444444444444,
"mmlu_eval_accuracy_world_religions": 0.7894736842105263,
"mmlu_loss": 1.1216171698610715,
"step": 800
}
],
"max_steps": 5000,
"num_train_epochs": 9,
"total_flos": 1.9806458895006106e+17,
"trial_name": null,
"trial_params": null
}