checkpoint-200/trainer_state.json

{
 "best_metric": 0.6043372750282288,
 "best_model_checkpoint": "./output_v2/7b_cluster021_Nous-Hermes-llama-2-7b_partitioned_v3_standardized_021/checkpoint-200",
 "epoch": 0.20392556716798368,
 "global_step": 200,
 "is_hyper_param_search": false,
 "is_local_process_zero": true,
 "is_world_process_zero": true,
 "log_history": [
 {
 "epoch": 0.01,
 "learning_rate": 0.0002,
 "loss": 0.6669,
 "step": 10
 },
 {
 "epoch": 0.02,
 "learning_rate": 0.0002,
 "loss": 0.6301,
 "step": 20
 },
 {
 "epoch": 0.03,
 "learning_rate": 0.0002,
 "loss": 0.7049,
 "step": 30
 },
 {
 "epoch": 0.04,
 "learning_rate": 0.0002,
 "loss": 0.676,
 "step": 40
 },
 {
 "epoch": 0.05,
 "learning_rate": 0.0002,
 "loss": 0.6972,
 "step": 50
 },
 {
 "epoch": 0.06,
 "learning_rate": 0.0002,
 "loss": 0.7754,
 "step": 60
 },
 {
 "epoch": 0.07,
 "learning_rate": 0.0002,
 "loss": 0.7056,
 "step": 70
 },
 {
 "epoch": 0.08,
 "learning_rate": 0.0002,
 "loss": 0.7018,
 "step": 80
 },
 {
 "epoch": 0.09,
 "learning_rate": 0.0002,
 "loss": 0.6906,
 "step": 90
 },
 {
 "epoch": 0.1,
 "learning_rate": 0.0002,
 "loss": 0.6263,
 "step": 100
 },
 {
 "epoch": 0.11,
 "learning_rate": 0.0002,
 "loss": 0.6163,
 "step": 110
 },
 {
 "epoch": 0.12,
 "learning_rate": 0.0002,
 "loss": 0.6824,
 "step": 120
 },
 {
 "epoch": 0.13,
 "learning_rate": 0.0002,
 "loss": 0.5913,
 "step": 130
 },
 {
 "epoch": 0.14,
 "learning_rate": 0.0002,
 "loss": 0.5903,
 "step": 140
 },
 {
 "epoch": 0.15,
 "learning_rate": 0.0002,
 "loss": 0.6851,
 "step": 150
 },
 {
 "epoch": 0.16,
 "learning_rate": 0.0002,
 "loss": 0.5648,
 "step": 160
 },
 {
 "epoch": 0.17,
 "learning_rate": 0.0002,
 "loss": 0.5615,
 "step": 170
 },
 {
 "epoch": 0.18,
 "learning_rate": 0.0002,
 "loss": 0.5532,
 "step": 180
 },
 {
 "epoch": 0.19,
 "learning_rate": 0.0002,
 "loss": 0.5662,
 "step": 190
 },
 {
 "epoch": 0.2,
 "learning_rate": 0.0002,
 "loss": 0.6302,
 "step": 200
 },
 {
 "epoch": 0.2,
 "eval_loss": 0.6043372750282288,
 "eval_runtime": 155.7668,
 "eval_samples_per_second": 6.42,
 "eval_steps_per_second": 3.21,
 "step": 200
 },
 {
 "epoch": 0.2,
 "mmlu_eval_accuracy": 0.45742535599141315,
 "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727,
 "mmlu_eval_accuracy_anatomy": 0.6428571428571429,
 "mmlu_eval_accuracy_astronomy": 0.375,
 "mmlu_eval_accuracy_business_ethics": 0.45454545454545453,
 "mmlu_eval_accuracy_clinical_knowledge": 0.41379310344827586,
 "mmlu_eval_accuracy_college_biology": 0.4375,
 "mmlu_eval_accuracy_college_chemistry": 0.25,
 "mmlu_eval_accuracy_college_computer_science": 0.45454545454545453,
 "mmlu_eval_accuracy_college_mathematics": 0.18181818181818182,
 "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": 0.4230769230769231,
 "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
 "mmlu_eval_accuracy_electrical_engineering": 0.375,
 "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536,
 "mmlu_eval_accuracy_formal_logic": 0.21428571428571427,
 "mmlu_eval_accuracy_global_facts": 0.3,
 "mmlu_eval_accuracy_high_school_biology": 0.375,
 "mmlu_eval_accuracy_high_school_chemistry": 0.4090909090909091,
 "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556,
 "mmlu_eval_accuracy_high_school_european_history": 0.5,
 "mmlu_eval_accuracy_high_school_geography": 0.8181818181818182,
 "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
 "mmlu_eval_accuracy_high_school_macroeconomics": 0.32558139534883723,
 "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": 0.7333333333333333,
 "mmlu_eval_accuracy_high_school_statistics": 0.2608695652173913,
 "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
 "mmlu_eval_accuracy_high_school_world_history": 0.5384615384615384,
 "mmlu_eval_accuracy_human_aging": 0.7391304347826086,
 "mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
 "mmlu_eval_accuracy_international_law": 0.8461538461538461,
 "mmlu_eval_accuracy_jurisprudence": 0.2727272727272727,
 "mmlu_eval_accuracy_logical_fallacies": 0.5555555555555556,
 "mmlu_eval_accuracy_machine_learning": 0.2727272727272727,
 "mmlu_eval_accuracy_management": 0.5454545454545454,
 "mmlu_eval_accuracy_marketing": 0.68,
 "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
 "mmlu_eval_accuracy_miscellaneous": 0.6511627906976745,
 "mmlu_eval_accuracy_moral_disputes": 0.42105263157894735,
 "mmlu_eval_accuracy_moral_scenarios": 0.24,
 "mmlu_eval_accuracy_nutrition": 0.48484848484848486,
 "mmlu_eval_accuracy_philosophy": 0.5,
 "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": 0.41935483870967744,
 "mmlu_eval_accuracy_professional_psychology": 0.4057971014492754,
 "mmlu_eval_accuracy_public_relations": 0.6666666666666666,
 "mmlu_eval_accuracy_security_studies": 0.48148148148148145,
 "mmlu_eval_accuracy_sociology": 0.5909090909090909,
 "mmlu_eval_accuracy_us_foreign_policy": 0.6363636363636364,
 "mmlu_eval_accuracy_virology": 0.5,
 "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
 "mmlu_loss": 0.8281081981918053,
 "step": 200
 }
 ],
 "max_steps": 5000,
 "num_train_epochs": 6,
 "total_flos": 3.62797173793751e+16,
 "trial_name": null,
 "trial_params": null
}