checkpoint-200/trainer_state.json

{
 "best_metric": 0.8394724130630493,
 "best_model_checkpoint": "./output_v2/7b_cluster015_Nous-Hermes-llama-2-7b_partitioned_v3_standardized_015/checkpoint-200",
 "epoch": 0.1340931947703654,
 "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": 0.969,
 "step": 10
 },
 {
 "epoch": 0.01,
 "learning_rate": 0.0002,
 "loss": 0.9032,
 "step": 20
 },
 {
 "epoch": 0.02,
 "learning_rate": 0.0002,
 "loss": 0.81,
 "step": 30
 },
 {
 "epoch": 0.03,
 "learning_rate": 0.0002,
 "loss": 0.876,
 "step": 40
 },
 {
 "epoch": 0.03,
 "learning_rate": 0.0002,
 "loss": 0.8858,
 "step": 50
 },
 {
 "epoch": 0.04,
 "learning_rate": 0.0002,
 "loss": 0.8608,
 "step": 60
 },
 {
 "epoch": 0.05,
 "learning_rate": 0.0002,
 "loss": 0.846,
 "step": 70
 },
 {
 "epoch": 0.05,
 "learning_rate": 0.0002,
 "loss": 0.8466,
 "step": 80
 },
 {
 "epoch": 0.06,
 "learning_rate": 0.0002,
 "loss": 0.8456,
 "step": 90
 },
 {
 "epoch": 0.07,
 "learning_rate": 0.0002,
 "loss": 0.8895,
 "step": 100
 },
 {
 "epoch": 0.07,
 "learning_rate": 0.0002,
 "loss": 0.862,
 "step": 110
 },
 {
 "epoch": 0.08,
 "learning_rate": 0.0002,
 "loss": 0.8193,
 "step": 120
 },
 {
 "epoch": 0.09,
 "learning_rate": 0.0002,
 "loss": 0.8588,
 "step": 130
 },
 {
 "epoch": 0.09,
 "learning_rate": 0.0002,
 "loss": 0.8516,
 "step": 140
 },
 {
 "epoch": 0.1,
 "learning_rate": 0.0002,
 "loss": 0.8428,
 "step": 150
 },
 {
 "epoch": 0.11,
 "learning_rate": 0.0002,
 "loss": 0.8829,
 "step": 160
 },
 {
 "epoch": 0.11,
 "learning_rate": 0.0002,
 "loss": 0.882,
 "step": 170
 },
 {
 "epoch": 0.12,
 "learning_rate": 0.0002,
 "loss": 0.8054,
 "step": 180
 },
 {
 "epoch": 0.13,
 "learning_rate": 0.0002,
 "loss": 0.8673,
 "step": 190
 },
 {
 "epoch": 0.13,
 "learning_rate": 0.0002,
 "loss": 0.8389,
 "step": 200
 },
 {
 "epoch": 0.13,
 "eval_loss": 0.8394724130630493,
 "eval_runtime": 191.2112,
 "eval_samples_per_second": 5.23,
 "eval_steps_per_second": 2.615,
 "step": 200
 },
 {
 "epoch": 0.13,
 "mmlu_eval_accuracy": 0.4626311671628311,
 "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
 "mmlu_eval_accuracy_anatomy": 0.5,
 "mmlu_eval_accuracy_astronomy": 0.375,
 "mmlu_eval_accuracy_business_ethics": 0.5454545454545454,
 "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": 0.5454545454545454,
 "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": 0.36363636363636365,
 "mmlu_eval_accuracy_conceptual_physics": 0.46153846153846156,
 "mmlu_eval_accuracy_econometrics": 0.16666666666666666,
 "mmlu_eval_accuracy_electrical_engineering": 0.5,
 "mmlu_eval_accuracy_elementary_mathematics": 0.3902439024390244,
 "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": 0.7777777777777778,
 "mmlu_eval_accuracy_high_school_european_history": 0.5555555555555556,
 "mmlu_eval_accuracy_high_school_geography": 0.6818181818181818,
 "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666,
 "mmlu_eval_accuracy_high_school_macroeconomics": 0.3023255813953488,
 "mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724,
 "mmlu_eval_accuracy_high_school_microeconomics": 0.34615384615384615,
 "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354,
 "mmlu_eval_accuracy_high_school_psychology": 0.7166666666666667,
 "mmlu_eval_accuracy_high_school_statistics": 0.30434782608695654,
 "mmlu_eval_accuracy_high_school_us_history": 0.7272727272727273,
 "mmlu_eval_accuracy_high_school_world_history": 0.5,
 "mmlu_eval_accuracy_human_aging": 0.6956521739130435,
 "mmlu_eval_accuracy_human_sexuality": 0.3333333333333333,
 "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": 0.18181818181818182,
 "mmlu_eval_accuracy_management": 0.6363636363636364,
 "mmlu_eval_accuracy_marketing": 0.72,
 "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273,
 "mmlu_eval_accuracy_miscellaneous": 0.6976744186046512,
 "mmlu_eval_accuracy_moral_disputes": 0.5,
 "mmlu_eval_accuracy_moral_scenarios": 0.24,
 "mmlu_eval_accuracy_nutrition": 0.5454545454545454,
 "mmlu_eval_accuracy_philosophy": 0.5294117647058824,
 "mmlu_eval_accuracy_prehistory": 0.42857142857142855,
 "mmlu_eval_accuracy_professional_accounting": 0.2903225806451613,
 "mmlu_eval_accuracy_professional_law": 0.3235294117647059,
 "mmlu_eval_accuracy_professional_medicine": 0.3870967741935484,
 "mmlu_eval_accuracy_professional_psychology": 0.4057971014492754,
 "mmlu_eval_accuracy_public_relations": 0.5833333333333334,
 "mmlu_eval_accuracy_security_studies": 0.5185185185185185,
 "mmlu_eval_accuracy_sociology": 0.6363636363636364,
 "mmlu_eval_accuracy_us_foreign_policy": 0.5454545454545454,
 "mmlu_eval_accuracy_virology": 0.4444444444444444,
 "mmlu_eval_accuracy_world_religions": 0.7368421052631579,
 "mmlu_loss": 1.0846105751924353,
 "step": 200
 }
 ],
 "max_steps": 5000,
 "num_train_epochs": 4,
 "total_flos": 4.435266218124902e+16,
 "trial_name": null,
 "trial_params": null
}