miqu-70b-sft / checkpoint-200 /trainer_state.json
anchen1011's picture
Upload folder using huggingface_hub
f0f7baf verified
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 0.36272953978689637,
"global_step": 200,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02,
"learning_rate": 0.0001,
"loss": 1.1655,
"step": 10
},
{
"epoch": 0.04,
"learning_rate": 0.0001,
"loss": 1.001,
"step": 20
},
{
"epoch": 0.05,
"learning_rate": 0.0001,
"loss": 1.0287,
"step": 30
},
{
"epoch": 0.07,
"learning_rate": 0.0001,
"loss": 1.1578,
"step": 40
},
{
"epoch": 0.09,
"learning_rate": 0.0001,
"loss": 1.2146,
"step": 50
},
{
"epoch": 0.11,
"learning_rate": 0.0001,
"loss": 0.997,
"step": 60
},
{
"epoch": 0.13,
"learning_rate": 0.0001,
"loss": 0.9024,
"step": 70
},
{
"epoch": 0.15,
"learning_rate": 0.0001,
"loss": 0.9901,
"step": 80
},
{
"epoch": 0.16,
"learning_rate": 0.0001,
"loss": 1.1264,
"step": 90
},
{
"epoch": 0.18,
"learning_rate": 0.0001,
"loss": 1.2038,
"step": 100
},
{
"epoch": 0.2,
"learning_rate": 0.0001,
"loss": 0.8935,
"step": 110
},
{
"epoch": 0.22,
"learning_rate": 0.0001,
"loss": 0.9178,
"step": 120
},
{
"epoch": 0.24,
"learning_rate": 0.0001,
"loss": 0.9746,
"step": 130
},
{
"epoch": 0.25,
"learning_rate": 0.0001,
"loss": 1.1566,
"step": 140
},
{
"epoch": 0.27,
"learning_rate": 0.0001,
"loss": 1.2877,
"step": 150
},
{
"epoch": 0.29,
"learning_rate": 0.0001,
"loss": 0.9146,
"step": 160
},
{
"epoch": 0.31,
"learning_rate": 0.0001,
"loss": 0.8895,
"step": 170
},
{
"epoch": 0.33,
"learning_rate": 0.0001,
"loss": 1.0121,
"step": 180
},
{
"epoch": 0.34,
"eval_loss": 1.0215636491775513,
"eval_runtime": 950.138,
"eval_samples_per_second": 1.052,
"eval_steps_per_second": 1.052,
"step": 187
},
{
"epoch": 0.34,
"mmlu_eval_accuracy": 0.731892294851104,
"mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182,
"mmlu_eval_accuracy_anatomy": 0.7857142857142857,
"mmlu_eval_accuracy_astronomy": 0.6875,
"mmlu_eval_accuracy_business_ethics": 0.7272727272727273,
"mmlu_eval_accuracy_clinical_knowledge": 0.896551724137931,
"mmlu_eval_accuracy_college_biology": 0.875,
"mmlu_eval_accuracy_college_chemistry": 0.5,
"mmlu_eval_accuracy_college_computer_science": 0.7272727272727273,
"mmlu_eval_accuracy_college_mathematics": 0.45454545454545453,
"mmlu_eval_accuracy_college_medicine": 0.9090909090909091,
"mmlu_eval_accuracy_college_physics": 0.6363636363636364,
"mmlu_eval_accuracy_computer_security": 0.6363636363636364,
"mmlu_eval_accuracy_conceptual_physics": 0.6538461538461539,
"mmlu_eval_accuracy_econometrics": 0.8333333333333334,
"mmlu_eval_accuracy_electrical_engineering": 0.8125,
"mmlu_eval_accuracy_elementary_mathematics": 0.7073170731707317,
"mmlu_eval_accuracy_formal_logic": 0.5714285714285714,
"mmlu_eval_accuracy_global_facts": 0.5,
"mmlu_eval_accuracy_high_school_biology": 0.8125,
"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": 0.9090909090909091,
"mmlu_eval_accuracy_high_school_government_and_politics": 0.9523809523809523,
"mmlu_eval_accuracy_high_school_macroeconomics": 0.7441860465116279,
"mmlu_eval_accuracy_high_school_mathematics": 0.3793103448275862,
"mmlu_eval_accuracy_high_school_microeconomics": 0.9615384615384616,
"mmlu_eval_accuracy_high_school_physics": 0.23529411764705882,
"mmlu_eval_accuracy_high_school_psychology": 0.9333333333333333,
"mmlu_eval_accuracy_high_school_statistics": 0.6956521739130435,
"mmlu_eval_accuracy_high_school_us_history": 0.9090909090909091,
"mmlu_eval_accuracy_high_school_world_history": 0.8076923076923077,
"mmlu_eval_accuracy_human_aging": 0.8260869565217391,
"mmlu_eval_accuracy_human_sexuality": 0.6666666666666666,
"mmlu_eval_accuracy_international_law": 1.0,
"mmlu_eval_accuracy_jurisprudence": 0.6363636363636364,
"mmlu_eval_accuracy_logical_fallacies": 0.7777777777777778,
"mmlu_eval_accuracy_machine_learning": 0.5454545454545454,
"mmlu_eval_accuracy_management": 0.9090909090909091,
"mmlu_eval_accuracy_marketing": 0.88,
"mmlu_eval_accuracy_medical_genetics": 0.9090909090909091,
"mmlu_eval_accuracy_miscellaneous": 0.7906976744186046,
"mmlu_eval_accuracy_moral_disputes": 0.8157894736842105,
"mmlu_eval_accuracy_moral_scenarios": 0.59,
"mmlu_eval_accuracy_nutrition": 0.7878787878787878,
"mmlu_eval_accuracy_philosophy": 0.7941176470588235,
"mmlu_eval_accuracy_prehistory": 0.8285714285714286,
"mmlu_eval_accuracy_professional_accounting": 0.6451612903225806,
"mmlu_eval_accuracy_professional_law": 0.6294117647058823,
"mmlu_eval_accuracy_professional_medicine": 0.8064516129032258,
"mmlu_eval_accuracy_professional_psychology": 0.8115942028985508,
"mmlu_eval_accuracy_public_relations": 0.6666666666666666,
"mmlu_eval_accuracy_security_studies": 0.8148148148148148,
"mmlu_eval_accuracy_sociology": 0.9545454545454546,
"mmlu_eval_accuracy_us_foreign_policy": 1.0,
"mmlu_eval_accuracy_virology": 0.5,
"mmlu_eval_accuracy_world_religions": 0.8421052631578947,
"mmlu_loss": 1.326305795171384,
"step": 187
},
{
"epoch": 0.34,
"learning_rate": 0.0001,
"loss": 1.1133,
"step": 190
},
{
"epoch": 0.36,
"learning_rate": 0.0001,
"loss": 1.2485,
"step": 200
}
],
"max_steps": 1875,
"num_train_epochs": 4,
"total_flos": 2.3241435438071808e+17,
"trial_name": null,
"trial_params": null
}