Nous-Hermes-llama-2-7b_7b_cluster028_partitioned_v3_standardized_028
/
checkpoint-400
/trainer_state.json
| { | |
| "best_metric": 0.6207628846168518, | |
| "best_model_checkpoint": "./output_v2/7b_cluster028_Nous-Hermes-llama-2-7b_partitioned_v3_standardized_028/checkpoint-400", | |
| "epoch": 0.4062976130015236, | |
| "global_step": 400, | |
| "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": 0.8709, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 0.0002, | |
| "loss": 0.9602, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7606, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7942, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 0.0002, | |
| "loss": 0.5849, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7161, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7699, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7264, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6845, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6638, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6089, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7681, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7489, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7472, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8521, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7223, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6727, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6434, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6754, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6945, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_loss": 0.6316617727279663, | |
| "eval_runtime": 120.7896, | |
| "eval_samples_per_second": 8.279, | |
| "eval_steps_per_second": 4.139, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "mmlu_eval_accuracy": 0.46933615423997516, | |
| "mmlu_eval_accuracy_abstract_algebra": 0.2727272727272727, | |
| "mmlu_eval_accuracy_anatomy": 0.5714285714285714, | |
| "mmlu_eval_accuracy_astronomy": 0.375, | |
| "mmlu_eval_accuracy_business_ethics": 0.5454545454545454, | |
| "mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552, | |
| "mmlu_eval_accuracy_college_biology": 0.4375, | |
| "mmlu_eval_accuracy_college_chemistry": 0.25, | |
| "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365, | |
| "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365, | |
| "mmlu_eval_accuracy_college_medicine": 0.4090909090909091, | |
| "mmlu_eval_accuracy_college_physics": 0.5454545454545454, | |
| "mmlu_eval_accuracy_computer_security": 0.36363636363636365, | |
| "mmlu_eval_accuracy_conceptual_physics": 0.38461538461538464, | |
| "mmlu_eval_accuracy_econometrics": 0.16666666666666666, | |
| "mmlu_eval_accuracy_electrical_engineering": 0.375, | |
| "mmlu_eval_accuracy_elementary_mathematics": 0.24390243902439024, | |
| "mmlu_eval_accuracy_formal_logic": 0.35714285714285715, | |
| "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": 0.4444444444444444, | |
| "mmlu_eval_accuracy_high_school_european_history": 0.6111111111111112, | |
| "mmlu_eval_accuracy_high_school_geography": 0.6818181818181818, | |
| "mmlu_eval_accuracy_high_school_government_and_politics": 0.6666666666666666, | |
| "mmlu_eval_accuracy_high_school_macroeconomics": 0.4186046511627907, | |
| "mmlu_eval_accuracy_high_school_mathematics": 0.27586206896551724, | |
| "mmlu_eval_accuracy_high_school_microeconomics": 0.4230769230769231, | |
| "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354, | |
| "mmlu_eval_accuracy_high_school_psychology": 0.7, | |
| "mmlu_eval_accuracy_high_school_statistics": 0.34782608695652173, | |
| "mmlu_eval_accuracy_high_school_us_history": 0.6818181818181818, | |
| "mmlu_eval_accuracy_high_school_world_history": 0.5, | |
| "mmlu_eval_accuracy_human_aging": 0.6956521739130435, | |
| "mmlu_eval_accuracy_human_sexuality": 0.5, | |
| "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": 0.2727272727272727, | |
| "mmlu_eval_accuracy_management": 0.5454545454545454, | |
| "mmlu_eval_accuracy_marketing": 0.72, | |
| "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273, | |
| "mmlu_eval_accuracy_miscellaneous": 0.6511627906976745, | |
| "mmlu_eval_accuracy_moral_disputes": 0.47368421052631576, | |
| "mmlu_eval_accuracy_moral_scenarios": 0.23, | |
| "mmlu_eval_accuracy_nutrition": 0.6060606060606061, | |
| "mmlu_eval_accuracy_philosophy": 0.47058823529411764, | |
| "mmlu_eval_accuracy_prehistory": 0.45714285714285713, | |
| "mmlu_eval_accuracy_professional_accounting": 0.2903225806451613, | |
| "mmlu_eval_accuracy_professional_law": 0.3411764705882353, | |
| "mmlu_eval_accuracy_professional_medicine": 0.45161290322580644, | |
| "mmlu_eval_accuracy_professional_psychology": 0.3188405797101449, | |
| "mmlu_eval_accuracy_public_relations": 0.5, | |
| "mmlu_eval_accuracy_security_studies": 0.48148148148148145, | |
| "mmlu_eval_accuracy_sociology": 0.6818181818181818, | |
| "mmlu_eval_accuracy_us_foreign_policy": 0.7272727272727273, | |
| "mmlu_eval_accuracy_virology": 0.3333333333333333, | |
| "mmlu_eval_accuracy_world_religions": 0.7368421052631579, | |
| "mmlu_loss": 0.9820772503398106, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6532, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7207, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7092, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6561, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6516, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6293, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6238, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6484, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6795, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 0.0002, | |
| "loss": 0.5931, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7188, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6823, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7286, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7396, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6779, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7003, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6721, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 0.0002, | |
| "loss": 0.736, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6221, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6736, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "eval_loss": 0.6207628846168518, | |
| "eval_runtime": 120.8451, | |
| "eval_samples_per_second": 8.275, | |
| "eval_steps_per_second": 4.138, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "mmlu_eval_accuracy": 0.4837331454649875, | |
| "mmlu_eval_accuracy_abstract_algebra": 0.36363636363636365, | |
| "mmlu_eval_accuracy_anatomy": 0.7142857142857143, | |
| "mmlu_eval_accuracy_astronomy": 0.3125, | |
| "mmlu_eval_accuracy_business_ethics": 0.5454545454545454, | |
| "mmlu_eval_accuracy_clinical_knowledge": 0.4827586206896552, | |
| "mmlu_eval_accuracy_college_biology": 0.5625, | |
| "mmlu_eval_accuracy_college_chemistry": 0.125, | |
| "mmlu_eval_accuracy_college_computer_science": 0.36363636363636365, | |
| "mmlu_eval_accuracy_college_mathematics": 0.36363636363636365, | |
| "mmlu_eval_accuracy_college_medicine": 0.45454545454545453, | |
| "mmlu_eval_accuracy_college_physics": 0.5454545454545454, | |
| "mmlu_eval_accuracy_computer_security": 0.5454545454545454, | |
| "mmlu_eval_accuracy_conceptual_physics": 0.38461538461538464, | |
| "mmlu_eval_accuracy_econometrics": 0.08333333333333333, | |
| "mmlu_eval_accuracy_electrical_engineering": 0.375, | |
| "mmlu_eval_accuracy_elementary_mathematics": 0.24390243902439024, | |
| "mmlu_eval_accuracy_formal_logic": 0.42857142857142855, | |
| "mmlu_eval_accuracy_global_facts": 0.5, | |
| "mmlu_eval_accuracy_high_school_biology": 0.375, | |
| "mmlu_eval_accuracy_high_school_chemistry": 0.3181818181818182, | |
| "mmlu_eval_accuracy_high_school_computer_science": 0.5555555555555556, | |
| "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": 0.37209302325581395, | |
| "mmlu_eval_accuracy_high_school_mathematics": 0.2413793103448276, | |
| "mmlu_eval_accuracy_high_school_microeconomics": 0.38461538461538464, | |
| "mmlu_eval_accuracy_high_school_physics": 0.4117647058823529, | |
| "mmlu_eval_accuracy_high_school_psychology": 0.7666666666666667, | |
| "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": 0.6521739130434783, | |
| "mmlu_eval_accuracy_human_sexuality": 0.5, | |
| "mmlu_eval_accuracy_international_law": 0.7692307692307693, | |
| "mmlu_eval_accuracy_jurisprudence": 0.5454545454545454, | |
| "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": 0.6627906976744186, | |
| "mmlu_eval_accuracy_moral_disputes": 0.5263157894736842, | |
| "mmlu_eval_accuracy_moral_scenarios": 0.25, | |
| "mmlu_eval_accuracy_nutrition": 0.5757575757575758, | |
| "mmlu_eval_accuracy_philosophy": 0.5, | |
| "mmlu_eval_accuracy_prehistory": 0.4857142857142857, | |
| "mmlu_eval_accuracy_professional_accounting": 0.2903225806451613, | |
| "mmlu_eval_accuracy_professional_law": 0.3235294117647059, | |
| "mmlu_eval_accuracy_professional_medicine": 0.45161290322580644, | |
| "mmlu_eval_accuracy_professional_psychology": 0.37681159420289856, | |
| "mmlu_eval_accuracy_public_relations": 0.4166666666666667, | |
| "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": 0.3333333333333333, | |
| "mmlu_eval_accuracy_world_religions": 0.7368421052631579, | |
| "mmlu_loss": 1.0501697772321128, | |
| "step": 400 | |
| } | |
| ], | |
| "max_steps": 5000, | |
| "num_train_epochs": 6, | |
| "total_flos": 5.044620663609754e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |