Nous-Hermes-llama-2-7b_7b_cluster015_partitioned_v3_standardized_015
/
checkpoint-400
/trainer_state.json
| { | |
| "best_metric": 0.8304864764213562, | |
| "best_model_checkpoint": "./output_v2/7b_cluster015_Nous-Hermes-llama-2-7b_partitioned_v3_standardized_015/checkpoint-400", | |
| "epoch": 0.2681863895407308, | |
| "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": 0.969, | |
| "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 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 0.0002, | |
| "loss": 0.9027, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8621, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8405, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8553, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8334, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8791, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8607, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8403, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8471, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8945, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8094, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8571, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8469, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8609, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8242, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8679, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8583, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8815, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 0.0002, | |
| "loss": 0.819, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8946, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "eval_loss": 0.8304864764213562, | |
| "eval_runtime": 191.0197, | |
| "eval_samples_per_second": 5.235, | |
| "eval_steps_per_second": 2.618, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "mmlu_eval_accuracy": 0.45184631712481954, | |
| "mmlu_eval_accuracy_abstract_algebra": 0.18181818181818182, | |
| "mmlu_eval_accuracy_anatomy": 0.5, | |
| "mmlu_eval_accuracy_astronomy": 0.4375, | |
| "mmlu_eval_accuracy_business_ethics": 0.45454545454545453, | |
| "mmlu_eval_accuracy_clinical_knowledge": 0.3793103448275862, | |
| "mmlu_eval_accuracy_college_biology": 0.4375, | |
| "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": 0.46153846153846156, | |
| "mmlu_eval_accuracy_econometrics": 0.16666666666666666, | |
| "mmlu_eval_accuracy_electrical_engineering": 0.5, | |
| "mmlu_eval_accuracy_elementary_mathematics": 0.36585365853658536, | |
| "mmlu_eval_accuracy_formal_logic": 0.2857142857142857, | |
| "mmlu_eval_accuracy_global_facts": 0.4, | |
| "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": 0.5555555555555556, | |
| "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": 0.38461538461538464, | |
| "mmlu_eval_accuracy_high_school_physics": 0.29411764705882354, | |
| "mmlu_eval_accuracy_high_school_psychology": 0.7, | |
| "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": 0.7391304347826086, | |
| "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": 0.6111111111111112, | |
| "mmlu_eval_accuracy_machine_learning": 0.18181818181818182, | |
| "mmlu_eval_accuracy_management": 0.45454545454545453, | |
| "mmlu_eval_accuracy_marketing": 0.72, | |
| "mmlu_eval_accuracy_medical_genetics": 0.7272727272727273, | |
| "mmlu_eval_accuracy_miscellaneous": 0.686046511627907, | |
| "mmlu_eval_accuracy_moral_disputes": 0.42105263157894735, | |
| "mmlu_eval_accuracy_moral_scenarios": 0.24, | |
| "mmlu_eval_accuracy_nutrition": 0.6060606060606061, | |
| "mmlu_eval_accuracy_philosophy": 0.5588235294117647, | |
| "mmlu_eval_accuracy_prehistory": 0.4, | |
| "mmlu_eval_accuracy_professional_accounting": 0.25806451612903225, | |
| "mmlu_eval_accuracy_professional_law": 0.3176470588235294, | |
| "mmlu_eval_accuracy_professional_medicine": 0.41935483870967744, | |
| "mmlu_eval_accuracy_professional_psychology": 0.391304347826087, | |
| "mmlu_eval_accuracy_public_relations": 0.5, | |
| "mmlu_eval_accuracy_security_studies": 0.5185185185185185, | |
| "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": 0.6842105263157895, | |
| "mmlu_loss": 1.1180029823286726, | |
| "step": 400 | |
| } | |
| ], | |
| "max_steps": 5000, | |
| "num_train_epochs": 4, | |
| "total_flos": 8.772803302765363e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |