task_2_3_results.json

{
  "task": "2.3 - Library-Based Quantization with bitsandbytes",
  "model_source": "ranjan56cse/gpt2-large-agnews-classifier",
  "int8_metrics": {
    "model_name": "INT8 (bitsandbytes)",
    "quantization_method": "bitsandbytes",
    "num_samples_evaluated": 2000,
    "accuracy": 0.9505,
    "precision": 0.9503862443765202,
    "recall": 0.9505,
    "f1_score": 0.9504235034869967,
    "confusion_matrix": [
      [
        494,
        4,
        11,
        2
      ],
      [
        1,
        521,
        2,
        2
      ],
      [
        12,
        0,
        409,
        28
      ],
      [
        11,
        2,
        24,
        477
      ]
    ],
    "per_class_metrics": {
      "World": {
        "precision": 0.9536679536679536,
        "recall": 0.9667318982387475,
        "f1": 0.9601554907677357
      },
      "Sports": {
        "precision": 0.9886148007590133,
        "recall": 0.9904942965779467,
        "f1": 0.9895536562203229
      },
      "Business": {
        "precision": 0.9170403587443946,
        "recall": 0.910913140311804,
        "f1": 0.9139664804469274
      },
      "Sci/Tech": {
        "precision": 0.93713163064833,
        "recall": 0.9280155642023347,
        "f1": 0.9325513196480938
      }
    },
    "memory_mb": 801.35498046875,
    "avg_latency_ms": 75.78567934036255,
    "std_latency_ms": 31.118263723559952
  },
  "nf4_metrics": {
    "model_name": "NF4 (bitsandbytes)",
    "quantization_method": "bitsandbytes",
    "num_samples_evaluated": 2000,
    "accuracy": 0.9495,
    "precision": 0.9493676564802229,
    "recall": 0.9495,
    "f1_score": 0.9494144946434884,
    "confusion_matrix": [
      [
        494,
        4,
        12,
        1
      ],
      [
        1,
        521,
        2,
        2
      ],
      [
        12,
        0,
        407,
        30
      ],
      [
        11,
        2,
        24,
        477
      ]
    ],
    "per_class_metrics": {
      "World": {
        "precision": 0.9536679536679536,
        "recall": 0.9667318982387475,
        "f1": 0.9601554907677357
      },
      "Sports": {
        "precision": 0.9886148007590133,
        "recall": 0.9904942965779467,
        "f1": 0.9895536562203229
      },
      "Business": {
        "precision": 0.9146067415730337,
        "recall": 0.9064587973273942,
        "f1": 0.9105145413870246
      },
      "Sci/Tech": {
        "precision": 0.9352941176470588,
        "recall": 0.9280155642023347,
        "f1": 0.931640625
      }
    },
    "memory_mb": 463.8560791015625,
    "avg_latency_ms": 43.911072731018066,
    "std_latency_ms": 1.4153050965884408
  },
  "comparison": {
    "accuracy_diff": 0.0010000000000000009,
    "memory_reduction_pct": 42.11602967386172
  }
}