eval/benchmark_summary.json

{
  "release": "IrishCore-DiffMask-135M-v1-rc4",
  "repo_id": "temsa/IrishCore-DiffMask-135M-v1-rc4",
  "architecture": {
    "family": "DistilBERT-size token-span extractor",
    "diffusion_style_training": true,
    "runtime_diffusion": false,
    "scanner_free": true,
    "validator_free": true,
    "heads": [
      "token_presence_head",
      "typed_start_boundary_head",
      "typed_end_boundary_head"
    ]
  },
  "base_model": "OpenMed/OpenMed-PII-mLiteClinical-Base-135M-v1",
  "notes": [
    "DiffMask uses a masked denoising training schedule, not a generative diffusion runtime.",
    "ONNX q8 is the recommended CPU deployment artifact.",
    "The release inference scripts emit [PII:LABEL] placeholders.",
    "rc4 selection added dedicated QA feedback and UAT replay suites for mixed PII, spaced IBAN, hyphen/local phones, and weak-context PPSN.",
    "The strict-boundary QA feedback score is reported from the full-checkpoint path with IoU 1.0 to avoid overstating partial-span success."
  ],
  "full": {
    "core_f1": 0.9733333333333333,
    "edge_f1": null,
    "multilingual_f1": null,
    "hardening_f1": null,
    "qa_feedback_exact_strict_boundary_f1": 0.6875
  },
  "onnx_q8": {
    "core_f1": 0.9733333333333333,
    "edge_f1": 0.9500000000000001,
    "finance_f1": 1.0,
    "finance_boundary_f1": 1.0,
    "user_ppsn_f1": null,
    "gaelic_weak_ppsn_f1": 1.0,
    "multilingual_f1": 0.937142857142857,
    "hardening_f1": 1.0,
    "core_examples_per_second": 29.466585672913574,
    "multilingual_examples_per_second": 38.41988858495913,
    "runtime_profile_examples_per_second": 158.49428243560962,
    "qa_feedback_exact_f1": 0.875,
    "uat_replay_exact_f1": 0.9032258064516129
  },
  "comparison": {
    "public_rc5_onnx_q8": {
      "core": 0.9668874172185431,
      "edge": 0.9743589743589743,
      "remaining_gaps": 0.888888888888889,
      "finance": 0.9361702127659575,
      "finance_boundary": 0.8750000000000001,
      "multilingual_ppsn": 0.9333333333333333,
      "user_ppsn": 1.0,
      "gaelic_weak_ppsn": 1.0,
      "overlap_ppsn": 1.0
    },
    "public_rc8_onnx_q8": {
      "min_score": 0.5,
      "irish_core_manual_f1": 0.9736842105263158,
      "irish_edge_f1": 1.0,
      "finance_suite_f1": 1.0,
      "finance_boundary_f1": 1.0,
      "user_raw_ppsn_f1": 1.0,
      "gaelic_weak_ppsn_f1": 1.0,
      "multilingual_ppsn_f1": 0.9176470588235294,
      "core_examples_per_second": 46.14201741375802,
      "multilingual_examples_per_second": 99.71655616732895
    }
  }
}