Disable GPTQModel in CI Dockerfile to stop pcre-induced core dumps

{
  "event_id": "event_1c97a315bd6f128e",
  "topic_id": "gpu_supply_and_compute_market",
  "event_type": "Pull Request",
  "event_time": "2026-06-04T15:43:31Z",
  "title": "Disable GPTQModel in CI Dockerfile to stop pcre-induced core dumps",
  "summary": "HuggingFace PEFT temporarily commented out GPTQModel integration in its CI Dockerfile because the pcre dependency caused a core dump when running tests on the CI GPU runner.",
  "contribution": "Prevents CI failures on GPU runners by temporarily removing a model integration that crashes due to a dependency conflict.",
  "impact": "Developers and maintainers using the PEFT library's CI pipelines will see fewer spurious test failures, making development velocity more predictable. The underlying issue is that GPTQModel's reliance on the pcre library triggers a core dump in the GPU-based CI environment, so disabling it is a workaround. Continued observation is needed to see if a root-cause fix for the pcre conflict is implemented or if this workaround becomes permanent, potentially delaying GPTQModel feature validation.",
  "maturity": "Code",
  "confidence": 0,
  "importance_score": 0.55,
  "risk_flags": [
    "Workaround Not Fix",
    "Gptqmodel Validation Disabled"
  ],
  "evidence_count": 1
}