feat(loader): add pipeline_tag fallback to task detection

[xieofxie]

Summary

When architecture-based task detection (Stage 1c) fails because config.architectures contains a generic name like 'Model' that doesn't map to any real transformers class, task resolution falls through to the last-resort feature-extraction default. However, HuggingFace Hub knows the correct task via pipeline_tag (e.g. audio-classification for audeering/wav2vec2-large-robust-24-ft-age-gender).

This PR adds a Stage 1e between Stage 1c (TasksManager architecture detection) and Stage 1d (last-resort default) that queries the Hub's pipeline_tag as a fallback.

Changes

src/winml/modelkit/loader/resolution.py

• Added TaskSource.PIPELINE_TAG = "pipeline-tag" enum value for provenance tracking
• Added KNOWN_TASKS to imports from .task
• Added Stage 1e block in resolve_task():
  • Only runs when opt_task is None (architecture detection failed) and model_id exists
  • Checks model_id is not a local path via _is_local_path from hub_utils
  • Queries HfApi().model_info(model_id).pipeline_tag
  • Normalizes the tag and validates against KNOWN_TASKS
  • Wrapped in try/except Exception for graceful fallthrough on network errors or invalid model IDs

tests/unit/loader/test_detect_task.py

Added 4 tests:

• Valid pipeline_tag is used when architecture detection fails
• Invalid/unknown pipeline_tag falls through to default
• API failure (network error) falls through gracefully
• Local path model_id skips the Hub API call entirely

[xieofxie]

a follow up of #1094

tested via winml export -m audeering/wav2vec2-large-robust-24-ft-age-gender -o temp/wav2vec2.onnx

[DingmaomaoBJTU]

Thanks for the pipeline_tag fallback — the overall shape is right: it only runs on the last-resort path (opt_task is None), guards against local paths, and falls through gracefully. A few things worth addressing before merge, left inline:

1. resolve_task now does blocking network I/O on a path that used to be purely local — no timeout, no caching, and the docstring's stage list wasn't updated.
2. The bare except Exception: pass drops all diagnostics.
3. The local-path negative test can't actually fail, because the AssertionError it raises is swallowed by that same except Exception.

[DingmaomaoBJTU]

Reviewed the pipeline_tag fallback — the core idea is sound and the resolve_task dispatch is well covered. A few things inline worth a look: the KNOWN_TASKS gate admits display-only tasks that aren't ONNX-exportable; this adds a Hub network call to the shared resolve_task path (build/eval/inference, not just inspect); the 1e/1d labels read out of execution order; and the local-path test mocks the function under test so it doesn't cover the skip.

One more, not inline: the PR description looks slightly stale — it says the HfApi().model_info(...) call is inlined in resolution.py and lists 2 changed files, but the implementation factors it into a get_pipeline_tag helper in utils/hub_utils.py (3 files). Worth updating so the description matches.