feat: Unified model-family inference engines (including image-classification) and KServe v2 API support

[cau-git]

Summary

This cumulative PR...

• Adds inference engine abstraction for the image classification model family, enabling runtime selection between ONNX Runtime and HuggingFace Transformers without model retraining. This follows the same pattern established in feat: Inference engines abstraction for object detection model family with Transformers and ONNXRuntime #2959 for object detection models.
• Implements a new Kserve-API engine type (ApiKserveV2ObjectDetectionEngine, ApiKserveV2ImageClassificationEngine) and the necessary options to comsume remote object-detection and image-classifier models in docling. It adheres to the Kserve v2 inference API spec.

Changes

New Infrastructure

• BaseImageClassificationEngine: Abstract interface with factory pattern
• OnnxRuntimeImageClassificationEngine: ONNX Runtime implementation
• TransformersImageClassificationEngine: HuggingFace Transformers implementation
• HfVisionModelMixin: Shared utilities for HF model loading and label conversion
• Engine-specific options classes with validation

Integration

• Updated DocumentPictureClassifier to use the new engine abstraction
• Added image classification model specs to stage model configuration
• Pipeline options now support engine selection for picture classification

Further changes

• Updated torch.compile support both on object detection and image classification model family transformer engine
• Added compile_model in the engine options, defaulting to False

Not in this PR

• AUTO_INLINE engine selection support

Checklist:

• Documentation has been updated, if necessary.
• Examples have been added, if necessary.
• Tests have been added, if necessary.

[github-actions]

❌ DCO Check Failed

Hi @cau-git, your pull request has failed the Developer Certificate of Origin (DCO) check.

This repository supports remediation commits, so you can fix this without rewriting history — but you must follow the required message format.

---

🛠 Quick Fix: Add a remediation commit

Run this command:

git commit --allow-empty -s -m "DCO Remediation Commit for Christoph Auer <cau@zurich.ibm.com>

I, Christoph Auer <cau@zurich.ibm.com>, hereby add my Signed-off-by to this commit: 4cdb01e6d3b34ae3444707cfe0b782e6b6a0cd16
I, Christoph Auer <cau@zurich.ibm.com>, hereby add my Signed-off-by to this commit: f1701b826a376c68dab6e077deb2fa7394ff27a9"
git push

---

🔧 Advanced: Sign off each commit directly

For the latest commit:

git commit --amend --signoff
git push --force-with-lease

For multiple commits:

git rebase --signoff origin/main
git push --force-with-lease

More info: DCO check report

[mergify]

Merge Protections

Your pull request matches the following merge protections and will not be merged until they are valid.

🟢 Enforce conventional commit

Wonderful, this rule succeeded.

Make sure that we follow https://www.conventionalcommits.org/en/v1.0.0/

• title ~= ^(fix|feat|docs|style|refactor|perf|test|build|ci|chore|revert)(?:\(.+\))?(!)?:

[codecov]

Codecov Report

❌ Patch coverage is 57.95181% with 349 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
...engines/image_classification/onnxruntime_engine.py	0.00%	87 Missing ⚠️
...e_engines/object_detection/api_kserve_v2_engine.py	34.93%	54 Missing ⚠️
...gines/image_classification/api_kserve_v2_engine.py	37.14%	44 Missing ⚠️
.../models/inference_engines/common/kserve_v2_http.py	51.68%	43 Missing ⚠️
docling/datamodel/stage_model_specs.py	70.42%	21 Missing ⚠️
...els/inference_engines/image_classification/base.py	75.00%	21 Missing ⚠️
.../models/inference_engines/common/hf_vision_base.py	75.34%	18 Missing ⚠️
.../inference_engines/image_classification/hf_base.py	55.26%	17 Missing ⚠️
...ngines/image_classification/transformers_engine.py	83.52%	14 Missing ⚠️
...ce_engines/object_detection/transformers_engine.py	0.00%	10 Missing ⚠️
... and 9 more

📢 Thoughts on this report? Let us know!

[dosubot]

Related Documentation

Checked 15 published document(s) in 1 knowledge base(s). No updates required.

^{How did I do? Any feedback?}  

[dolfim-ibm]

lgtm