Add HuggingFace Model Wrapper for Multi-Class Classification

[harimkang]

Summary

• Add HuggingFaceModelForMulticlassCls
• Training, Testing, Exporting works

*The results of running the existing models are from workstation, and for the hugging face model, the results are from a different machine, so it is difficult to compare the time directly. Just a note.

Checked models

• "nateraw/vit-base-beans", # Available
• "google/vit-base-patch16-224", # Available
• "google/mobilenet_v1_1.0_224", # Available
• "google/efficientnet-b0", # Available
• "microsoft/beit-base-patch16-224-pt22k-ft22k", # Available
• "microsoft/swin-tiny-patch4-window7-224", # Available
• "microsoft/resnet-18", # Available (but show NAN loss)
• "facebook/convnextv2-tiny-1k-224", # Available
• "facebook/dinov2-small-imagenet1k-1-layer", # Available
• "facebook/deit-tiny-patch16-224", # Available

I haven't checked them all (too many models), but most of the models in the list below are available.
https://huggingface.co/models?pipeline_tag=image-classification

How to test

Install transformers
pip install transformers

API Examples

from otx.algo.classification.huggingface_model import HuggingFaceModelForMulticlassCls
from otx.engine import Engine

data_root = "/datasets/otx_v2_dataset/multiclass_classification/multiclass_food101_large"

otx_model = HuggingFaceModelForMulticlassCls(
    model_name_or_path="facebook/deit-tiny-patch16-224",
    label_info=20,
)

engine = Engine(
    data_root=data_root,
    model=otx_model,
    work_dir="otx-workspace-hf",
)

engine.train(max_epochs=2)
engine.test()
engine.export()
engine.test(checkpoint="otx-workspace-hf/exported_model.xml")

CLI Examples

(venv) ~/otx$ otx train \
    --model otx.algo.classification.huggingface_model.HuggingFaceModelForMulticlassCls \
    --model.model_name_or_path facebook/deit-tiny-patch16-224 \
    --data_root /datasets/otx_v2_dataset/multiclass_classification/multiclass_food101_large \
    --work_dir otx-workspace-hf-cli --max_epochs 3

Checklist

• I have added unit tests to cover my changes.​
• I have added integration tests to cover my changes.​
• I have ran e2e tests and there is no issues.
• I have added the description of my changes into CHANGELOG in my target branch (e.g., CHANGELOG in develop).​
• I have updated the documentation in my target branch accordingly (e.g., documentation in develop).
• I have linked related issues.

License

• I submit my code changes under the same Apache License that covers the project.
  Feel free to contact the maintainers if that's a concern.
• I have updated the license header for each file (see an example below).

# Copyright (C) 2024 Intel Corporation
# SPDX-License-Identifier: Apache-2.0

[eunwoosh]

Thanks for your work :) I left minor comment, Please take a look.