Add HuggingFace Model Wrapper for Semantic-Segmentation

[harimkang]

Summary

*pipeline: /base/data/semantic_segmentation.yaml & optimizer: AdamW (lr 0.0001)
*The current settings are probably not optimal for each model, so we should be able to improve performance based on the more optimal settings.

How to test

Install transformers
pip install transformers
or
pip install -e .[transformers]

API Examples

from otx.algo.segmentation.huggingface_model import HuggingFaceModelForSegmentation
from otx.engine import Engine

data_root = "/datasets/otx_v2_dataset/semantic_seg/kvasir_medium"

otx_model = HuggingFaceModelForSegmentation(
    model_name_or_path="nvidia/segformer-b0-finetuned-ade-512-512",
    label_info=2,
)

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.segmentation.huggingface_model.HuggingFaceModelForSegmentation \
    --model.model_name_or_path nvidia/segformer-b0-finetuned-ade-512-512 \
    --data_root /datasets/otx_v2_dataset/semantic_seg/kvasir_medium \
    --work_dir otx-workspace-hf-cli --max_epochs 3

Unit-Test:

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

[kprokofi]

Thank you for your work

[wonjuleee]

LGTM!