Add DoRA support for ViT classification model.

[gyuilLim]

Summary

Added DoRA support and replaced lora falg with peft argument using Literal["lora", "dora"].

How to 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) 2025 Intel Corporation
# SPDX-License-Identifier: Apache-2.0

[sovrasov]

@gyuilLim thanks for your efforts! Could you add a brief version of the table with results of the experiments? That'd be useful in the future, so everyone can see the impact of the PR by simply looking at description.
Also, the PR doesn't contain any tests. Newly added modules should be covered with unit tests. It'd be good to check that peft parameter actually works, like Kirill said. It might be enough to check if a multiclass cls ViT model can be created and forwarded with peft not None. Actual training seems to be a bit redundant and we have don't have a redy-to-use location to store that kind of tests

[gyuilLim]

@sovrasov Thank you! I'll add unit tests.
and here is the table that shows exprimental results using DINOv2 on six datasets

• GPU memory usage(GB)

Method	FGVC_aircraft	FOOD_101	Stanford_Cars	CUB-200	HAM10000	RESISC45	Avg Memory
Full Fine-Tuning	4.050	4.070	4.010	4.020	3.990	4.050	4.032
Freeze backbone	0.550	0.560	0.550	0.560	0.530	0.540	0.548
LoRA(q, v)	2.580	2.640	2.580	2.590	2.560	2.570	2.587
DoRA(q, v)	2.680	2.700	2.680	2.730	2.670	2.670	2.688

• Test Accuracy(%)

Method	FGVC_aircraft	FOOD_101	Stanford_Cars	CUB-200	HAM10000	RESISC45	Avg Top-1 Acc
Full Fine-Tuning	47.11	83.87	75.0	75.41	78.16	92.14	75.28
Freeze backbone	38.52	87.12	66.24	81.98	59.76	80.87	69.08
LoRA(q, v)	49.12	88.53	76.51	80.36	78.62	93.05	77.70
DoRA(q, v)	50.04	88.87	75.70	81.05	79.02	92.76	77.90