[CI] [GHA] Use snapshot_download for HF models

[akashchi]

Should lower the number of API requests to the HF servers.

It was introduced and tested in openvinotoolkit/openvino/pull/32282 and openvinotoolkit/openvino/pull/32458

Ticket:

• 180694

[Copilot]

Pull request overview

This PR updates the test suite to use snapshot_download from huggingface_hub instead of directly calling model loading functions with model IDs. This change helps reduce the number of API requests to HuggingFace servers by downloading all model files at once and caching them locally before loading models and tokenizers.

Changes:

• Replaced direct model ID usage with snapshot_download() calls to pre-cache models and reduce HF API rate limit issues
• Added huggingface_hub.snapshot_download imports across multiple test files
• Applied the pattern consistently across test files for Whisper, VLM, tokenizer, and parser tests

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
tests/python_tests/utils/hugging_face.py	Updated get_huggingface_models() and GGUF loading functions to use snapshot_download for caching models before loading
tests/python_tests/test_whisper_pipeline_static.py	Added snapshot_download call in load_and_save_whisper_model() to cache model before loading processor and tokenizer
tests/python_tests/test_whisper_pipeline.py	Added snapshot_download call in save_to_temp() to cache model before loading tokenizer, model, and processor
tests/python_tests/test_vlm_pipeline.py	Added snapshot_download calls in multiple functions to cache VLM models before loading processors and tokenizers
tests/python_tests/test_vllm_parsers_wrapper.py	Added inline snapshot_download calls when creating parsers with tokenizers
tests/python_tests/test_tokenizer.py	Added snapshot_download calls in multiple test functions to cache models before loading tokenizers
tests/python_tests/test_text_streamer.py	Added inline snapshot_download calls when loading tokenizers for text streaming tests
tests/python_tests/test_parsers.py	Added snapshot_download call in fixture to cache model before loading tokenizer

[Copilot]

Pull request overview

Copilot reviewed 8 out of 8 changed files in this pull request and generated 14 comments.

[Copilot]

Pull request overview

Copilot reviewed 8 out of 8 changed files in this pull request and generated 6 comments.