fix(llm_simple_qa): replace hardcoded paths and fix basemodel compatibility

[rakshaak29]

What this PR does

Fixes the llm_simple_qa example which was broken due to hardcoded paths and missing methods.
Fixes #387

Changes

• YAML configs: Replaced all hardcoded absolute paths (/home/icyfeather/...) with relative paths (./examples/..., ./dataset/..., ./workspace)
• basemodel.py:
  • Use HuggingFace Hub model ID (Qwen/Qwen2.5-0.5B-Instruct) instead of local path
  • Added missing preprocess() method required by sedna's SingleTaskLearning paradigm
  • Added MPS/CPU device detection for cross-platform compatibility

Testing

Verified the benchmark runs end-to-end successfully:

$ ianvs -f examples/llm_simple_qa/benchmarkingjob.yaml

Loading weights: 100%|████████████████████| 290/290 [00:01<00:00, 250.13it/s]
BaseModel doesn't need to preprocess
BaseModel doesn't need to train
BaseModel doesn't need to save
BaseModel load
BaseModel predict
+------+-------------------------------+-----+--------------------+-----------+---------------------+
| rank |           algorithm           | acc |      paradigm      | basemodel |         time        |
+------+-------------------------------+-----+--------------------+-----------+---------------------+
|  1   | simple_qa_singletask_learning | 0.8 | singletasklearning |    gen    | 2026-04-12 05:26:45 |
+------+-------------------------------+-----+--------------------+-----------+---------------------+

[kubeedge-bot]

Welcome @rakshaak29! It looks like this is your first PR to kubeedge/ianvs 🎉

[kubeedge-bot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: rakshaak29
To complete the pull request process, please assign jaypume after the PR has been reviewed.
You can assign the PR to them by writing /assign @jaypume in a comment when ready.

The full list of commands accepted by this bot can be found here.

Details

Needs approval from an approver in each of these files:

• OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[gemini-code-assist]

Code Review

This pull request updates configuration paths to relative links, adds local artifacts to .gitignore, and switches the base model to a public Hugging Face repository. Feedback identifies a regression in basemodel.py where CUDA support was removed from device detection, potentially causing runtime errors or performance degradation. Additionally, the new preprocess method returns None, which may break the data pipeline if subsequent stages expect processed data.

[rakshaak29]

/assign @jaypume