perf: Update HuggingFace datasets subsystem benchmark baseline to use 4 dataloader workers

[yuxin00j]

Overview
This PR updates the num_workers parameter in the baseline configuration for the HuggingFace datasets subsystem benchmarks, reducing it from 16 to 4.

Details

• Target File: gcsfs/tests/perf/subsystembenchmarks/dataloading/huggingface_datasets/configs.yaml
• Change: Adjusted the num_workers value in the common.baseline configuration from 16 to 4.
• Impact: This establishes 4 workers as the new baseline for our dataloading subsystem benchmark sweeps. The scaling variants sweep (which tests 0, 1, and 8 workers) remains intact, providing a clean comparison [0, 1, 4 (baseline), 8] to measure performance scaling across different dataloader concurrency levels.

[gemini-code-assist]

Code Review

This pull request updates the Hugging Face datasets benchmark configuration by reducing the number of workers (num_workers) from 16 to 4. There are no review comments, and I have no feedback to provide.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.68%. Comparing base (e6c107b) to head (ccffbf8).
⚠️ Report is 6 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #986   +/-   ##
=======================================
  Coverage   89.68%   89.68%           
=======================================
  Files          16       16           
  Lines        3579     3579           
=======================================
  Hits         3210     3210           
  Misses        369      369           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.