[CI/Build] Increase pooling tolerance to pass CI

[DarkLight1337]

Essential Elements of an Effective PR Description Checklist

• The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
• The test plan, such as providing test command.
• The test results, such as pasting the results comparison before and after, or e2e results
• (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.

Purpose

Increase the tolerance to pass CI. atol=0.02 is not too bad for cosine similarity

cc @maxdebayser @noooop

FIX #22836

Test Plan

Test Result

(Optional) Documentation Update

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[gemini-code-assist]

Code Review

This pull request increases the tolerance in two MTEB tests to pass CI. While this resolves the immediate issue, significantly increasing test tolerances (to 0.02 and 0.01) is a risky practice as it can hide underlying correctness issues or regressions. I've added critical comments suggesting to document these changes with a TODO to investigate the root cause of the discrepancies. This will ensure the test suite remains robust in the long term.

[DarkLight1337]

Let's pass the CI first, then open a follow-up PR to debug it on CI since I can't really repro this locally

[maxdebayser]

I prefer relative tolerance in general, but I'm ok with this.

[mgoin]

It seems snowflake is still failing

[2025-08-13T20:23:40Z] FAILED models/language/pooling/test_snowflake_arctic_embed.py::test_embed_models_mteb[model_info5] - AssertionError 

Let me try bumping it the same

[yewentao256]

Looks good to me, thanks for the work!