chore: trim tests in CI

[garrett4wade]

No description provided.

[gemini-code-assist]

Summary of Changes

Hello @garrett4wade, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request focuses on optimizing the continuous integration (CI) pipeline by strategically reducing the execution time of tests. This is primarily achieved by marking several resource-intensive or time-consuming tests with @pytest.mark.slow, ensuring they are skipped in standard CI runs. Additionally, a vLLM-related test configuration is removed from an example test, as its functionality is already validated in other dedicated test suites, further streamlining the CI process.

Highlights

• CI Test Optimization for vLLM: The test_gsm8k_grpo in areal/tests/test_examples.py has been updated to skip vLLM-related test configurations in CI environments, as vLLM is too slow to launch and its integration is covered by other dedicated tests.
• Marking FSDP Engine Test as Slow: The test_fsdpengine_nccl_weight_update_to_remote in areal/tests/test_fsdp_engine_nccl.py is now marked with @pytest.mark.slow, indicating it will be skipped in standard CI runs due to existing integration tests.
• Megatron Engine Distributed Tests Marked Slow: Several multi-GPU tests for the Megatron engine, including test_qwen3_tensor_parallel, test_qwen3_pipeline_parallel, test_qwen3_context_parallel, test_qwen3_virtual_pipeline_parallel, and test_qwen3moe_expert_parallel in areal/tests/test_megatron_engine_distributed.py, have been marked with @pytest.mark.slow.
• Offload Test Marked Slow: The test_engine_offload_and_onload in areal/tests/test_offload.py has been marked with @pytest.mark.slow to optimize CI execution time.
• Train Engine Save/Load Test Marked Slow: The test_dcp_save_load_weights in areal/tests/test_train_engine.py has been updated to use @pytest.mark.slow instead of @torch.no_grad(), indicating it will be skipped in faster CI runs.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request aims to trim tests in CI to speed up the process. Most changes involve adding @pytest.mark.slow to long-running tests or removing slow test parameterizations, which is appropriate. However, I found one issue in areal/tests/test_train_engine.py where the @torch.no_grad() decorator was incorrectly removed from a test, which can lead to unnecessary gradient computations and resource usage. I've suggested a fix for this.