Add more useful options in OptimizerConfig

[garyzhang99]

Description

As the title says.

Checklist

Please check the following items before code is ready to be reviewed.

• Code has passed all tests
• Docstrings have been added/updated in Google Style
• Documentation has been updated
• Code is ready for review

[gemini-code-assist]

Summary of Changes

Hello @garyzhang99, 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 enhances the OptimizerConfig by adding more granular control over the optimization process. It introduces parameters for gradient clipping and various learning rate scheduling strategies, such as warmup, decay steps, decay style, and minimum learning rate. These additions provide greater flexibility for tuning model training, with corresponding test updates to ensure proper configuration propagation.

Highlights

• New Optimizer Configuration Options: Introduced several new fields to the OptimizerConfig class, including clip_grad, lr_warmup_init, lr_decay_steps, lr_decay_style, and min_lr.
• Updated Configuration Tests: Modified test_optimizer_config_propagation to verify that the newly added optimizer configuration options are correctly propagated through the system.

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. ↩

[garyzhang99]

/unittest-module-common

[gemini-code-assist]

Code Review

This pull request adds several new configuration options to OptimizerConfig, such as clip_grad, lr_decay_style, and min_lr, and updates the tests to verify their propagation. The changes are good, but I've suggested adding docstrings for the new options and clarifying the relationship between min_lr and the existing min_lr_ratio to improve maintainability and prevent potential misconfiguration.

[pan-x-c]

/unittest-module-common

[github-actions]

Summary

Tests 📝	Passed ✅	Failed ❌	Skipped ⏭️	Other ❓	Flaky 🍂	Duration ⏱️
34	34	0	0	0	0	4m 55s

Tests

Test Name	Status	Flaky	Duration
tests/common/config_test.py::TestConfig::test_all_examples_are_valid	✅		41.4s
tests/common/config_test.py::TestConfig::test_config_flatten	✅		39ms
tests/common/config_test.py::TestConfig::test_continue_from_checkpoint_is_valid	✅		188ms
tests/common/config_test.py::TestConfig::test_default_workflow	✅		89ms
tests/common/config_test.py::TestConfig::test_load_default_config	✅		3.3s
tests/common/config_test.py::TestConfig::test_max_token_len_per_gpu_set_correctly	✅		90ms
tests/common/config_test.py::TestConfig::test_optimizer_config_propagation	✅		89ms
tests/common/config_test.py::TestConfig::test_update_config_from_ray_cluster	✅		1.8s
tests/common/experience_test.py::TestEID::test_eid_properties	✅		1ms
tests/common/experience_test.py::TestExperience::test_action_mask_and_logprobs_type	✅		1ms
tests/common/experience_test.py::TestExperience::test_assertions	✅		1ms
tests/common/experience_test.py::TestExperience::test_dpo_experience	✅		1ms
tests/common/experience_test.py::TestExperience::test_gather	✅		1ms
tests/common/experience_test.py::TestExperience::test_hf_datasets_conversion	✅		15ms
tests/common/experience_test.py::TestExperience::test_multi_turn_experience	✅		1ms
tests/common/experience_test.py::TestExperience::test_serialize_deserialize	✅		2ms
tests/common/experience_test.py::TestExperience::test_single_turn_experience	✅		1ms
tests/common/experience_test.py::TestExperience::test_to_dict	✅		1ms
tests/common/experience_test.py::TestExperienceConversion::test_batch_conversion	✅		1ms
tests/common/experience_test.py::TestExperienceConversion::test_dpo_experience_batch_conversion	✅		1ms
tests/common/experience_test.py::TestExperienceConversion::test_experience_model_experience_conversion	✅		1ms
tests/common/experience_test.py::TestExperienceConversion::test_gather_experiences_with_custom_fields	✅		1ms
tests/common/experience_test.py::TestExperienceConversion::test_multiturn_experience_batch_converstion	✅		1ms
tests/common/vllm_test.py::ModelWrapperTest_0::test_generate	✅		51.8s
tests/common/vllm_test.py::ModelWrapperTest_1::test_generate	✅		32.0s
tests/common/vllm_test.py::ModelWrapperTest_2::test_generate	✅		43.1s
tests/common/vllm_test.py::TestModelLen_0::test_model_len	✅		17.1s
tests/common/vllm_test.py::TestModelLen_1::test_model_len	✅		17.4s
tests/common/vllm_test.py::TestAPIServer::test_api	✅		22.4s
tests/common/vllm_test.py::TestAsyncAPIServer::test_api_async	✅		22.4s
tests/common/vllm_test.py::TestTokenizer::test_action_mask	✅		236ms
tests/common/vllm_test.py::TestTokenizer::test_action_mask_with_tools	✅		229ms
tests/common/vllm_test.py::TestAPIServerToolCall_0_deepseek_r1::test_api_tool_calls	✅		19.7s
tests/common/vllm_test.py::TestAPIServerToolCall_1::test_api_tool_calls	✅		17.5s

Github Test Reporter by CTRF 💚

[github-actions]

Summary

Tests 📝	Passed ✅	Failed ❌	Skipped ⏭️	Other ❓	Flaky 🍂	Duration ⏱️
34	34	0	0	0	0	4m 55s

Tests

Test Name	Status	Flaky	Duration
tests/common/config_test.py::TestConfig::test_all_examples_are_valid	✅		41.1s
tests/common/config_test.py::TestConfig::test_config_flatten	✅		39ms
tests/common/config_test.py::TestConfig::test_continue_from_checkpoint_is_valid	✅		188ms
tests/common/config_test.py::TestConfig::test_default_workflow	✅		89ms
tests/common/config_test.py::TestConfig::test_load_default_config	✅		3.3s
tests/common/config_test.py::TestConfig::test_max_token_len_per_gpu_set_correctly	✅		92ms
tests/common/config_test.py::TestConfig::test_optimizer_config_propagation	✅		89ms
tests/common/config_test.py::TestConfig::test_update_config_from_ray_cluster	✅		1.9s
tests/common/experience_test.py::TestEID::test_eid_properties	✅		1ms
tests/common/experience_test.py::TestExperience::test_action_mask_and_logprobs_type	✅		1ms
tests/common/experience_test.py::TestExperience::test_assertions	✅		1ms
tests/common/experience_test.py::TestExperience::test_dpo_experience	✅		1ms
tests/common/experience_test.py::TestExperience::test_gather	✅		1ms
tests/common/experience_test.py::TestExperience::test_hf_datasets_conversion	✅		15ms
tests/common/experience_test.py::TestExperience::test_multi_turn_experience	✅		1ms
tests/common/experience_test.py::TestExperience::test_serialize_deserialize	✅		1ms
tests/common/experience_test.py::TestExperience::test_single_turn_experience	✅		1ms
tests/common/experience_test.py::TestExperience::test_to_dict	✅		1ms
tests/common/experience_test.py::TestExperienceConversion::test_batch_conversion	✅		1ms
tests/common/experience_test.py::TestExperienceConversion::test_dpo_experience_batch_conversion	✅		1ms
tests/common/experience_test.py::TestExperienceConversion::test_experience_model_experience_conversion	✅		1ms
tests/common/experience_test.py::TestExperienceConversion::test_gather_experiences_with_custom_fields	✅		1ms
tests/common/experience_test.py::TestExperienceConversion::test_multiturn_experience_batch_converstion	✅		1ms
tests/common/vllm_test.py::ModelWrapperTest_0::test_generate	✅		51.4s
tests/common/vllm_test.py::ModelWrapperTest_1::test_generate	✅		31.3s
tests/common/vllm_test.py::ModelWrapperTest_2::test_generate	✅		43.6s
tests/common/vllm_test.py::TestModelLen_0::test_model_len	✅		17.3s
tests/common/vllm_test.py::TestModelLen_1::test_model_len	✅		17.3s
tests/common/vllm_test.py::TestAPIServer::test_api	✅		23.4s
tests/common/vllm_test.py::TestAsyncAPIServer::test_api_async	✅		22.6s
tests/common/vllm_test.py::TestTokenizer::test_action_mask	✅		264ms
tests/common/vllm_test.py::TestTokenizer::test_action_mask_with_tools	✅		242ms
tests/common/vllm_test.py::TestAPIServerToolCall_0_deepseek_r1::test_api_tool_calls	✅		19.4s
tests/common/vllm_test.py::TestAPIServerToolCall_1::test_api_tool_calls	✅		17.5s

Github Test Reporter by CTRF 💚