Rename warmup_style to lr_scheduler_type

[chenyushuo]

Description

1. Rename warmup_style to lr_scheduler_type.
2. Remove lr_warmup_init, lr_decay_steps, lr_decay_style and min_lr in OptimizerConfig.
3. Bug fix in tinker_model.py.

Unittest Results

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 @chenyushuo, 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 enhancing the clarity and maintainability of the codebase by standardizing the naming convention for learning rate scheduler configurations. The parameter previously known as warmup_style has been renamed to lr_scheduler_type, which more accurately describes its function. This change is propagated throughout various configuration files and updated in the documentation, guiding users towards the new, more descriptive parameter and ensuring a consistent approach to defining learning rate schedules.

Highlights

• Configuration Parameter Rename: The configuration parameter warmup_style has been renamed to lr_scheduler_type for improved clarity and consistency in defining learning rate scheduling.
• Extensive Configuration File Updates: This rename has been applied across numerous YAML configuration files, including benchmark templates, example configurations, and test configurations, ensuring all relevant files use the new parameter name.
• Documentation Updates: The documentation (both English and Chinese versions) has been updated to reflect this change, deprecating warmup_style and introducing lr_scheduler_type with its supported values (constant, cosine).
• Codebase Adaptation: The core configuration definition in trinity/common/verl_config.py has been updated to include both the deprecated warmup_style (marked as such) and the new lr_scheduler_type.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

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 refactors warmup_style to lr_scheduler_type across various configuration files and documentation. The changes are mostly consistent. I've pointed out a typo in the documentation and a more significant issue regarding backward compatibility in trinity/common/verl_config.py, where the deprecated warmup_style field is not handled, potentially causing issues with older configurations.

[pan-x-c]

/unittest-diff

[github-actions]

Summary

Tests 📝	Passed ✅	Failed ❌	Skipped ⏭️	Other ❓	Flaky 🍂	Duration ⏱️
68	62	0	6	0	0	52m 36s

Skipped

Tests	Status
tests/common/vllm_test.py::TestTinkerAsyncAPIServer::test_api_async	skipped ⏭️
tests/trainer/trainer_test.py::TestMultiModalGRPO::test_trainer	skipped ⏭️
tests/trainer/trainer_test.py::TestMultiModalSFT::test_trainer	skipped ⏭️
tests/trainer/trainer_test.py::TestTinkerTrainer::test_trainer	skipped ⏭️
tests/trainer/trainer_test.py::TestTinkerTrainer::test_trainer_class	skipped ⏭️
tests/trainer/trainer_test.py::AgentScopeTunerTest::test_agentscope_tuner	skipped ⏭️

Tests

Test Name	Status	Flaky	Duration
tests/common/config_test.py::TestConfig::test_all_examples_are_valid	✅		34.9s
tests/common/config_test.py::TestConfig::test_chat_template_path	✅		74ms
tests/common/config_test.py::TestConfig::test_config_flatten	✅		30ms
tests/common/config_test.py::TestConfig::test_continue_from_checkpoint_is_valid	✅		154ms
tests/common/config_test.py::TestConfig::test_default_workflow	✅		71ms
tests/common/config_test.py::TestConfig::test_load_default_config	✅		3.9s
tests/common/config_test.py::TestConfig::test_max_token_len_per_gpu_set_correctly	✅		73ms
tests/common/config_test.py::TestConfig::test_optimizer_config_propagation	✅		73ms
tests/common/config_test.py::TestConfig::test_update_config_from_ray_cluster	✅		2.0s
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_gather_with_token_level_reward	✅		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	✅		1m 17s
tests/common/vllm_test.py::ModelWrapperTest_1::test_generate	✅		1m 5s
tests/common/vllm_test.py::ModelWrapperTest_2::test_generate	✅		35.0s
tests/common/vllm_test.py::TestModelLen_0::test_model_len	✅		58.5s
tests/common/vllm_test.py::TestModelLen_1::test_model_len	✅		22.1s
tests/common/vllm_test.py::TestModelLen_2::test_model_len	✅		51.4s
tests/common/vllm_test.py::TestModelLenWithoutPromptTruncation::test_model_len	✅		51.8s
tests/common/vllm_test.py::TestAPIServer::test_api	✅		24.0s
tests/common/vllm_test.py::TestLogprobs::test_logprobs_api	✅		51.2s
tests/common/vllm_test.py::TestAsyncAPIServer::test_api_async	✅		23.8s
tests/common/vllm_test.py::TestTinkerAsyncAPIServer::test_api_async	⏭️		1ms
tests/common/vllm_test.py::TestTokenizer::test_action_mask	✅		256ms
tests/common/vllm_test.py::TestTokenizer::test_action_mask_with_tools	✅		254ms
tests/common/vllm_test.py::TestAPIServerToolCall_0_deepseek_r1::test_api_tool_calls	✅		53.3s
tests/common/vllm_test.py::TestAPIServerToolCall_1::test_api_tool_calls	✅		51.9s
tests/common/vllm_test.py::TestSuperLongGeneration::test_generate	✅		3m 11s
tests/common/vllm_test.py::TestTinkerAPI::test_tinker_api	✅		1m 13s
tests/trainer/trainer_test.py::TestTrainerCountdown_0_fsdp::test_trainer	✅		2m 46s
tests/trainer/trainer_test.py::TestTrainerCountdown_1_megatron::test_trainer	✅		4m 21s
tests/trainer/trainer_test.py::TestStepAheadAsyncRL::test_trainer	✅		1m 47s
tests/trainer/trainer_test.py::TestTrainerGSM8K_0_fsdp::test_trainer	✅		1m 20s
tests/trainer/trainer_test.py::TestTrainerGSM8K_1_fsdp2::test_trainer	✅		48.6s
tests/trainer/trainer_test.py::TestTrainerGSM8K_2_fsdp::test_trainer	✅		52.3s
tests/trainer/trainer_test.py::TestTrainerGSM8K_3_fsdp2::test_trainer	✅		58.8s
tests/trainer/trainer_test.py::TestTrainerSFTWarmupGSM8K::test_trainer	✅		1m 58s
tests/trainer/trainer_test.py::TestTrainerDPO::test_trainer	✅		35.5s
tests/trainer/trainer_test.py::TestTrainerSFT::test_trainer	✅		31.0s
tests/trainer/trainer_test.py::TestTrainerToolsSFT::test_trainer_tools	✅		30.1s
tests/trainer/trainer_test.py::TestFullyAsyncMode_0_fsdp::test_fully_async_mode	✅		1m 36s
tests/trainer/trainer_test.py::TestFullyAsyncMode_1_fsdp::test_fully_async_mode	✅		1m 31s
tests/trainer/trainer_test.py::TestFullyAsyncMode_2_megatron::test_fully_async_mode	✅		2m 20s
tests/trainer/trainer_test.py::TestTrainerCheckpointSave_0_fsdp::test_trainer	✅		2m 24s
tests/trainer/trainer_test.py::TestTrainerCheckpointSave_1_megatron::test_trainer	✅		5m 23s
tests/trainer/trainer_test.py::TestTrainerMIX::test_trainer	✅		1m 35s
tests/trainer/trainer_test.py::TestServeWithTrainer::test_serve_with_trainer	✅		1m 48s
tests/trainer/trainer_test.py::TestMultiModalGRPO::test_trainer	⏭️		2.0s
tests/trainer/trainer_test.py::TestMultiModalSFT::test_trainer	⏭️		554ms
tests/trainer/trainer_test.py::TestTrainerLoRA::test_trainer	✅		2m 43s
tests/trainer/trainer_test.py::TestOverRollout::test_trainer	✅		46.9s
tests/trainer/trainer_test.py::TestTrainerPromptTruncation::test_trainer	✅		1m 12s
tests/trainer/trainer_test.py::TestTinkerTrainer::test_trainer	⏭️		1ms
tests/trainer/trainer_test.py::TestTinkerTrainer::test_trainer_class	⏭️		1ms
tests/trainer/trainer_test.py::AgentScopeTunerTest::test_agentscope_tuner	⏭️		1ms

Github Test Reporter by CTRF 💚