Add `rope_scaling` and `rope_theta` to config #390

chenyushuo · 2025-11-18T06:52:40Z

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 · 2025-11-18T06:52:55Z

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 introduces dedicated configuration fields for Rotary Positional Embedding (RoPE) parameters, rope_scaling and rope_theta, across various model configuration structures. The changes ensure these parameters are consistently defined, propagated, and applied during model initialization, particularly for VLLM engines and within actor/auxiliary model setups. This provides enhanced control over positional encoding strategies for improved model performance and flexibility.

Highlights

Configuration Extension: Added rope_scaling (dictionary) and rope_theta (float) as optional configuration parameters to ModelConfig, InferenceModelConfig, and ActorModel classes to support Rotary Positional Embedding settings.
Configuration Propagation: Updated the check_and_update method in Config to ensure rope_scaling and rope_theta are correctly propagated to rollout_model and auxiliary_models.
VLLM Integration: Integrated the new rope_scaling and rope_theta parameters into the initialization of vllm.AsyncEngineArgs within vllm_model.py, allowing VLLM to utilize these RoPE settings.
Actor Model Synchronization: Modified synchronize_config in verl_config.py to synchronize rope_scaling and rope_theta from the main configuration to the actor_rollout_ref.model.
FSDP Worker Patching: Implemented logic in fsdp_workers.py to patch the actor_model_config with rope_scaling and rope_theta if these parameters are defined in the model configuration.

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.

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 adds support for rope_scaling and rope_theta configurations. The changes are well-implemented across the configuration files, the vLLM model wrapper, and the FSDP trainer worker, ensuring the new parameters are correctly propagated. I have one suggestion to improve maintainability by centralizing the list of RoPE argument names to avoid duplication.

trinity/common/models/vllm_model.py

trinity/trainer/verl/megatron_workers.py

chenyushuo · 2025-11-18T09:54:04Z

/unittest-module-explorer

github-actions · 2025-11-18T10:07:41Z

Summary

Tests 📝	Passed ✅	Failed ❌	Skipped ⏭️	Other ❓	Flaky 🍂	Duration ⏱️
36	35	0	1	0	0	12m 1s

Skipped

Tests	Status
tests/explorer/workflow_test.py::TestAgentScopeWorkflowAdapter::test_adapter	skipped ⏭️

Tests

Test Name	Status	Duration
tests/explorer/explorer_test.py::TestExplorerCountdownEval::test_explorer	✅	1m 19s
tests/explorer/explorer_test.py::TestExplorerGSM8KRULERNoEval::test_explorer	✅	1m 42s
tests/explorer/explorer_test.py::TestExplorerGSM8k::test_explorer	✅	3m 38s
tests/explorer/explorer_test.py::ServeTest::test_serve	✅	1m 24s
tests/explorer/scheduler_test.py::SchedulerTest::test_async_workflow	✅	12.6s
tests/explorer/scheduler_test.py::SchedulerTest::test_concurrent_operations	✅	12.1s
tests/explorer/scheduler_test.py::SchedulerTest::test_get_results	✅	30.8s
tests/explorer/scheduler_test.py::SchedulerTest::test_multi_step_execution	✅	12.5s
tests/explorer/scheduler_test.py::SchedulerTest::test_non_repeatable_workflow	✅	12.5s
tests/explorer/scheduler_test.py::SchedulerTest::test_scheduler_all_methods	✅	22.2s
tests/explorer/scheduler_test.py::SchedulerTest::test_scheduler_restart_after_stop	✅	23.4s
tests/explorer/scheduler_test.py::SchedulerTest::test_split_tasks	✅	15.5s
tests/explorer/scheduler_test.py::SchedulerTest::test_stepwise_experience_eid	✅	12.5s
tests/explorer/scheduler_test.py::SchedulerTest::test_wait_all	✅	15.5s
tests/explorer/scheduler_test.py::SchedulerTest::test_wait_all_timeout_with_multi_batch	✅	21.0s
tests/explorer/step_wise_workflow_test.py::WorkflowTest::test_reward_propagation_workflow_0	✅	2ms
tests/explorer/step_wise_workflow_test.py::WorkflowTest::test_reward_propagation_workflow_1	✅	602ms
tests/explorer/step_wise_workflow_test.py::WorkflowTest::test_step_wise_reward_workflow_0	✅	1ms
tests/explorer/step_wise_workflow_test.py::WorkflowTest::test_step_wise_reward_workflow_1	✅	1.0s
tests/explorer/step_wise_workflow_test.py::WorkflowTest::test_workflows_raise_error	✅	1ms
tests/explorer/step_wise_workflow_test.py::WorkflowTest::test_workflows_stop_at_max_env_steps	✅	1.0s
tests/explorer/workflow_test.py::WorkflowTest::test_gsm8k_workflow	✅	35ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_boxed_workflow	✅	24ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_complex_workflow	✅	174ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_eval_workflow	✅	3ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_fraction_workflow	✅	13ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_workflow	✅	8ms
tests/explorer/workflow_test.py::WorkflowTest::test_rm_gallery_workflow	✅	86ms
tests/explorer/workflow_test.py::WorkflowTest::test_workflow_repeatable_0	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_workflow_repeatable_1	✅	100ms
tests/explorer/workflow_test.py::WorkflowTest::test_workflow_resettable_0	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_workflow_resettable_1	✅	201ms
tests/explorer/workflow_test.py::MultiTurnWorkflowTest_0::test_multi_turn_workflow	✅	15.1s
tests/explorer/workflow_test.py::MultiTurnWorkflowTest_1::test_multi_turn_workflow	✅	14.7s
tests/explorer/workflow_test.py::TestAgentScopeWorkflowAdapter::test_adapter	⏭️	1ms
tests/explorer/workflow_test.py::TestWorkflowRunner::test_workflow_runner	✅	294ms

Github Test Reporter by CTRF 💚

Add rope_scaling and rope_theta to config

7ef71a2

gemini-code-assist bot reviewed Nov 18, 2025

View reviewed changes

trinity/common/models/vllm_model.py Show resolved Hide resolved

chenyushuo mentioned this pull request Nov 18, 2025

支持yarn扩展rollout模型上下文长度吗 #389

Closed

Add rope_scaling and rope_theta to megatron workers

c00811b

pan-x-c reviewed Nov 18, 2025

View reviewed changes

trinity/trainer/verl/megatron_workers.py Show resolved Hide resolved

add unittest

d3e42bf

pan-x-c approved these changes Nov 18, 2025

View reviewed changes

pan-x-c merged commit 0c58b5e into agentscope-ai:main Nov 18, 2025
1 check passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Add `rope_scaling` and `rope_theta` to config #390

Add `rope_scaling` and `rope_theta` to config #390

chenyushuo commented Nov 18, 2025

Uh oh!

gemini-code-assist bot commented Nov 18, 2025

Uh oh!

gemini-code-assist bot left a comment

Uh oh!

Uh oh!

Uh oh!

chenyushuo commented Nov 18, 2025

Uh oh!

github-actions bot commented Nov 18, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Add rope_scaling and rope_theta to config #390

Add rope_scaling and rope_theta to config #390

Conversation

chenyushuo commented Nov 18, 2025

Description

Checklist

Uh oh!

gemini-code-assist bot commented Nov 18, 2025

Summary of Changes

Highlights

Footnotes

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

Uh oh!

Uh oh!

chenyushuo commented Nov 18, 2025

Uh oh!

github-actions bot commented Nov 18, 2025

Summary

Skipped

Tests

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Add `rope_scaling` and `rope_theta` to config #390

Add `rope_scaling` and `rope_theta` to config #390