Split group of tasks to multiple runners #116

pan-x-c · 2025-07-10T06:47:30Z

Description

Add explorer.max_repeat_times, which limits the number of times a single runner can repeat a task, thereby splitting tasks with larger repeat_times into multiple subtasks to be executed on different runners. When explorer.max_repeat_times is not set, no splitting will be performed.

This method can reduce the overall latency of tasks with large repeat_times and lower the cost of failed retries; however, it will also cause the rollout results of the same group to be scattered. Users must decide whether to enable this function based on their actual needs.

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

pan-x-c · 2025-07-10T06:53:33Z

/run-unittest

github-actions · 2025-07-10T07:13:51Z

Summary

Tests 📝	Passed ✅	Failed ❌	Skipped ⏭️	Other ❓	Flaky 🍂	Duration ⏱️
45	44	1	0	0	0	1.2s

Failed Tests

Failed Tests ❌	Fail Message
❌ tests/buffer/queue_test.py::TestQueueBuffer::test_queue_buffer	The test failed in the call phase

Tests

Test Name	Status	Duration
tests/algorithm/policy_loss_test.py::VerlPolicyLossTest::test_dpo_policy_loss	✅	1ms
tests/algorithm/policy_loss_test.py::VerlPolicyLossTest::test_mix_policy_loss	✅	1ms
tests/algorithm/policy_loss_test.py::VerlPolicyLossTest::test_opmd_policy_loss	✅	1ms
tests/algorithm/policy_loss_test.py::VerlPolicyLossTest::test_ppo_policy_loss	✅	1ms
tests/algorithm/policy_loss_test.py::VerlPolicyLossTest::test_sft_policy_loss	✅	1ms
tests/buffer/file_test.py::TestFileBuffer::test_file_buffer	✅	3ms
tests/buffer/file_test.py::TestFileBuffer::test_file_reader	✅	1ms
tests/buffer/file_test.py::TestFileBuffer::test_file_writer	✅	2ms
tests/buffer/queue_test.py::TestQueueBuffer::test_queue_buffer	❌	5ms
tests/buffer/sql_test.py::TestSQLBuffer::test_create_sql_buffer	✅	4ms
tests/common/config_test.py::TestConfig::test_all_examples_are_valid	✅	1ms
tests/common/config_test.py::TestConfig::test_load_default_config	✅	4ms
tests/common/experience_test.py::TestExperienceConversion::test_batch_conversion	✅	1ms
tests/common/experience_test.py::TestExperienceConversion::test_experience_model_experience_conversion	✅	1ms
tests/common/vllm_test.py::TestModelWrapperSyncV0::test_generate	✅	42ms
tests/common/vllm_test.py::TestModelWrapperAsyncV0::test_generate	✅	41ms
tests/common/vllm_test.py::TestModelWrapperAsyncTPV0::test_generate	✅	51ms
tests/common/vllm_test.py::TestModelWrapperAsyncTPV1::test_generate	✅	52ms
tests/common/vllm_test.py::TestModelWrapperAsyncV1::test_generate	✅	39ms
tests/common/vllm_test.py::TestAPIServer::test_api	✅	24ms
tests/common/vllm_test.py::TestTokenizer::test_assistant_token_mask	✅	1ms
tests/explorer/explorer_test.py::BaseExplorerCase::test_explorer	✅	1ms
tests/explorer/explorer_test.py::TestExplorerCountdownEval::test_explorer	✅	84ms
tests/explorer/explorer_test.py::TestExplorerCountdownNoEval::test_explorer	✅	99ms
tests/explorer/scheduler_test.py::SchedulerTest::test_concurrent_operations	✅	4ms
tests/explorer/scheduler_test.py::SchedulerTest::test_get_results	✅	20ms
tests/explorer/scheduler_test.py::SchedulerTest::test_scheduler_all_methods	✅	14ms
tests/explorer/scheduler_test.py::SchedulerTest::test_scheduler_restart_after_stop	✅	8ms
tests/explorer/scheduler_test.py::SchedulerTest::test_split_tasks	✅	7ms
tests/explorer/scheduler_test.py::SchedulerTest::test_wait_all	✅	7ms
tests/explorer/scheduler_test.py::SchedulerTest::test_wait_all_timeout_with_multi_batch	✅	13ms
tests/explorer/workflow_test.py::WorkflowTest::test_gsm8k_workflow	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_boxed_workflow	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_complex_workflow	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_fraction_workflow	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_workflow	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_workflow_resettable	✅	1ms
tests/trainer/trainer_test.py::BaseTrainerCase::test_trainer	✅	1ms
tests/trainer/trainer_test.py::TestTrainerCountdown::test_trainer	✅	253ms
tests/trainer/trainer_test.py::TestStepAheadAsyncRL::test_trainer	✅	83ms
tests/trainer/trainer_test.py::TestTrainerGSM8K::test_trainer	✅	60ms
tests/trainer/trainer_test.py::TestTrainerSFTWarmupGSM8K::test_trainer	✅	104ms
tests/trainer/trainer_test.py::TestTrainerDPO::test_trainer	✅	42ms
tests/trainer/trainer_test.py::TestFullyAsyncMode::test_fully_async_mode	✅	93ms
tests/utils/plugin_test.py::TestPluginLoader::test_load_plugins	✅	4ms

Github Test Reporter by CTRF 💚

trinity/common/config.py

pan-x-c · 2025-07-10T09:00:16Z

/run-unittest

github-actions · 2025-07-10T09:20:01Z

Summary

Tests 📝	Passed ✅	Failed ❌	Skipped ⏭️	Other ❓	Flaky 🍂	Duration ⏱️
45	45	0	0	0	0	1.1s

Tests

Test Name	Status	Duration
tests/algorithm/policy_loss_test.py::VerlPolicyLossTest::test_dpo_policy_loss	✅	1ms
tests/algorithm/policy_loss_test.py::VerlPolicyLossTest::test_mix_policy_loss	✅	1ms
tests/algorithm/policy_loss_test.py::VerlPolicyLossTest::test_opmd_policy_loss	✅	1ms
tests/algorithm/policy_loss_test.py::VerlPolicyLossTest::test_ppo_policy_loss	✅	1ms
tests/algorithm/policy_loss_test.py::VerlPolicyLossTest::test_sft_policy_loss	✅	1ms
tests/buffer/file_test.py::TestFileBuffer::test_file_buffer	✅	3ms
tests/buffer/file_test.py::TestFileBuffer::test_file_reader	✅	1ms
tests/buffer/file_test.py::TestFileBuffer::test_file_writer	✅	1ms
tests/buffer/queue_test.py::TestQueueBuffer::test_queue_buffer	✅	9ms
tests/buffer/sql_test.py::TestSQLBuffer::test_create_sql_buffer	✅	4ms
tests/common/config_test.py::TestConfig::test_all_examples_are_valid	✅	1ms
tests/common/config_test.py::TestConfig::test_load_default_config	✅	4ms
tests/common/experience_test.py::TestExperienceConversion::test_batch_conversion	✅	1ms
tests/common/experience_test.py::TestExperienceConversion::test_experience_model_experience_conversion	✅	1ms
tests/common/vllm_test.py::TestModelWrapperSyncV0::test_generate	✅	43ms
tests/common/vllm_test.py::TestModelWrapperAsyncV0::test_generate	✅	40ms
tests/common/vllm_test.py::TestModelWrapperAsyncTPV0::test_generate	✅	51ms
tests/common/vllm_test.py::TestModelWrapperAsyncTPV1::test_generate	✅	51ms
tests/common/vllm_test.py::TestModelWrapperAsyncV1::test_generate	✅	38ms
tests/common/vllm_test.py::TestAPIServer::test_api	✅	24ms
tests/common/vllm_test.py::TestTokenizer::test_assistant_token_mask	✅	1ms
tests/explorer/explorer_test.py::BaseExplorerCase::test_explorer	✅	1ms
tests/explorer/explorer_test.py::TestExplorerCountdownEval::test_explorer	✅	89ms
tests/explorer/explorer_test.py::TestExplorerCountdownNoEval::test_explorer	✅	84ms
tests/explorer/scheduler_test.py::SchedulerTest::test_concurrent_operations	✅	4ms
tests/explorer/scheduler_test.py::SchedulerTest::test_get_results	✅	19ms
tests/explorer/scheduler_test.py::SchedulerTest::test_scheduler_all_methods	✅	14ms
tests/explorer/scheduler_test.py::SchedulerTest::test_scheduler_restart_after_stop	✅	8ms
tests/explorer/scheduler_test.py::SchedulerTest::test_split_tasks	✅	7ms
tests/explorer/scheduler_test.py::SchedulerTest::test_wait_all	✅	7ms
tests/explorer/scheduler_test.py::SchedulerTest::test_wait_all_timeout_with_multi_batch	✅	13ms
tests/explorer/workflow_test.py::WorkflowTest::test_gsm8k_workflow	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_boxed_workflow	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_complex_workflow	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_fraction_workflow	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_math_workflow	✅	1ms
tests/explorer/workflow_test.py::WorkflowTest::test_workflow_resettable	✅	1ms
tests/trainer/trainer_test.py::BaseTrainerCase::test_trainer	✅	1ms
tests/trainer/trainer_test.py::TestTrainerCountdown::test_trainer	✅	242ms
tests/trainer/trainer_test.py::TestStepAheadAsyncRL::test_trainer	✅	96ms
tests/trainer/trainer_test.py::TestTrainerGSM8K::test_trainer	✅	56ms
tests/trainer/trainer_test.py::TestTrainerSFTWarmupGSM8K::test_trainer	✅	81ms
tests/trainer/trainer_test.py::TestTrainerDPO::test_trainer	✅	44ms
tests/trainer/trainer_test.py::TestFullyAsyncMode::test_fully_async_mode	✅	88ms
tests/utils/plugin_test.py::TestPluginLoader::test_load_plugins	✅	4ms

Github Test Reporter by CTRF 💚

support split tasks

998056d

fix tests

3f0aa4c

yanxi-chen reviewed Jul 10, 2025

View reviewed changes

trinity/common/config.py Outdated Show resolved Hide resolved

fix comments

fa2ec9d

chenyushuo approved these changes Jul 10, 2025

View reviewed changes

pan-x-c merged commit 73e8d8f into modelscope:main Jul 10, 2025
2 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Split group of tasks to multiple runners #116

Split group of tasks to multiple runners #116

Uh oh!

pan-x-c commented Jul 10, 2025 •

edited

Loading

Uh oh!

pan-x-c commented Jul 10, 2025

Uh oh!

github-actions bot commented Jul 10, 2025

Uh oh!

Uh oh!

pan-x-c commented Jul 10, 2025

Uh oh!

github-actions bot commented Jul 10, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Split group of tasks to multiple runners #116

Split group of tasks to multiple runners #116

Uh oh!

Conversation

pan-x-c commented Jul 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Checklist

Uh oh!

pan-x-c commented Jul 10, 2025

Uh oh!

github-actions bot commented Jul 10, 2025

Summary

Failed Tests

Tests

Uh oh!

Uh oh!

pan-x-c commented Jul 10, 2025

Uh oh!

github-actions bot commented Jul 10, 2025

Summary

Tests

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

pan-x-c commented Jul 10, 2025 •

edited

Loading