Commit aa4eed2
File tree
- assets/js
- docs
- DesignImplementation
- AgenticPipeline
- RLVRPipeline
- DevelopmentGuide
- customer_env_en
- prompt_intro_en
- support_new_models_en
- English
- QuickStart
- debug_guide
- image_address
- installation
- UserGuide
- agentic
- agentic_GiGPO
- agentic_StarPO
- algorithms
- GRPO
- GSPO
- LitePPO
- PPO
- RAFT_Plus_Plus
- Reinforce_Plus_Plus
- Reward_FL
- TOPR
- ascend/ascend_usage
- async_parallel_rollout
- backend
- deepspeed
- sglang
- vllm
- checkpoint_and_resume
- megatron_convert_2_hf
- start
- QuickStart
- config_guide
- config_system
- debug_guide
- image_address
- installation
- multi_nodes_quick_start
- qa_issues
- single_node_quick_start
- UserGuide
- agentic
- Tool_Use
- agentic_GiGPO
- agentic_StarPO
- agentic_engineer_practice
- algorithms
- GRPO
- GSPO
- LitePPO
- PPO
- RAFT_Plus_Plus
- Reinforce_Plus_Plus
- Reward_FL
- TOPR
- offpolicy_setting
- ascend/ascend_usage
- async_parallel_rollout
- async_training
- backend
- deepspeed
- fp8_rollout
- lora
- megatron
- sglang
- vllm
- checkpoint_and_resume
- device_mapping
- megatron_convert_2_hf
- offload_reload_control
- pipeline
- agent_pipeline_start
- agentic_pipeline_start
- distill_pipeline_start
- dpo_pipeline_start
- rlvr_pipeline_start
- vl_rlvr_pipeline_start
- start
- trackers_and_metrics
- 简体中文
- 使用指南
- agentic
- agentic_GiGPO
- agentic_StarPO
- algorithms
- GRPO
- GSPO
- LitePPO
- PPO
- RAFT_Plus_Plus
- Reinforce_Plus_Plus
- Reward_FL
- TOPR
- rl_options
- ascend/ascend_usage
- async_parallel_rollout
- async_training
- backend
- deepspeed
- lora
- megatron
- sglang
- vllm
- checkpoint_and_resume
- device_mapping
- megatron_convert_2_hf
- offload_reload_control
- pipeline
- agentic_pipeline_start
- distill_pipeline_start_cn
- dpo_pipeline_start
- sft_pipeline_start
- 快速开始
- debug_guide
- image_address
- installation
- multi_nodes_quick_start_cn
- qa_issues
- single_node_quick_start_cn
- 性能调优指南
- megatron_config_simple_guide
- roll_config_tune_guide
- 扩展开发手册
- custom_env_cn
- custom_pipeline
- custom_reward_cn
- custom_reward_worker
- multi_turn_env_manager
- prompt_intro_cn
- support_new_models_cn
- 设计实现文档
- agentic_async_traning_design
- agentic_rollout_env_manager
- resource_system
- rlvr_dynamic_sampling
- strategy_backend_design
- markdown-page
- search
- zh-Hans
- assets
- css
- images
- js
- docs
- DesignImplementation
- AgenticPipeline
- RLVRPipeline
- DevelopmentGuide
- customer_env_en
- prompt_intro_en
- support_new_models_en
- QuickStart
- config_guide
- config_system
- debug_guide
- image_address
- installation
- multi_nodes_quick_start
- qa_issues
- single_node_quick_start
- UserGuide
- agentic
- Tool_Use
- agentic_GiGPO
- agentic_StarPO
- agentic_engineer_practice
- algorithms
- GRPO
- GSPO
- LitePPO
- PPO
- RAFT_Plus_Plus
- Reinforce_Plus_Plus
- Reward_FL
- TOPR
- offpolicy_setting
- ascend/ascend_usage
- async_parallel_rollout
- async_training
- backend
- deepspeed
- fp8_rollout
- lora
- megatron
- sglang
- vllm
- checkpoint_and_resume
- device_mapping
- megatron_convert_2_hf
- offload_reload_control
- pipeline
- agent_pipeline_start
- agentic_pipeline_start
- distill_pipeline_start
- dpo_pipeline_start
- rlvr_pipeline_start
- vl_rlvr_pipeline_start
- start
- trackers_and_metrics
- img
- markdown-page
- search
Some content is hidden
Large Commits have some content hidden by default. Use the searchbox below for content that may be hidden.
Large diffs are not rendered by default.
This file was deleted.
This file was deleted.
This file was deleted.
0 commit comments