kang sheng bd756c15c8 [BREAKING][rollout] feat: allow users pass all vllm/sglang engine args (#3037) ...

This PR allows users to pass all vllm/sglang engine args and optimizes
qwen3 rollout speed through vllm Engine argument.

1. deprecate the default value of previous engine_kwargs
2. pass all the engine_kwargs to vllm/sglang engine
3. optimize Qwen3-235B rollout speed by setting TP=8 and enabling expert
parallel.

From top to bottom: tp=16 without EP, tp=8 without EP and tp=8 with EP.
<img width="1000" height="808" alt="image"
src="https://github.com/user-attachments/assets/6b096be4-3896-4e96-8916-d8d6e13a58cc"
/>

PS: The DeepSeek-V3's rollout slows down after enabling expert
parallelism.

2025-08-14 19:12:26 +08:00

..

config.rst

[BREAKING][rollout] feat: allow users pass all vllm/sglang engine args (#3037)

2025-08-14 19:12:26 +08:00

gsm8k_example.rst

[doc] fix: quickstart example can't work on zsh (#2509)

2025-07-14 13:26:32 +08:00

multi_modal_example.rst

[doc] fix: add time info for each doc, assert sphinx warning in CI (#2255)

2025-06-29 11:58:35 +08:00

ppo_code_architecture.rst

[trainer] fix: Allow FSDP2 when doing strategy check (#2497)

2025-07-12 16:31:31 -07:00

sandbox_fusion_example.rst

[doc] fix: add time info for each doc, assert sphinx warning in CI (#2255)

2025-06-29 11:58:35 +08:00