Skip to content
View LingYi-Liang's full-sized avatar
:shipit:
活下去，
:shipit:
活下去，
2 followers · 1 following

Achievements

Achievement: Pull Shark

Achievements

Achievement: Pull Shark

Block or report LingYi-Liang

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
LingYi-Liang/README.md

English切换到中文

LingYi — LLM reasoning & efficient learning · mathematics researcher

site mail

AI infra · full-stack: merged pull requests and the organisations I contribute to

langgenius/dify: refactor(api): separate console conversation variable queries bytedance/deer-flow: fix(channels): preserve inbound files on concurrent uploads bytedance/deer-flow: fix(agents): avoid executor starvation in file-gate waits bytedance/deer-flow: fix(tui): isolate callbacks from interrupted runs

Tencent/WeKnora: fix(docreader): preserve PPTX fallback image associations Tencent/WeKnora: fix(dataset): allow selecting the QA answer model Tencent-Hunyuan/UniRL: refactor(train): share stack lifecycle and update reporting apache/fory-site: Fix edit links for versioned English documentation

Pull requests in review (click to expand)

vllm-project/vllm: [Bugfix] Reserve sparse MLA indexer metadata before sizing the KV cache bytedance/deer-flow: fix(frontend): restore focus after fallback clipboard copy bytedance/deer-flow: fix(frontend): ignore message edit shortcuts during IME composition bytedance/deer-flow: test(frontend): account for native clipboard line endings bytedance/deer-flow: fix(frontend): align one-time DST conversion with the scheduler bytedance/deer-flow: fix(mcp): keep OAuth refresh waiters off the default executor bytedance/UI-TARS-desktop: fix(cli): prevent 401/403 errors in cached headless runs zai-org/Open-AutoGLM: fix: restore the input method when Type fails or is interrupted kvcache-ai/Mooncake: [Store] Persist invalid memory descriptor removals during restore kvcache-ai/Mooncake: [Store] Preserve pending task quota after terminal history pruning modelscope/ms-agent: fix(llm): merge streamed tool calls by index trpc-group/trpc-agent-go: {session, server/agui}: preserve same-time events during snapshot follow MiniMax-AI/MiniMax-MCP: Fix save failures caused by invalid filename characters llm-d/llm-d-router: fix: check projected request cost before flow-control release

Orbit I mathematics, orbit II my projects, orbit III the projects my pull requests went to

The past year of contributions as a starfield, swept by a small black hole

Until the next observation.

✦ AI use disclosure · I may use AI tools to assist my work, including code and pull requests. Everything is carefully reviewed and read by me before it goes out — I make sure I understand it, and I take responsibility for the result.

Pinned Loading

  1. squarefree-fuglede-lean squarefree-fuglede-lean Public

    Lean 4 formalization of square-free cyclic spectral-to-tiling and coprime prime-step descent, with explicit verification scope and reproducible checks.

    Lean