Vaibhav JindalandVaibhav Jindal
809baf44ed
Move .claude/skills to .agents/skills as the canonical (vendor-neutral) location, with .claude/skills now a symlink so Claude Code still auto-discovers them. This mirrors the pattern NVIDIA TileGym uses. Skill content updates: - SKILL.md stage intros rephrased from "Spawn an X agent" to "Follow the X workflow in <file>", which describes the workflow without mandating any single runtime's subagent mechanism. A note explains that runtimes with parallel subagent support may still delegate. - finalizer.md commit/PR templates no longer hardcode a Claude-specific Co-Authored-By trailer or "Generated with Claude Code" footer; runtime attribution is left to the host runtime's own convention. Adds a root-level AGENTS.md (a convention picked up by Codex, Cursor, Aider, Gemini CLI, etc.) pointing readers to .agents/skills/ and summarising the three skills. ## Summary <!--- This is a required section; please describe the main purpose of this proposed code change. ---> <!--- ## Details This is an optional section; is there anything specific that reviewers should be aware of? ---> ## Testing Done <!--- This is a required section; please describe how this change was tested. ---> <!-- Replace BLANK with your device type. For example, A100-80G-PCIe Complete the following tasks before sending your PR, and replace `[ ]` with `[x]` to indicate you have done them. --> - Hardware Type: <BLANK> - [ ] run `make test` to ensure correctness - [ ] run `make checkstyle` to ensure code style - [ ] run `make test-convergence` to ensure convergence Co-authored-by: Vaibhav Jindal <[email protected]>
2.3 KiB
AGENTS.md
Guidance for AI coding assistants working in this repository.
Skills
Repository-specific workflow guides ("skills") live in .agents/skills/. Each subdirectory is a self-contained guide for a multi-stage workflow. Read the SKILL.md at the top of each subdirectory for an overview; the other files in that directory are referenced from SKILL.md and should be read on demand.
| Skill | What it does |
|---|---|
liger-kernel-dev |
Develops new Triton kernels from a PyTorch reference (or modifies existing kernels). 3-stage pipeline: Analyze → Generate → Validate. NVIDIA GPUs only. |
liger-autopatch |
Adds Liger Kernel support for a new HuggingFace Transformers model, or modifies an existing monkey-patch. 3-stage pipeline: Analyze → Generate → Validate. |
liger-kernel-perf |
Optimizes the performance of an existing Liger Triton kernel. 3-stage pipeline: Profile → Optimize → Finalize. NVIDIA GPUs only. |
The skills are written to be runtime-agnostic — they describe the workflow as a sequence of stages a competent agent (or human) can follow. Where a stage says "Follow the X workflow in x.md", that's a directive to read and execute that file's instructions; runtimes that support parallel subagents may delegate the stage, but it is not required.
Vendor-specific shortcuts
For convenience, some assistants auto-discover skills from vendor-specific paths. These point at the canonical .agents/skills/ directory:
.claude/skills→ symlink →.agents/skills(for Claude Code)
If you're adding support for another assistant, add a symlink (or your tool's preferred adapter) pointing to .agents/skills/. Do not duplicate the content.
Repo conventions
- Source layout:
src/liger_kernel/{ops,transformers}/for Triton ops andnn.Module/ HF wrappers respectively - Tests:
test/transformers/(unit) andtest/convergence/{bf16,fp32}/(model convergence) - Benchmarks:
benchmark/scripts/(scripts) andbenchmark/data/all_benchmark_data.csv(results) - Lint/format:
make checkstyle(usesruff) - Install dev mode:
pip install -e ".[dev]"
See README.md for the project overview and contribution guide.