benchmark-integrations
Guides RoboVerse benchmark metadata, dataset and demo conversion, and optional LIBERO, ManiSkill, MJLab, robosuite, RobotWin, and SimplerEnv integrations.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- benchmark-integrations
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Guides RoboVerse benchmark metadata, dataset and demo conversion, and optional LIBERO, ManiSkill, MJLab, robosuite, RobotWin, and SimplerEnv integrations.
- disable-model-invocation
- true
02 资源文件 3
- 参考文档references/data-formats.md1.4 KB
- 参考文档references/troubleshooting.md2.4 KB
- 参考文档references/workflows.md3.0 KB
03 SKILL.md 正文
Benchmark Integrations
Use this route when the task mentions benchmark task specs, dataset layouts, LIBERO, ManiSkill, MJLab, robosuite, RobotWin, SimplerEnv, passthrough, replay, demo conversion, or cross-framework evaluation.
Route
- Read workflows.md to select metadata-only, local-fixture conversion, replay, passthrough, or native integration.
- Read data-formats.md before converting or replaying data. Establish task/robot/simulator/camera/action identity and episode boundaries first.
- Treat each external stack as an optional backend with its own dependency, asset, license, data, display, GPU, and version gate. The package's CPU import does not verify a native integration.
- Run the tiny fixture or inventory path before any download, rollout, render, sweep, or policy evaluation. Record PASS, SKIP_NOT_SELECTED, or the actual blocker; never disguise missing assets as a test failure.
- Use troubleshooting.md for locator, format, passthrough, replay, and backend failures.
The BenchmarkTaskSpec metadata API is lightweight and robot-agnostic; use it
for planning even when a native benchmark backend is unavailable. Task changes
route to task-development; measured parity
routes to parity-and-tooling.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).