mllm-structure-style
Skill1 starsChanged 4 months agoNo licenceView on GitHub
MLLM 推理工程架构、评估接入、Hydra 配置工程、既有项目迁移整理与代码风格基础规范。Use when Codex needs to initialize an MLLM inference/evaluation project, scaffold a Hydra + torchrun + strategy-method codebase, migrate an existing unstructured MLLM project into the template layout, design Lightning-Hydra-style config groups without Trainer, integrate VLMEvalKit via adapters, add a new multimodal inference algorithm, or review/refactor MLLM inference/evaluation code for project structure, module boundaries, distributed routing, device placement, experiment logging, and style compliance.
In this repository27 other files it gives its agents.
- brightdataSkill
- cavemanSkill
- codegraph-project-readerSkill
- codex2codexSkill
- codex-subagent-plan2doSkill
- context-engineeringSkill
- debugging-and-error-recoverySkill
- debug-skillSkill
- deprecation-and-migrationSkill
- frontend-design-skillSkill
- git-workflow-and-versioningSkill
- gpu-schedulerSkill
- huggingface-local-cacheSkill
- idea-refineSkill
- interview-meSkill
- literature-ragSkill
- plan2doSkill
- ponytailSkill
- reviewerSkill
- skill-routerSkill
- skill-tokenlessSkill
- spec2planSkill
- structural-editSkill
- teachSkill
- test-driven-developmentSkill
- tool-health-checkSkill
- uv-mirror-envSkill
The file
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.