Create, transform, and optimize MLIR modules for AI frameworks and compiler backends.
npx skills add https://github.com/jstzwj/ai-infra-plugins --skill mlir
Please help me install this Agent Skill. Skill: mlir Source: https://github.com/jstzwj/ai-infra-plugins/tree/main/plugins/mlir/skills/mlir Command: npx skills add https://github.com/jstzwj/ai-infra-plugins --skill mlir
This Skill enables efficient manipulation and analysis of MLIR representations to simplify, optimize, and generate code for machine learning and hardware targets.
Initialize an MLIR context, construct or load IR modules, and apply transformation passes to prepare code for codegen, all through programmatic API commands.
High-intent search queries and answers about installing and using this skill.
Lower MLIR modules to LLVM IR by applying dialect conversion and lowering passes. This translates high-level operations into executable code for target-specific hardware compilation workflows.
Register and manage custom dialects in MLIR by initializing an MLIR context and using dialect APIs to define operations. This enables seamless IR manipulation within C++ projects.
Pattern rewriting optimizes MLIR representations by applying canonicalization and dialect conversion passes. This improves IR quality before lowering high-level modules to LLVM IR for codegen.
Automate ML model compilation pipelines by loading IR modules and applying transformation passes. This translates machine learning models into optimized, target-specific code for CPUs, GPUs, or accelerators.
You need LLVM project libraries and MLIR core dialect APIs to transform modules in C++. These dependencies provide the necessary environment for seamless IR management and code emitting workflows.
ROCmAutomate building and installing FlyDSL from source on remote hosts or containers.
jstzwjImport models from PyTorch, ONNX, and TFLite into Relax IR for optimization.
rainoftimeGenerate LLVM Intermediate Representation and build compiler backends.
copyleftdevDesign JIT compilers and optimize managed runtimes with advanced compiler techniques.
