Guide edge ML deployment with quantization, pruning, and runtime selection.
npx skills add https://github.com/infantesromeroadrian/arca-claude-code --skill edge-ml
Please help me install this Agent Skill. Skill: edge-ml Source: https://github.com/infantesromeroadrian/arca-claude-code/tree/main/skills/edge-ml Command: npx skills add https://github.com/infantesromeroadrian/arca-claude-code --skill edge-ml
Edge ML deployment requires systematic optimization and suitable runtimes for on-device inference across mobile, IoT, and embedded systems.
Guide me through converting a PyTorch model to a mobile-optimized edge-ML pipeline.
High-intent search queries and answers about installing and using this skill.
To convert a PyTorch model to edge ML, apply model optimization techniques like quantization and pruning, then convert and profile it for mobile runtimes like TensorFlow Lite, Core ML, or ONNX Runtime Mobile to execute on-device inference.
The best way to reduce model size and latency for on-device inference is applying quantization, pruning, and distillation during model optimization, achieving substantial size reduction and latency improvements before deployment.
Yes, this edge ML workflow supports deployment on IoT systems and embedded platforms by providing platform-aware conversions and profiling steps tailored for on-device inference across mobile and embedded devices.
Knowledge distillation for edge ML optimizes models by transferring knowledge from a larger model to a smaller one, reducing model size and improving inference latency while maintaining accuracy for on-device deployment.
Choosing between TensorFlow Lite and ONNX Runtime Mobile for edge deployment depends on your target platform and model format, requiring platform-aware runtime selection and profiling to ensure optimal on-device inference performance.
gracefullightExport, quantize, and serve TensorFlow models via SavedModel, TFLite, and TensorFlow Serving.
pytorchProfile ExecuTorch model execution with etdump and Inspector for operator timing.
jstzwjQuantize weights and optimizer states for PyTorch model training.
datathingsReference the llama.cpp C API for model loading and inference workflows.
