skills.rest
pradeepmouli
pradeepmouliCOMMUNITY
Explore all skills by this author →
GitHub

axiom-ios-ml

Convert PyTorch models to CoreML and optimize for iOS/macOS deployment.

Updated Dec 23, 2025
Search Tags:#ios#macos#model-conversion#coreml#on-device-ml#kv-cache#mltensor
One-click install
npx skills add https://github.com/pradeepmouli/swift-template --skill axiom-ios-ml-pradeepmouli
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: axiom-ios-ml
Source: https://github.com/pradeepmouli/swift-template/tree/main/.agents/skills/axiom-ios-ml
Command: npx skills add https://github.com/pradeepmouli/swift-template --skill axiom-ios-ml-pradeepmouli

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill provides a cohesive on-device machine learning workflow for iOS/macOS, enabling developers to convert models to CoreML, compress weights, and manage stateful KV-cache with multi-function support.

Core Features & Use Cases

  • CoreML model conversion and deployment across Apple platforms.
  • Stateful models with KV-cache for LLMs and multi-function adapters.
  • MLTensor-based pipeline stitching for modular AI apps.
  • Guidance on deployment targets, performance profiling, and safe concurrency.

Quick Start

Convert a PyTorch model to CoreML and run a quick on-device test.

Frequently Asked Questions about axiom-ios-ml

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I convert a PyTorch model to CoreML for on-device inference?▼

To convert a PyTorch model to CoreML, you use CoreML tooling to translate the model architecture and weights into a deployable format, then run an on-device test to verify inference and performance on iOS or macOS.

How does KV-cache work for stateful LLMs in CoreML?▼

KV-cache in CoreML enables stateful models by storing key-value tensors during inference, allowing multi-function adapters and LLMs to manage context efficiently without recomputing previous tokens on iOS or macOS.

Can I use MLTensor to stitch multiple ML models into a single iOS pipeline?▼

Yes, MLTensor-based pipeline stitching allows you to connect modular AI components and multi-function adapters, enabling cohesive on-device ML workflows across iOS and macOS applications.

What is the best way to optimize model compression and performance profiling for macOS deployment?▼

Optimizing model compression involves reducing weight sizes using CoreML tooling, while performance profiling ensures efficient on-device inference by evaluating deployment targets and safe concurrency on macOS.

Do I need safe concurrency and multi-function adapters for real-world app pipelines?▼

Yes, safe concurrency and multi-function adapters are required to handle stateful KV-cache and modular AI operations efficiently within real-world app pipelines on iOS and macOS.

Related Skills

Explore More Software Engineering Skills→
tuliopc23tuliopc23

axiom-ios-ml

Convert PyTorch and TensorFlow models to CoreML for on-device iOS inference.

Community
Advanced
pradeepmoulipradeepmouli

coreml-diag

Diagnose and remediate CoreML model load, performance, memory, and accuracy issues.

Community
Advanced
DFly7DFly7

coreml

Integrate Core ML models into iOS apps for on-device inference.

Community
Advanced
melissa-pereira-deelmelissa-pereira-deel

On-Device ML Optimization

Optimize on-device ML models for speed, size, and energy efficiency.

Community
Advanced