Detect WebNN support and initialize an MLContext for in-browser ML acceleration.
npx skills add https://github.com/webmaxru/agent-skills --skill webnn
Please help me install this Agent Skill. Skill: webnn Source: https://github.com/webmaxru/agent-skills/tree/main/skills/webnn Command: npx skills add https://github.com/webmaxru/agent-skills --skill webnn
WebNN enables in-browser machine learning with local acceleration, reducing reliance on server-side inference and enabling private, offline-capable ML in web apps.
Detect WebNN support and initialize a context with a simple graph to validate the setup.
High-intent search queries and answers about installing and using this skill.
To use WebNN for in-browser ML acceleration, detect navigator.ml support and initialize an MLContext within a secure context. You can configure context options like powerPreference and accelerated to optimize local inference.
WebNN is used for in-browser machine learning to enable local hardware acceleration. It reduces reliance on server-side inference by allowing web apps to load models, construct graphs with MLGraphBuilder, and execute tensor I/O directly on the client.
When WebNN is unavailable, you can fallback to non-WebNN runtimes to maintain browser inference capabilities. This ensures your web app continues functioning by dispatching tensors through alternative local execution paths.
Yes, you can reuse MLGraph and tensors during in-browser ML execution to optimize performance. The WebNN integration path supports graph construction, tensor write/read operations, and proper resource cleanup to prevent memory leaks.
Yes, WebNN requires a secure context for MLContext initialization to ensure safe in-browser machine learning. Enforcing secure contexts protects local model loading, tensor dispatch, and graph construction operations within the browser.
WebNN context loss can occur during local model loading due to system resource constraints. You should monitor context loss events and implement proper cleanup procedures to dispose of MLGraph and tensor resources when the MLContext is lost.
liminal-aiScaffold Convex projects and wire ConvexProvider with environment variables.
spacecat2002Diagnose CUDA crashes in SGLang by logging tensors at kernel API boundaries.
FaysilAlshareefCompose Polly v8 resilience pipelines for HTTP and non-HTTP .NET operations.
ClyptAIRun machine learning models in JavaScript with WebGPU/WASM acceleration.
