VMAFx
Official@vmafx
VMAFX — modernized perceptual video quality assessment. Cloud-native fork of Netflix VMAF with GPU backends, tiny-AI models, k8s-native deployment.
Agent Skills by VMAFx
Showing 28 vetted skills indexed across 1 GitHub repositories.
VMAFxlint-all
Run multiple static analysis tools and merge findings into one report.
VMAFxformat-all
Format C, C++, CUDA, Python, and shell files across a repository.
VMAFxvalidate-scores
Compare reference and distorted inputs across scoring backends and report pairwise numerical deltas.
VMAFxadd-feature-extractor
Scaffold a new libvmaf feature extractor with source, header, registry, and smoke test files.
VMAFxbuild-vmaf
Build libvmaf with Meson and Ninja across cpu, cuda, sycl, and hip backends.
VMAFxsync-upstream
Reconcile a forked Git repository with its upstream branch.
VMAFxadd-simd-path
Scaffolds SIMD implementations for features with runtime dispatch and bit-exact tests.
VMAFxaudit-modernization
Run a read-only modernization audit and write a dated Markdown report to /tmp.
VMAFxport-upstream-commit
Cherry-pick an upstream commit onto a fork's master branch with provenance.
VMAFxbisect-regression
Run git bisect to find the first bad commit using a named test predicate.
VMAFxadd-mcp-tool
Scaffold new MCP tools with matched handlers across VMAFx Go and Python servers.
VMAFxadd-gpu-backend
Scaffold a GPU backend package for libvmaf with build wiring and smoke tests.
VMAFxcross-backend-diff
Detect cross-backend VMAF score drift using per-frame JSON and ULP distance.
VMAFxprep-release
Dry-run release-please to preview version, changelog, and blockers.
VMAFxrun-netflix-bench
Run the Netflix benchmark suite and compare results against the committed baseline.
VMAFxadd-k8s-resource
Scaffolds Kubernetes CRD resources and operator wiring for the VMAFx project.
VMAFxadd-model
Validate and register VMAF model artifacts into build targets and smoke tests.
VMAFxai-run-manifest
Add replay-manifest sidecars to AI-generated artifacts with run provenance.
VMAFxbuild-ffmpeg-with-vmaf
Build FFmpeg against a local libvmaf checkout and run smoke tests.
VMAFxdev-llm-commitmsg
Draft Conventional Commits messages from staged Git changes.
VMAFxrefresh-ffmpeg-patches
Rebase an ffmpeg patch series onto upstream changes and regenerate clean patches.
VMAFxregen-docs
Rebuild mkdocs-material sites and detect broken links and ADR drift.
VMAFxdev-llm-docgen
Generate Doxygen docblocks for C, C++, and CUDA functions.
VMAFxregen-snapshots
Regenerate snapshot JSON files for CPU, SYCL, and Netflix benchmarks.
Frequently Asked Questions About VMAFx
FAQPage SchemaWhat specific tasks does VMAFx enable for video engineers?▼
VMAFx enables precise perceptual video quality assessment by providing infrastructure to build libvmaf with multiple GPU backends, validate model quality against statistical gates, and perform cross-backend drift detection using per-frame numerical analysis.
Which technical personas benefit from using VMAFx?▼
VMAFx is designed for video codec engineers, performance optimization specialists, and infrastructure developers who require high-fidelity quality metrics and need to deploy video processing workloads within containerized environments.
What are the primary dependencies for running VMAFx?▼
VMAFx requires a build environment configured with Meson and Ninja, alongside specific GPU drivers for CUDA, SYCL, or HIP backends. Users must also have a compatible FFmpeg checkout and the necessary runtime libraries for ONNX model execution.