This project is a TypeScript + Plotly frontend dashboard for comparing LLM profiling results across open-source models, hardware targets, and software/runtime builds. It includes interactive filters, KPI cards, and multiple visualizations (scatter, bar, box, and heatmap) to help developers quickly inspect throughput, latency, memory use, and efficiency across benchmark runs.
Screen.Recording.2026-03-01.at.7.48.41.PM.mov
Use npm install first to install all dependencies from package.json into node_modules. Run this command after cloning the repo and anytime dependencies change so the project has the exact libraries it needs for local development and builds.
Use npm run dev to start the Vite development server for live local testing, and use npm run build to generate a production build after your changes are ready. The dev command is for fast iteration with hot reload, while the build command verifies TypeScript and outputs optimized static assets for deployment.