📈 GH Trending Explorer 🎲 Surprise me

Search: patch · Language: HTML · clear

patchy631 / time-to-first-token

A 10-week, 30-minutes-a-day roadmap for LLM inference serving and optimization. vLLM, SGLang, quantization, speculative decoding, benchmarking.

▲ ~0/7d ★ 997 HTML 2mo ago
marcosblg / rackscope

Document your rack cabling in a single HTML file. Click any port to see exactly where it goes. No server, no dependencies, works offline.