Tag
A tweet compiling key resources for learning LLM inference, anchored by a detailed blog post on transformer inference arithmetic covering KV cache, model parallelism, latency calculations, batch sizing, and FLOPs counting with practical benchmarks.