LLM Inference Performance Estimator

Estimate LLM inference performance metrics including TTFT, decode speed, and VRAM requirements based on model architecture, GPU specs, and quantization format.