Inference Optimization

Deep dive into speculative sampling techniques for accelerating LLM inference through draft model prediction and rejection sampling.