Linked Presentation: Parrot: Efficient Serving of LLM-based Applications with Semantic VariableLadder: Enabling Efficient Low-Precision Deep Learning Computing through Hardware-aware Tensor Transformation
