AutoCompress: Automated Hardware-Guided Model Compression
Comments
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules
Summary
AutoCompress is a model-compression framework that uses differentiable architecture search to prune LLMs under explicit latency, throughput, memory, and parameter constraints measured on the target hardware and serving engine, rather than by parameter count alone.