New
microsoft/llmlingua v0.2.2 — Prompt and KV-Cache Compression
The LLMLingua Series is a project that removes unnecessary tokens from prompts using a small language model, providing compression of up to 20x with only small performance loss; the latest release is v0.2.2.