Whichlang: A fast language detection library for Rust
While the current trend is about smart but slow large language models (LLM), we hope you will find it refreshing to read a blog post about a tiny, stupid but fast model. Less is more, isn’t it?
While the current trend is about smart but slow large language models (LLM), we hope you will find it refreshing to read a blog post about a tiny, stupid but fast model. Less is more, isn’t it?
At Quickwit, we are building the most cost-efficient search engine for logs and traces. Such an engine typically ingests massive amounts of data while serving a comparatively low number of search queries. Under this workload, most of your CPU is spent on indexing, making our indexing pipeline a crucial component.