Codendum, shared generative coding on local infrastructure
The release of Codendum 0.1.0 under the Apache 2.0 licence, a service for sharing one inference machine running local open-weight models for generative coding. How it is built, how to size the memory and what I measured under load.
Read more