ChromaDB leaves.
Instead I will use the same approach on the current Nexus version as for the base version Cyberdeck and use an own most tiny database system being no more than a Class in a pytrhon script and .json file assembling the most basic and core version of a Database.
ChromaDB has heavy communication issues with Ollama making the core idea of parallel batch processing hard to impossible to achieve over a computer cluster.
To scale open source systems, such as my Cyberdeck Knowledge System from low cost hardware to large systems and architectures the Open Source Community needs an Apache grade database for Ollama architecture in order to fully utilize the reasoning capability of LLM models.
#cyberpunkcoltoure
