Hi,
There are several threads regarding the problems to get Ollama LLM and Embeddings to work with Knowledge covering problems with authentication referring to OpenAI and problems with initiating. All with different suggestions of workarounds and claimed solutions, but I have not been able to find any consensus that the problem is resolved.
Is this still an open problem? If not, what is the agreed solution to allow us relying on pdf-based sources to use CrewAI?
Thanks in advance,
Magnus