I once again play around with the GPT4ALL Local Models and the workflow from here (Creating a Local LLM Vector Store from PDFs with KNIME and GPT4All | by Markus Lauber | Low Code for Data Science | Medium)
It worked fine in the past but now the FAISS Vector store creator does not accept any Models from GPT4ALL except one. Even the Models which where working in the past are not working.
Error: Failed to generate Embeddings. Not an Embedding Model. I Cannot figure out why this happens.
@gator_ali welcome to the KNIME forum. I also see some GPT4All models failing. What does work on macOS is to tell the embeddings connector to download a model. One older model does work but some others do not. I have no idea if they changed something in the formats (again).
I can confirm on Windows. The download works.
As well as the mentioned all-mini model as local.
As this is a F16 quantization model i suspected this first. Which is not the case. Other embedding F16 models also fail.
I wanted to compare the performance/results of the models which I have to postpone it seems.
Thanks for the quick reply.
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
