The workflow runs on the KNIME platform of the macbook pro intel cpu with errors. The same workflow runs successfully on Windows operating system platforms. The embedding model (nomic-embed-text-v1.5.f16.gguf) used is also exactly the same.
Hey there,
not a lot to go off - can you provide some more information - error logs, screenshots…
Some initial thoughts: how are you running the embedding model and connecting to it? GPT4All? Ollama? Another inference service that is OpenAI API compatible?
Given that you talk about Mac and that Gen AI extension depends on Python you may want to take a look at this announcement:
2 Likes
Hi there,
I think in GPT4All Embeddings Connector you manually have to enter the path correct?
I am fairly sure that the way it is referenced is incorrect - that said I am not familiar with Mac at all.
What you can try and do:
use this node to select your file:
Then pass in the path via the generated flow variables into the node:
1 Like
@MRAZ besides the already mentioned issue you can try this example. Best to download the whole workflow group and try to give the paths to the GPT4All models (the .gguf files) and select one suitable for embedding like “all-minilm-l6-v2_f16.gguf”.
1 Like