This is related to running TokenSmith on Intel Mac's. Current issue was mentioned by @karvyq regarding an issue with PyTorch version deprecating Intel mac support but HF/sentence_transformers requiring it.
Fix would be switching reranker to LLAMA.cpp too and remove ST/HF requirement altogether.
This is related to running TokenSmith on Intel Mac's. Current issue was mentioned by @karvyq regarding an issue with PyTorch version deprecating Intel mac support but HF/sentence_transformers requiring it.
Fix would be switching reranker to LLAMA.cpp too and remove ST/HF requirement altogether.