Version: llamafile v0.8.1 (same issue since 0.7.x, earlier version 0.6.2 works perfectly fine)
This issues only exists if you use llamafile with external weights and with your own UI
Everything works perfectly fine of Intel Mac but on M1 it starts fine and then on localhost:8080 llama server simply renders 'File not found'. This custom UI is nothing more than just a simple index.html
Older llamafile version 0.6.2 with the exact command shown above works perfectly fine on Intel and M1.
Version: llamafile v0.8.1 (same issue since 0.7.x, earlier version 0.6.2 works perfectly fine) This issues only exists if you use llamafile with external weights and with your own UI
Eg: ./llamafile -m mistral.gguf -ngl 9999 --server --nobrowser -ngl 9999 --ctx-size 2048 --path ./ui
Everything works perfectly fine of Intel Mac but on M1 it starts fine and then on localhost:8080 llama server simply renders 'File not found'. This custom UI is nothing more than just a simple index.html
Older llamafile version 0.6.2 with the exact command shown above works perfectly fine on Intel and M1.