microsoft

/

Phi-3-mini-4k-instruct-gguf

like 387

There is an error for model file Phi-3-mini-4k-instruct-q4.gguf loaded with llama.cpp
ggml_metal_graph_compute_block_invoke: error: unsupported op 'RMS_NORM'
GGML_ASSERT: /Users/xiaoxingqiu/GitHub/llama.cpp/ggml-metal.m:918: !"unsupported op"