Can't run
#1
by
bartowski
- opened
Even with llama.cpp master, running this errors. How did you make an imatrix? It fails in a similar way
It crashes with: ggml/src/ggml.c:6399: GGML_ASSERT(c->ne[0] >= n_dims / 2) failed
@NikolayKozloff please test your quants that you make with GGUF My Repo first. This doesn't work in llama.cpp/LM Studio/Ollama. Would suggest making the repo private.