From 6daa69ee81851ab26ca8aefca1a4202941fc0262 Mon Sep 17 00:00:00 2001 From: Jared Van Bortel Date: Mon, 29 Jan 2024 17:11:27 -0500 Subject: [PATCH] kompute : fix fallback to CPU (#5201) --- llama.cpp | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/llama.cpp b/llama.cpp index 9631506c6..796aaa895 100644 --- a/llama.cpp +++ b/llama.cpp @@ -4136,7 +4136,7 @@ static int llama_model_load(const std::string & fname, llama_model & model, llam } #ifdef GGML_USE_KOMPUTE - if (ggml_vk_has_device() && params.n_gpu_layers > 0 && ( + if (params.n_gpu_layers > 0 && ( !(model.arch == LLM_ARCH_LLAMA || model.arch == LLM_ARCH_FALCON) || !( model.ftype == LLAMA_FTYPE_ALL_F32 || @@ -4145,8 +4145,8 @@ static int llama_model_load(const std::string & fname, llama_model & model, llam model.ftype == LLAMA_FTYPE_MOSTLY_Q4_1 ) )) { - // disable Vulkan due to unsupported model architecture or quantization type // TODO(cebtenzzre): propagate this error outside of llama_load_model_from_file + LLAMA_LOG_WARN("%s: disabling Kompute due to unsupported model arch or quantization\n", __func__); params.n_gpu_layers = 0; } #endif