From 2d74714535e99e57add2aecbf7609f0a5cc209f1 Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Sat, 27 Jul 2024 13:32:44 +0300 Subject: [PATCH] llama : disable 405B max_nodes path due to lack of complaints ggml-ci --- src/llama.cpp | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/src/llama.cpp b/src/llama.cpp index 2f000cb5a..8a557b5b2 100644 --- a/src/llama.cpp +++ b/src/llama.cpp @@ -3657,10 +3657,10 @@ namespace GGUFMeta { using llama_buf_map = std::unordered_map; // TODO: update when needed or think of some clever automatic way to do this -static size_t llama_model_max_nodes(const llama_model & model) { - if (model.arch == LLM_ARCH_LLAMA && model.hparams.n_layer > 400) { // llama-3 405B - return 32768; - } +static size_t llama_model_max_nodes(const llama_model & /*model*/) { + //if (model.arch == LLM_ARCH_LLAMA && model.hparams.n_layer > ??) { // llama-3 405B + // return 32768; + //} return 8192; }