From: Eric Curtin <redacted>
Date: Thu, 4 Sep 2025 09:49:44 +0000 (+0100)
Subject: Document the new max GPU layers default in help (#15771)
X-Git-Tag: upstream/0.0.6527~148
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=badb80cadbc40e047b30c43611aba575fc8d6845;p=pkg%2Fggml%2Fsources%2Fllama.cpp

Document the new max GPU layers default in help (#15771)

This is a key change, just letting users know.

Signed-off-by: Eric Curtin <redacted>
---

diff --git a/common/arg.cpp b/common/arg.cpp
index fcee0c44..7507c811 100644
--- a/common/arg.cpp
+++ b/common/arg.cpp
@@ -2466,7 +2466,7 @@ common_params_context common_params_parser_init(common_params & params, llama_ex
     ).set_examples({LLAMA_EXAMPLE_SPECULATIVE, LLAMA_EXAMPLE_SERVER}).set_env("LLAMA_ARG_N_CPU_MOE_DRAFT"));
     add_opt(common_arg(
         {"-ngl", "--gpu-layers", "--n-gpu-layers"}, "N",
-        "number of layers to store in VRAM",
+        string_format("max. number of layers to store in VRAM (default: %d)", params.n_gpu_layers),
         [](common_params & params, int value) {
             params.n_gpu_layers = value;
             if (!llama_supports_gpu_offload()) {