From: Christian Demsar <redacted>
Date: Sun, 23 Jul 2023 11:56:34 +0000 (-0400)
Subject: llama : print max tensor size to stderr (#2336)
X-Git-Tag: gguf-v0.4.0~421
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=a940458e4814e87bd0d3fbdb3f3d2733b4a3ccb1;p=pkg%2Fggml%2Fsources%2Fllama.cpp

llama : print max tensor size to stderr (#2336)
---

diff --git a/llama.cpp b/llama.cpp
index 135aa9fe..0731c75a 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -2795,7 +2795,7 @@ struct llama_context * llama_new_context_with_model(
 
         const size_t max_size = ggml_get_max_tensor_size(ctx->model.ctx);
 
-        printf("%s: max tensor size = %8.2f MB\n", __func__, max_size/1024.0/1024.0);
+        fprintf(stderr, "%s: max tensor size = %8.2f MB\n", __func__, max_size/1024.0/1024.0);
 
 #define LLAMA_METAL_CHECK_BUF(result)                                          \
     if (!(result)) {                                                           \