From: Georgi Gerganov <redacted>
Date: Fri, 21 Apr 2023 07:23:36 +0000 (+0300)
Subject: llama : fix comment for "output.weight" tensor
X-Git-Tag: gguf-v0.4.0~907
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=d40fded93e1a533e969768e1e335c15c61c296ce;p=pkg%2Fggml%2Fsources%2Fllama.cpp

llama : fix comment for "output.weight" tensor
---

diff --git a/llama.cpp b/llama.cpp
index 4a646eb9..33ee4fbb 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -1618,8 +1618,8 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
         // quantize only 2D tensors
         quantize &= (tensor.ne.size() == 2);
 
-        // GG: uncomment this to keep the output layer in FP16
-        //if (tensor.name.rfind("output")) {
+        // uncomment this to keep the output layer in FP16
+        //if (tensor.name == "output.weight") {
         //    quantize = false;
         //}