From 91f6499393d2d999331fbfdba47a7f8b9f913f0d Mon Sep 17 00:00:00 2001
From: Kerfuffle <44031344+KerfuffleV2@users.noreply.github.com>
Date: Thu, 16 Nov 2023 19:14:37 -0700
Subject: Respect tokenizer.ggml.add_bos_token value when tokenizing (#4040)

* gguf-py: gguf-dump: Respect --no-tensor flag in JSON mode.

* Respect add_bos_token GGUF metadata value

* gguf-py: Try to fix SpecialVocab giving up too easily for the Nth time
---
 examples/infill/infill.cpp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'examples/infill/infill.cpp')

diff --git a/examples/infill/infill.cpp b/examples/infill/infill.cpp
index 62f5ce3c..11f7410e 100644
--- a/examples/infill/infill.cpp
+++ b/examples/infill/infill.cpp
@@ -230,7 +230,7 @@ int main(int argc, char ** argv) {
         LOG_TEE("\n");
         LOG_TEE("%s\n", get_system_info(params).c_str());
     }
-    const bool add_bos = llama_vocab_type(model) == LLAMA_VOCAB_TYPE_SPM;
+    const bool add_bos = llama_should_add_bos_token(model);
     LOG("add_bos: %d\n", add_bos);
 
     bool suff_rm_leading_spc = params.escape;
-- 
cgit v1.2.3