happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

convert : set "add bos" == True for Gemma 4 (#21500) 

* convert : set "add bos" == True for Gemma 4

* cont : handle old GGUFs
This commit is contained in:
Georgi Gerganov 2026-04-06 13:52:07 +03:00 committed by GitHub
parent f51fd36d79
commit 400ac8e194
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 9 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
convert_hf_to_gguf.py
View File

@ -7472,7 +7472,7 @@ class Gemma4Model(Gemma3Model):
special_vocab = gguf.SpecialVocab(self.dir_model, load_merges=True)
special_vocab.add_to_gguf(self.gguf_writer)
self.gguf_writer.add_add_space_prefix(False)
self.gguf_writer.add_add_bos_token(False) # already added via the chat template
self.gguf_writer.add_add_bos_token(True)
def set_gguf_parameters(self):
super().set_gguf_parameters()

8
src/llama-vocab.cpp
View File

@ -2325,6 +2325,14 @@ void llama_vocab::impl::load(llama_model_loader & ml, const LLM_KV & kv) {
if (ml.get_key(LLM_KV_TOKENIZER_ADD_SEP, temp, false)) {
add_sep = temp;
}
// workaround for Gemma 4
// ref: https://github.com/ggml-org/llama.cpp/pull/21500
if (pre_type == LLAMA_VOCAB_PRE_TYPE_GEMMA4 && !add_bos) {
add_bos = true;
LLAMA_LOG_WARN("%s: override '%s' to 'true' for Gemma4\n", __func__, kv(LLM_KV_TOKENIZER_ADD_BOS).c_str());
}
}
// auto-detect special tokens by text