Skip to content

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

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
ggerganov merged 2 commits into from
Apr 6, 2026
+9 −1
Merged

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

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3f1950e
convert : set "add bos" == True for Gemma 4
ggerganov Apr 6, 2026
4e19abc
cont : handle old GGUFs
ggerganov Apr 6, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion convert_hf_to_gguf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7472,7 +7472,7 @@ def set_vocab(self):
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()
Expand Down
8 changes: 8 additions & 0 deletions src/llama-vocab.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -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
Expand Down
Loading