'viking' detokenizer clean spaces

This commit is contained in:
jaime-m-p 2024-07-04 22:30:48 +02:00
parent 8072089c4e
commit 8f5e1e0c76

View file

@ -5158,6 +5158,7 @@ static void llm_load_vocab(
} else if (
tokenizer_pre == "viking") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_VIKING;
vocab.tokenizer_clean_spaces = false;
} else if (
tokenizer_pre == "jais") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_JAIS;