vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

add comment regarding special token regex in chameleon pre-tokenizer

Browse source
This commit is contained in:
nopperl 2024-07-22 13:44:24 +02:00
parent 6e0ded3637
commit 05f138551f
1 changed files with 4 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

4
src/llama.cpp
View file

@ -15843,6 +15843,10 @@ struct llm_tokenizer_bpe {
}; };
break; break;
case LLAMA_VOCAB_PRE_TYPE_CHAMELEON: case LLAMA_VOCAB_PRE_TYPE_CHAMELEON:
// Note: in theory, the special token (sentinel and image token) regex_exprs below
// are unnecessary, as they are split in `tokenizer_st_partition` anyway.
// However, since the upstream pre-tokenizer uses them, they are also
// included here (see https://huggingface.co/facebook/chameleon-7b).
regex_exprs = { regex_exprs = {
"<sentinel:[0-9]+>", // Sentinel tokens "<sentinel:[0-9]+>", // Sentinel tokens
"(IMGIMG)((A|B|C|D|E|F|G|H|I){1,4})Z", // Image tokens "(IMGIMG)((A|B|C|D|E|F|G|H|I){1,4})Z", // Image tokens