Skip to content
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

Jump to bottom

llama : Added support for Tekken pre-tokenizer (#8577) #8579

Merged
merged 4 commits into from
Jul 20, 2024
Merged

llama : Added support for Tekken pre-tokenizer (#8577) #8579

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
7fc8505
llama : Added support for Tekken pre-tokenizer (#8577)
m18coppola Jul 18, 2024
447c080
llama : fix order of pre-tokenizers
ggerganov Jul 19, 2024
8506e13
Merge branch 'ggerganov:master' into master
m18coppola Jul 19, 2024
dd5a0bf
* Tekken pre-tokenizer no longer uses clean_up_tokenization_spaces
m18coppola Jul 19, 2024
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
3 changes: 3 additions & 0 deletions convert_hf_to_gguf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -593,6 +593,9 @@ def get_vocab_base_pre(self, tokenizer) -> str:
if chkhsh == "b53802fb28e26d645c3a310b34bfe07da813026ec7c7716883404d5e0f8b1901":
# ref: https://huggingface.co/core42/jais-13b
res = "jais"
if chkhsh == "aa78fe8b04bc622b077520b1fb3d3a5c6f7a53dd375e2361e62599be3cf58de1":
m18coppola marked this conversation as resolved.
Show resolved Hide resolved
# ref: https://huggingface.co/mistralai/Mistral-Nemo-Base-2407
res = "tekken"

if res is None:
logger.warning("\n")
Expand Down
1 change: 1 addition & 0 deletions convert_hf_to_gguf_update.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -91,6 +91,7 @@ class TOKENIZER_TYPE(IntEnum):
{"name": "gemma-2", "tokt": TOKENIZER_TYPE.SPM, "repo": "https://huggingface.co/google/gemma-2-9b", },
{"name": "jais", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/core42/jais-13b", },
{"name": "t5", "tokt": TOKENIZER_TYPE.UGM, "repo": "https://huggingface.co/google-t5/t5-small", },
{"name": "tekken", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/mistralai/Mistral-Nemo-Base-2407", },
]


Expand Down
1 change: 1 addition & 0 deletions include/llama.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -92,6 +92,7 @@ extern "C" {
LLAMA_VOCAB_PRE_TYPE_CHATGLM4 = 17,
LLAMA_VOCAB_PRE_TYPE_VIKING = 18,
LLAMA_VOCAB_PRE_TYPE_JAIS = 19,
LLAMA_VOCAB_PRE_TYPE_TEKKEN = 20,
};

// note: these values should be synchronized with ggml_rope
Expand Down
12 changes: 12 additions & 0 deletions src/llama.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5517,6 +5517,11 @@ static void llm_load_vocab(
tokenizer_pre == "viking") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_VIKING;
vocab.tokenizer_clean_spaces = false;
} else if (
tokenizer_pre == "tekken") {
compilade marked this conversation as resolved.
Show resolved Hide resolved
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_TEKKEN;
vocab.tokenizer_ignore_merges = true;
vocab.tokenizer_add_bos = true;
} else if (
tokenizer_pre == "jais") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_JAIS;
Expand Down Expand Up @@ -15581,6 +15586,13 @@ struct llm_tokenizer_bpe {
"\\p{N}",
};
break;
case LLAMA_VOCAB_PRE_TYPE_TEKKEN:
// original regex from tokenizer.json
// "[^\\r\\n\\p{L}\\p{N}]?[\\p{Lu}\\p{Lt}\\p{Lm}\\p{Lo}\\p{M}]*[\\p{Ll}\\p{Lm}\\p{Lo}\\p{M}]+|[^\\r\\n\\p{L}\\p{N}]?[\\p{Lu}\\p{Lt}\\p{Lm}\\p{Lo}\\p{M}]+[\\p{Ll}\\p{Lm}\\p{Lo}\\p{M}]*|\\p{N}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n/]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+"
regex_exprs = {
"[^\\r\\n\\p{L}\\p{N}]?((?=[\\p{L}])([^a-z]))*((?=[\\p{L}])([^A-Z]))+|[^\\r\\n\\p{L}\\p{N}]?((?=[\\p{L}])([^a-z]))+((?=[\\p{L}])([^A-Z]))*|\\p{N}| ?[^\\s\\p{L}\\p{N}]+[\\r\\n/]*|\\s*[\\r\\n]+|\\s+(?!\\S)|\\s+",
};
break;
default:
// default regex for BPE tokenization pre-processing
regex_exprs = {
Expand Down
Loading