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

Add phi-2 tokenizer #7300

Closed
wants to merge 2 commits into from
Closed

Add phi-2 tokenizer #7300

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
0c6ae12
add phi 2 hash
BramVanroy May 15, 2024
4ac22a8
Update convert-hf-to-gguf-update.py
BramVanroy May 15, 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
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 @@ -77,6 +77,7 @@ class TOKENIZER_TYPE(IntEnum):
{"name": "jina-v2-en", "tokt": TOKENIZER_TYPE.WPM, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-en", }, # WPM!
{"name": "jina-v2-es", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-es", },
{"name": "jina-v2-de", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/jinaai/jina-embeddings-v2-base-de", },
{ "name": "phi-2", "tokt": TOKENIZER_TYPE.BPE, "repo": "https://huggingface.co/microsoft/phi-2", },
]

# make directory "models/tokenizers" if it doesn't exist
Expand Down
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 @@ -469,6 +469,9 @@ def get_vocab_base_pre(self, tokenizer) -> str:
if chkhsh == "27949a2493fc4a9f53f5b9b029c82689cfbe5d3a1929bb25e043089e28466de6":
# ref: https://huggingface.co/jinaai/jina-embeddings-v2-base-de
res = "jina-v2-de"
if chkhsh == "fcace8b9cac38ce847670c970cd5892031a753a1ef381abd1d9af00f713da085":
# ref: https://huggingface.co/microsoft/phi-2
res = "phi-2"
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This new pre-tokenizer has to be handled in llama.cpp:

llama.cpp/llama.cpp

Lines 4414 to 4475 in e18bc6a

// for now, only BPE models have pre-tokenizers
if (vocab.type == LLAMA_VOCAB_TYPE_BPE) {
if (tokenizer_pre.empty()) {
LLAMA_LOG_WARN("%s: missing pre-tokenizer type, using: 'default'\n", __func__);
LLAMA_LOG_WARN("%s: \n", __func__);
LLAMA_LOG_WARN("%s: ************************************ \n", __func__);
LLAMA_LOG_WARN("%s: GENERATION QUALITY WILL BE DEGRADED! \n", __func__);
LLAMA_LOG_WARN("%s: CONSIDER REGENERATING THE MODEL \n", __func__);
LLAMA_LOG_WARN("%s: ************************************ \n", __func__);
LLAMA_LOG_WARN("%s: \n", __func__);
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
} else if (
tokenizer_pre == "default") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
} else if (
tokenizer_pre == "llama3" ||
tokenizer_pre == "llama-v3" ||
tokenizer_pre == "llama-bpe") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_LLAMA3;
} else if (
tokenizer_pre == "deepseek-llm") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_LLM;
} else if (
tokenizer_pre == "deepseek-coder") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEEPSEEK_CODER;
} else if (
tokenizer_pre == "falcon") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_FALCON;
} else if (
tokenizer_pre == "mpt") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_MPT;
} else if (
tokenizer_pre == "starcoder") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_STARCODER;
} else if (
tokenizer_pre == "gpt-2" ||
tokenizer_pre == "jina-es" ||
tokenizer_pre == "jina-de" ||
tokenizer_pre == "jina-v2-es" ||
tokenizer_pre == "jina-v2-de") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_GPT2;
} else if (
tokenizer_pre == "refact") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_REFACT;
} else if (
tokenizer_pre == "command-r") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_COMMAND_R;
} else if (
tokenizer_pre == "qwen2") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_QWEN2;
} else if (
tokenizer_pre == "olmo") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_OLMO;
} else if (
tokenizer_pre == "dbrx") {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DBRX;
} else {
throw std::runtime_error(format("unknown pre-tokenizer type: '%s'", tokenizer_pre.c_str()));
}
} else {
vocab.type_pre = LLAMA_VOCAB_PRE_TYPE_DEFAULT;
}

Copy link
Contributor

@teleprint-me teleprint-me May 17, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@ggerganov Thats not necessary. I solved this already in #7219 and ##7117.


if res is None:
logger.warning("\n")
Expand Down