CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-03 09:22:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add byte token type when tokenizer.model is not exists (#4641)

Browse Source 
* Add byte token type to hf format

* remove unused variable
This commit is contained in:
wonjun Jang
2023-12-27 17:37:25 +09:00
committed by GitHub
parent dc68f0054c
commit f56d6077d0
1 changed files with 6 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
convert.py
View File

@@ -357,6 +357,7 @@ class VocabLoader:
for tok in self.tokenizer.all_special_tokens for tok in self.tokenizer.all_special_tokens
} }
self.special_ids: set[int] = set(self.tokenizer.all_special_ids) self.special_ids: set[int] = set(self.tokenizer.all_special_ids)
self.reverse_vocab = {id: encoded_tok for encoded_tok, id in self.tokenizer.get_vocab().items()}
self.vocab_size_base: int = self.tokenizer.vocab_size self.vocab_size_base: int = self.tokenizer.vocab_size
self.vocab_size: int = self.vocab_size_base + len(self.added_tokens_dict) self.vocab_size: int = self.vocab_size_base + len(self.added_tokens_dict)
self.fname_tokenizer: Path = fname_tokenizer self.fname_tokenizer: Path = fname_tokenizer
@@ -370,15 +371,13 @@ class VocabLoader:
self.spm = None self.spm = None
def hf_tokens(self) -> Iterable[tuple[bytes, float, gguf.TokenType]]: def hf_tokens(self) -> Iterable[tuple[bytes, float, gguf.TokenType]]:
tokenizer = self.tokenizer
reverse_vocab = {id: encoded_tok for encoded_tok, id in tokenizer.get_vocab().items()}
added_tokens_ids = set(self.added_tokens_dict.values()) added_tokens_ids = set(self.added_tokens_dict.values())
for i in range(self.vocab_size_base): for i in range(self.vocab_size_base):
if i in added_tokens_ids: if i in added_tokens_ids:
continue continue
text = reverse_vocab[i].encode("utf-8") text = self.reverse_vocab[i].encode("utf-8")
yield text, self.get_token_score(i), self.get_token_type(i) yield text, self.get_token_score(i), self.get_token_type(i)
def get_token_type(self, token_id: int) -> gguf.TokenType: def get_token_type(self, token_id: int) -> gguf.TokenType:
@@ -394,10 +393,13 @@ class VocabLoader:
if self.spm.is_byte(token_id): if self.spm.is_byte(token_id):
toktype = gguf.TokenType.BYTE toktype = gguf.TokenType.BYTE
else: else:
token = self.reverse_vocab[token_id]
if token_id == self.unk_token_id: if token_id == self.unk_token_id:
toktype = gguf.TokenType.UNKNOWN toktype = gguf.TokenType.UNKNOWN
if token_id in self.special_ids: elif token_id in self.special_ids:
toktype = gguf.TokenType.CONTROL toktype = gguf.TokenType.CONTROL
elif len(token) == 6 and token.startswith("<0x") and token.endswith(">"):
toktype = gguf.TokenType.BYTE
return toktype return toktype