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

chore: Apply deduped token type references

Browse source
This commit is contained in:
teleprint-me 2024-05-21 00:45:06 -04:00
parent a35b76755f
commit 6296206392
No known key found for this signature in database
GPG key ID: B0D11345E65C4D48
1 changed files with 27 additions and 37 deletions
Download patch file Download diff file Expand all files Collapse all files

64
convert-hf-to-gguf.py
View file

@ -40,16 +40,6 @@ logger = logging.getLogger("hf-to-gguf")
###### MODEL DEFINITIONS ###### ###### MODEL DEFINITIONS ######
class SentencePieceTokenTypes(IntEnum):
NORMAL = 1
UNKNOWN = 2
CONTROL = 3
USER_DEFINED = 4
UNUSED = 5
BYTE = 6
AnyModel = TypeVar("AnyModel", bound="type[Model]") AnyModel = TypeVar("AnyModel", bound="type[Model]")
@ -538,22 +528,22 @@ class Model:
tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)] tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)]
scores: list[float] = [-10000.0] * vocab_size scores: list[float] = [-10000.0] * vocab_size
toktypes: list[int] = [SentencePieceTokenTypes.UNKNOWN] * vocab_size toktypes: list[int] = [gguf.TokenType.UNKNOWN] * vocab_size
for token_id in range(tokenizer.vocab_size()): for token_id in range(tokenizer.vocab_size()):
piece = tokenizer.IdToPiece(token_id) piece = tokenizer.IdToPiece(token_id)
text = piece.encode("utf-8") text = piece.encode("utf-8")
score = tokenizer.GetScore(token_id) score = tokenizer.GetScore(token_id)
toktype = SentencePieceTokenTypes.NORMAL toktype = gguf.TokenType.NORMAL
if tokenizer.IsUnknown(token_id): if tokenizer.IsUnknown(token_id):
toktype = SentencePieceTokenTypes.UNKNOWN toktype = gguf.TokenType.UNKNOWN
elif tokenizer.IsControl(token_id): elif tokenizer.IsControl(token_id):
toktype = SentencePieceTokenTypes.CONTROL toktype = gguf.TokenType.CONTROL
elif tokenizer.IsUnused(token_id): elif tokenizer.IsUnused(token_id):
toktype = SentencePieceTokenTypes.UNUSED toktype = gguf.TokenType.UNUSED
elif tokenizer.IsByte(token_id): elif tokenizer.IsByte(token_id):
toktype = SentencePieceTokenTypes.BYTE toktype = gguf.TokenType.BYTE
tokens[token_id] = text tokens[token_id] = text
scores[token_id] = score scores[token_id] = score
@ -571,7 +561,7 @@ class Model:
tokens[token_id] = key.encode("utf-8") tokens[token_id] = key.encode("utf-8")
scores[token_id] = -1000.0 scores[token_id] = -1000.0
toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED toktypes[token_id] = gguf.TokenType.USER_DEFINED
if vocab_size > len(tokens): if vocab_size > len(tokens):
pad_count = vocab_size - len(tokens) pad_count = vocab_size - len(tokens)
@ -579,7 +569,7 @@ class Model:
for i in range(1, pad_count + 1): for i in range(1, pad_count + 1):
tokens.append(bytes(f"[PAD{i}]", encoding="utf-8")) tokens.append(bytes(f"[PAD{i}]", encoding="utf-8"))
scores.append(-1000.0) scores.append(-1000.0)
toktypes.append(SentencePieceTokenTypes.UNUSED) toktypes.append(gguf.TokenType.UNUSED)
self.gguf_writer.add_tokenizer_model("llama") self.gguf_writer.add_tokenizer_model("llama")
self.gguf_writer.add_tokenizer_pre("default") self.gguf_writer.add_tokenizer_pre("default")
@ -1663,7 +1653,7 @@ class Phi3MiniModel(Model):
tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)] tokens: list[bytes] = [f"[PAD{i}]".encode("utf-8") for i in range(vocab_size)]
scores: list[float] = [-10000.0] * vocab_size scores: list[float] = [-10000.0] * vocab_size
toktypes: list[int] = [SentencePieceTokenTypes.UNKNOWN] * vocab_size toktypes: list[int] = [gguf.TokenType.UNKNOWN] * vocab_size
for token_id in range(tokenizer.vocab_size()): for token_id in range(tokenizer.vocab_size()):
@ -1671,15 +1661,15 @@ class Phi3MiniModel(Model):
text = piece.encode("utf-8") text = piece.encode("utf-8")
score = tokenizer.GetScore(token_id) score = tokenizer.GetScore(token_id)
toktype = SentencePieceTokenTypes.NORMAL toktype = gguf.TokenType.NORMAL
if tokenizer.IsUnknown(token_id): if tokenizer.IsUnknown(token_id):
toktype = SentencePieceTokenTypes.UNKNOWN toktype = gguf.TokenType.UNKNOWN
elif tokenizer.IsControl(token_id): elif tokenizer.IsControl(token_id):
toktype = SentencePieceTokenTypes.CONTROL toktype = gguf.TokenType.CONTROL
elif tokenizer.IsUnused(token_id): elif tokenizer.IsUnused(token_id):
toktype = SentencePieceTokenTypes.UNUSED toktype = gguf.TokenType.UNUSED
elif tokenizer.IsByte(token_id): elif tokenizer.IsByte(token_id):
toktype = SentencePieceTokenTypes.BYTE toktype = gguf.TokenType.BYTE
tokens[token_id] = text tokens[token_id] = text
scores[token_id] = score scores[token_id] = score
@ -1698,7 +1688,7 @@ class Phi3MiniModel(Model):
tokens[token_id] = key.encode("utf-8") tokens[token_id] = key.encode("utf-8")
scores[token_id] = -1000.0 scores[token_id] = -1000.0
toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED toktypes[token_id] = gguf.TokenType.USER_DEFINED
tokenizer_config_file = self.dir_model / 'tokenizer_config.json' tokenizer_config_file = self.dir_model / 'tokenizer_config.json'
if tokenizer_config_file.is_file(): if tokenizer_config_file.is_file():
@ -1708,13 +1698,13 @@ class Phi3MiniModel(Model):
for token_id, foken_data in added_tokens_decoder.items(): for token_id, foken_data in added_tokens_decoder.items():
token_id = int(token_id) token_id = int(token_id)
token = foken_data["content"].encode("utf-8") token = foken_data["content"].encode("utf-8")
if toktypes[token_id] != SentencePieceTokenTypes.UNKNOWN: if toktypes[token_id] != gguf.TokenType.UNKNOWN:
assert(tokens[token_id] == token) assert(tokens[token_id] == token)
tokens[token_id] = token tokens[token_id] = token
scores[token_id] = -1000.0 scores[token_id] = -1000.0
toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED toktypes[token_id] = gguf.TokenType.USER_DEFINED
if foken_data.get("special"): if foken_data.get("special"):
toktypes[token_id] = SentencePieceTokenTypes.CONTROL toktypes[token_id] = gguf.TokenType.CONTROL
tokenizer_file = self.dir_model / 'tokenizer.json' tokenizer_file = self.dir_model / 'tokenizer.json'
if tokenizer_file.is_file(): if tokenizer_file.is_file():
@ -1724,13 +1714,13 @@ class Phi3MiniModel(Model):
for foken_data in added_tokens: for foken_data in added_tokens:
token_id = int(foken_data["id"]) token_id = int(foken_data["id"])
token = foken_data["content"].encode("utf-8") token = foken_data["content"].encode("utf-8")
if toktypes[token_id] != SentencePieceTokenTypes.UNKNOWN: if toktypes[token_id] != gguf.TokenType.UNKNOWN:
assert(tokens[token_id] == token) assert(tokens[token_id] == token)
tokens[token_id] = token tokens[token_id] = token
scores[token_id] = -1000.0 scores[token_id] = -1000.0
toktypes[token_id] = SentencePieceTokenTypes.USER_DEFINED toktypes[token_id] = gguf.TokenType.USER_DEFINED
if foken_data.get("special"): if foken_data.get("special"):
toktypes[token_id] = SentencePieceTokenTypes.CONTROL toktypes[token_id] = gguf.TokenType.CONTROL
self.gguf_writer.add_tokenizer_model("llama") self.gguf_writer.add_tokenizer_model("llama")
self.gguf_writer.add_tokenizer_pre("default") self.gguf_writer.add_tokenizer_pre("default")
@ -1889,15 +1879,15 @@ class InternLM2Model(Model):
logger.warning(f"InternLM2 convert token '{text}' to '🐉'!") logger.warning(f"InternLM2 convert token '{text}' to '🐉'!")
text = "🐉".encode("utf-8") text = "🐉".encode("utf-8")
toktype = SentencePieceTokenTypes.NORMAL toktype = gguf.TokenType.NORMAL
if tokenizer.IsUnknown(token_id): if tokenizer.IsUnknown(token_id):
toktype = SentencePieceTokenTypes.UNKNOWN toktype = gguf.TokenType.UNKNOWN
elif tokenizer.IsControl(token_id): elif tokenizer.IsControl(token_id):
toktype = SentencePieceTokenTypes.CONTROL toktype = gguf.TokenType.CONTROL
elif tokenizer.IsUnused(token_id): elif tokenizer.IsUnused(token_id):
toktype = SentencePieceTokenTypes.UNUSED toktype = gguf.TokenType.UNUSED
elif tokenizer.IsByte(token_id): elif tokenizer.IsByte(token_id):
toktype = SentencePieceTokenTypes.BYTE toktype = gguf.TokenType.BYTE
tokens.append(text) tokens.append(text)
scores.append(score) scores.append(score)
@ -1911,7 +1901,7 @@ class InternLM2Model(Model):
for key in added_tokens_json: for key in added_tokens_json:
tokens.append(key.encode("utf-8")) tokens.append(key.encode("utf-8"))
scores.append(-1000.0) scores.append(-1000.0)
toktypes.append(SentencePieceTokenTypes.USER_DEFINED) toktypes.append(gguf.TokenType.USER_DEFINED)
self.gguf_writer.add_tokenizer_model("llama") self.gguf_writer.add_tokenizer_model("llama")
self.gguf_writer.add_tokenizer_pre("default") self.gguf_writer.add_tokenizer_pre("default")