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

Clarify logic in conversion

Browse source
This commit is contained in:
goerch 2023-10-22 20:35:50 +02:00
parent 1dc13168ff
commit 74204ccbae
1 changed files with 8 additions and 7 deletions
Download patch file Download diff file Expand all files Collapse all files

15
convert-mpt-hf-to-gguf.py
View file

@ -132,15 +132,16 @@ added_vocab = tokenizer.get_added_vocab()
reverse_vocab = {id: encoded_tok for encoded_tok, id in tokenizer.vocab.items()} reverse_vocab = {id: encoded_tok for encoded_tok, id in tokenizer.vocab.items()}
for i in range(vocab_size): for i in range(vocab_size):
if i in reverse_vocab: if i not in reverse_vocab:
tokens.append(reverse_vocab[i])
if reverse_vocab[i] not in added_vocab:
toktypes.append(gguf.TokenType.NORMAL)
else:
toktypes.append(gguf.TokenType.USER_DEFINED)
else:
tokens.append(f"[PAD{i}]") tokens.append(f"[PAD{i}]")
toktypes.append(gguf.TokenType.USER_DEFINED) toktypes.append(gguf.TokenType.USER_DEFINED)
elif reverse_vocab[i] in added_vocab:
# NOTE: wouldn't we like to distinguish CONTROL tokens here?
tokens.append(reverse_vocab[i])
toktypes.append(gguf.TokenType.USER_DEFINED)
else:
tokens.append(reverse_vocab[i])
toktypes.append(gguf.TokenType.NORMAL)
gguf_writer.add_token_list(tokens) gguf_writer.add_token_list(tokens)
gguf_writer.add_token_types(toktypes) gguf_writer.add_token_types(toktypes)