ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

llama : improve infill support and special token detection (#9798)

Browse source 
* llama : improve infill support

ggml-ci

* llama : add more FIM token strings

ggml-ci

* server : update prompt on slot restore (#9800)

* gguf : deprecate old FIM token KVs
This commit is contained in:
Georgi Gerganov 2024-10-12 08:21:51 +03:00 • committed by GitHub
parent 943d20b411
commit 11ac9800af
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
12 changed files with 601 additions and 427 deletions
Download patch file Download diff file Expand all files Collapse all files

9
gguf-py/gguf/gguf_writer.py
View file

@ -843,15 +843,6 @@ class GGUFWriter:
self.add_string(Keys.Tokenizer.CHAT_TEMPLATE, value)
def add_prefix_token_id(self, id: int) -> None:
self.add_uint32(Keys.Tokenizer.PREFIX_ID, id)
def add_suffix_token_id(self, id: int) -> None:
self.add_uint32(Keys.Tokenizer.SUFFIX_ID, id)
def add_middle_token_id(self, id: int) -> None:
self.add_uint32(Keys.Tokenizer.MIDDLE_ID, id)
def add_eot_token_id(self, id: int) -> None:
self.add_uint32(Keys.Tokenizer.EOT_ID, id)