CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-04 09:32:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama : add infill sampler (#9896)

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2024-10-15 16:35:33 +03:00
committed by GitHub
parent 223c25a72f
commit 755a9b2bf0
9 changed files with 300 additions and 29 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
src/llama-vocab.h
View File

@@ -48,7 +48,7 @@ struct llama_vocab {
id special_cls_id = LLAMA_TOKEN_NULL;
id special_mask_id = LLAMA_TOKEN_NULL;
id linefeed_id = 13;
id linefeed_id = 13;
// fim tokens
id special_fim_pre_id = LLAMA_TOKEN_NULL;
@@ -149,6 +149,12 @@ int32_t llama_token_to_piece_impl(
int32_t lstrip,
bool special);
// check if token0 is contained as a prefix in token1
bool llama_token_is_prefix_impl(
const struct llama_vocab & vocab,
llama_token token0,
llama_token token1);
int32_t llama_detokenize_impl(
const struct llama_vocab & vocab,
const llama_token * tokens,