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

llama_tokenize should accept strings containing NUL now

Browse source
This commit is contained in:
goerch 2023-09-14 21:26:08 +02:00
parent a90bf494c9
commit e41209a95f
1 changed files with 2 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

4
common/common.cpp
View file

@ -804,10 +804,10 @@ std::vector<llama_token> llama_tokenize(
// upper limit for the number of tokens // upper limit for the number of tokens
int n_tokens = text.length() + add_bos; int n_tokens = text.length() + add_bos;
std::vector<llama_token> result(n_tokens); std::vector<llama_token> result(n_tokens);
n_tokens = llama_tokenize(ctx, text.c_str(), text.length(), result.data(), result.size(), add_bos); n_tokens = llama_tokenize(ctx, text.data(), text.length(), result.data(), result.size(), add_bos);
if (n_tokens < 0) { if (n_tokens < 0) {
result.resize(-n_tokens); result.resize(-n_tokens);
int check = llama_tokenize(ctx, text.c_str(), text.length(), result.data(), result.size(), add_bos); int check = llama_tokenize(ctx, text.data(), text.length(), result.data(), result.size(), add_bos);
GGML_ASSERT(check == -n_tokens); GGML_ASSERT(check == -n_tokens);
} else { } else {
result.resize(n_tokens); result.resize(n_tokens);