Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-27 06:39:25 +01:00
Code Issues Packages Projects Releases Wiki Activity

infill : assert prefix/suffix tokens + remove old space logic (#8351)

Browse Source
This commit is contained in:
Georgi Gerganov 2024-07-08 09:34:35 +03:00 committed by GitHub
parent ffd00797d8
commit 6f0dbf6ab0
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 9 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
examples/infill/infill.cpp
View File

@ -204,21 +204,17 @@ int main(int argc, char ** argv) {
GGML_ASSERT(llama_add_eos_token(model) != 1); GGML_ASSERT(llama_add_eos_token(model) != 1);
LOG("add_bos: %d\n", add_bos); LOG("add_bos: %d\n", add_bos);
bool suff_rm_leading_spc = params.escape;
if (suff_rm_leading_spc && params.input_suffix.find_first_of(' ') == 0 && params.input_suffix.size() > 1) {
params.input_suffix.erase(0, 1);
suff_rm_leading_spc = false;
}
std::vector<llama_token> embd_inp; std::vector<llama_token> embd_inp;
std::vector<llama_token> embd_end; std::vector<llama_token> embd_end;
std::vector<llama_token> inp_pfx = ::llama_tokenize(ctx, params.input_prefix, false); std::vector<llama_token> inp_pfx = ::llama_tokenize(ctx, params.input_prefix, false);
std::vector<llama_token> inp_sfx = ::llama_tokenize(ctx, params.input_suffix, false); std::vector<llama_token> inp_sfx = ::llama_tokenize(ctx, params.input_suffix, false);
const int space_token = 29871;
if (suff_rm_leading_spc && inp_sfx[0] == space_token) { GGML_ASSERT(llama_token_prefix(model) >= 0);
inp_sfx.erase(inp_sfx.begin()); GGML_ASSERT(llama_token_suffix(model) >= 0);
}
inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(model)); inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(model));
inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(model)); inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(model));
embd_inp = params.spm_infill ? inp_sfx : inp_pfx; embd_inp = params.spm_infill ? inp_sfx : inp_pfx;
embd_end = params.spm_infill ? inp_pfx : inp_sfx; embd_end = params.spm_infill ? inp_pfx : inp_sfx;
if (add_bos) { if (add_bos) {
@ -516,19 +512,14 @@ int main(int argc, char ** argv) {
string_process_escapes(params.input_prefix); string_process_escapes(params.input_prefix);
string_process_escapes(params.input_suffix); string_process_escapes(params.input_suffix);
} }
suff_rm_leading_spc = params.escape;
if (suff_rm_leading_spc && params.input_suffix.find_first_of(' ') == 0 && params.input_suffix.size() > 1) {
params.input_suffix.erase(0, 1);
suff_rm_leading_spc = false;
}
// tokenize new prefix and suffix // tokenize new prefix and suffix
std::vector<llama_token> inp_pfx = ::llama_tokenize(ctx, params.input_prefix, false); std::vector<llama_token> inp_pfx = ::llama_tokenize(ctx, params.input_prefix, false);
std::vector<llama_token> inp_sfx = ::llama_tokenize(ctx, params.input_suffix, false); std::vector<llama_token> inp_sfx = ::llama_tokenize(ctx, params.input_suffix, false);
if (suff_rm_leading_spc && inp_sfx[0] == space_token) {
inp_sfx.erase(inp_sfx.begin());
}
inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(model)); inp_pfx.insert(inp_pfx.begin(), llama_token_prefix(model));
inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(model)); inp_sfx.insert(inp_sfx.begin(), llama_token_suffix(model));
embd_inp = params.spm_infill ? inp_sfx : inp_pfx; embd_inp = params.spm_infill ? inp_sfx : inp_pfx;
embd_end = params.spm_infill ? inp_pfx : inp_sfx; embd_end = params.spm_infill ? inp_pfx : inp_sfx;
if (add_bos) { if (add_bos) {