happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

impl : use 6 digits for tensor dims (#20094) 

Many models have vocabulary sizes, and thus tensor shapes, with more
than 5 digits (ex: Gemma 3's vocab size is 262,208).

I already fixed this for `llama_format_tensor_shape` but missed it for
`llama_format_tensor_shape` until now. Oops.
This commit is contained in:
ddh0 2026-03-04 02:53:38 -06:00 committed by GitHub
parent cb8f4fa3f8
commit c99909dd0b
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 2 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
src/llama-impl.cpp
View File

@ -100,9 +100,9 @@ std::string format(const char * fmt, ...) {
std::string llama_format_tensor_shape(const std::vector<int64_t> & ne) {
char buf[256];
snprintf(buf, sizeof(buf), "%5" PRId64, ne.at(0));
snprintf(buf, sizeof(buf), "%6" PRId64, ne.at(0));
for (size_t i = 1; i < ne.size(); i++) {
snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), ", %5" PRId64, ne.at(i));
snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), ", %6" PRId64, ne.at(i));
}
return buf;
}