nikola/llama_cpp
1
0
Fork 0
Code Issues Pull Requests Actions 22 Packages Projects Releases Wiki Activity

llama_context: synchronize before reallocating output buffer (#17974)

Browse Source
This commit is contained in:
Jeff Bolz
2025-12-13 09:19:51 -06:00
committed by GitHub
parent 4d5ae24c0a
commit 5266379bca
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-context.cpp
+1
View File
@@ -1318,6 +1318,7 @@ uint32_t llama_context::output_reserve(int32_t n_outputs) {
// This doesn't happen often, but may be annoying in some cases (like the HellaSwag benchmark) // This doesn't happen often, but may be annoying in some cases (like the HellaSwag benchmark)
LLAMA_LOG_INFO("%s: reallocating output buffer from size %.02f MiB to %.02f MiB\n", __func__, prev_size / 1024.0 / 1024.0, new_size / 1024.0 / 1024.0); LLAMA_LOG_INFO("%s: reallocating output buffer from size %.02f MiB to %.02f MiB\n", __func__, prev_size / 1024.0 / 1024.0, new_size / 1024.0 / 1024.0);
#endif #endif
synchronize();
buf_output = nullptr; buf_output = nullptr;
logits = nullptr; logits = nullptr;
embd = nullptr; embd = nullptr;