tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-28 04:15:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

main : fix session loading bug (#3400)

Browse Source
This commit is contained in:
Georgi Gerganov
2023-10-11 23:55:08 +03:00
parent a8bdd65525
commit b8fe4b5cc9
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
examples/main/main.cpp
View File

@ -297,6 +297,9 @@ int main(int argc, char ** argv) {
LOG_TEE("%s: session file matches %zu / %zu tokens of prompt\n",
__func__, n_matching_session_tokens, embd_inp.size());
}
// remove any "future" tokens that we might have inherited from the previous session
llama_kv_cache_tokens_rm(ctx, n_matching_session_tokens, -1);
}
LOGLN(
@ -545,9 +548,6 @@ int main(int argc, char ** argv) {
if (i > 0) {
embd.erase(embd.begin(), embd.begin() + i);
}
// remove any "future" tokens that we might have inherited from the session from the KV cache
llama_kv_cache_tokens_rm(ctx, n_past, -1);
}
// evaluate tokens in batches