tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-08-15 04:33:06 -04:00
Code Issues Packages Projects Releases Wiki Activity

server : fix crash with multimodal models without BOS token (#4904)

Browse Source
This commit is contained in:
makomk
2024-01-13 14:16:11 +00:00
committed by GitHub
parent 15ebe59210
commit ee8243adaa
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
examples/server/server.cpp
View File

@@ -1835,7 +1835,7 @@ struct llama_server_context
slot.cache_tokens = prompt_tokens; slot.cache_tokens = prompt_tokens;
if (slot.n_past == slot.num_prompt_tokens) if (slot.n_past == slot.num_prompt_tokens && slot.n_past > 0)
{ {
// we have to evaluate at least 1 token to generate logits. // we have to evaluate at least 1 token to generate logits.
LOG_TEE("slot %d : we have to evaluate at least 1 token to generate logits\n", slot.id); LOG_TEE("slot %d : we have to evaluate at least 1 token to generate logits\n", slot.id);