CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-10-30 08:42:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

server : clear the KV cache beyond n_past before llama_decode

Browse Source
This commit is contained in:
Georgi Gerganov
2023-09-28 18:12:39 +03:00
parent 2b8830af71
commit ce2d995af2
1 changed files with 4 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
examples/server/server.cpp
View File

@@ -434,6 +434,10 @@ struct llama_server_context
{ {
n_eval = params.n_batch; n_eval = params.n_batch;
} }
// since #3228 we now have to manually manage the KV cache
llama_kv_cache_tokens_rm(ctx, n_past, -1);
if (llama_decode(ctx, llama_batch_get_one(&embd[n_past], n_eval, n_past, 0), params.n_threads)) if (llama_decode(ctx, llama_batch_get_one(&embd[n_past], n_eval, n_past, 0), params.n_threads))
{ {
LOG_ERROR("failed to eval", { LOG_ERROR("failed to eval", {