CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-07 09:57:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

context : do not cap the size of the context

Browse Source
This commit is contained in:
Georgi Gerganov
2025-11-01 17:44:52 +02:00
parent ff684363fa
commit c08d0d148a
2 changed files with 11 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
tools/server/server.cpp
View File

@@ -2497,12 +2497,20 @@ struct server_context {
void init() {
SRV_INF("initializing slots, n_slots = %d\n", params_base.n_parallel);
const int n_ctx_train = llama_model_n_ctx_train(model);
int n_ctx_slot = llama_n_ctx_seq(ctx);
if (n_ctx_slot > n_ctx_train) {
SRV_WRN("the slot context (%d) exceeds the training context of the model (%d) - capping\n", n_ctx_slot, n_ctx_train);
n_ctx_slot = n_ctx_train;
}
for (int i = 0; i < params_base.n_parallel; i++) {
server_slot slot;
slot.id = i;
slot.ctx = ctx;
slot.n_ctx = llama_n_ctx_seq(ctx);
slot.n_ctx = n_ctx_slot;
slot.mctx = mctx;
slot.prompt.tokens.has_mtmd = mctx != nullptr;