CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-02 09:12:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

examples : replace llama_kv_cache_seq_* with llama_past_seq_*

Browse Source
This commit is contained in:
Francis Couture-Harpin
2024-06-10 14:44:42 -04:00
parent 372482dffe
commit 43d8d4bf9e
23 changed files with 125 additions and 112 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
examples/parallel/parallel.cpp
View File

@@ -200,7 +200,7 @@ int main(int argc, char ** argv) {
// assign the system KV cache to all parallel sequences
for (int32_t i = 1; i <= n_clients; ++i) {
llama_kv_cache_seq_cp(ctx, 0, i, -1, -1);
llama_past_seq_cp(ctx, 0, i, -1, -1);
}
LOG_TEE("\n");
@@ -232,9 +232,9 @@ int main(int argc, char ** argv) {
if (batch.n_tokens == 0) {
// all sequences have ended - clear the entire KV cache
for (int i = 1; i <= n_clients; ++i) {
llama_kv_cache_seq_rm(ctx, i, -1, -1);
llama_past_seq_rm(ctx, i, -1, -1);
// but keep the system prompt
llama_kv_cache_seq_cp(ctx, 0, i, -1, -1);
llama_past_seq_cp(ctx, 0, i, -1, -1);
}
LOG_TEE("%s: clearing the KV cache\n", __func__);
@@ -371,8 +371,8 @@ int main(int argc, char ** argv) {
}
// delete only the generated part of the sequence, i.e. keep the system prompt in the cache
llama_kv_cache_seq_rm(ctx, client.id + 1, -1, -1);
llama_kv_cache_seq_cp(ctx, 0, client.id + 1, -1, -1);
llama_past_seq_rm(ctx, client.id + 1, -1, -1);
llama_past_seq_cp(ctx, 0, client.id + 1, -1, -1);
const auto t_main_end = ggml_time_us();