tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-20 17:49:18 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'master' into gg/llama-kv-cache

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2025-01-27 14:00:56 +02:00
parent a0c500b4dc df984e0147
commit e665b57fa2
6 changed files with 106 additions and 96 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
src/llama-context.h
View File

@ -28,6 +28,9 @@ struct llama_batch_manager_i {
virtual void restore() = 0;
virtual void update() = 0;
virtual void finalize() = 0;
// TODO: might be temporary
int64_t n_outputs_all = 0;
};
// TODO: make implementation details private
@ -98,7 +101,7 @@ struct llama_context {
void * abort_callback_data = nullptr;
// TODO: do not pass logits_all explicitly
std::unique_ptr<llama_batch_manager_i> prepare_batch(const llama_batch & batch, bool logits_all);
std::unique_ptr<llama_batch_manager_i> prepare_batch(const llama_batch & batch);
// returns the result of ggml_backend_sched_graph_compute_async execution
enum ggml_status compute_graph(