vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llama : flash_attn cparam + fix defrag

Browse source
This commit is contained in:
Georgi Gerganov 2024-04-17 12:00:35 +03:00
parent 2c41180e88
commit 599ce84a71
No known key found for this signature in database
GPG key ID: 449E073F9DC10735
4 changed files with 198 additions and 163 deletions
Download patch file Download diff file Expand all files Collapse all files

1
llama.h
View file

@ -270,6 +270,7 @@ extern "C" {
bool logits_all; // the llama_decode() call computes all logits, not just the last one (DEPRECATED - set llama_batch.logits instead)
bool embeddings; // if true, extract embeddings (together with logits)
bool offload_kqv; // whether to offload the KQV ops (including the KV cache) to GPU
bool flash_attn; // whether to use flash attention
// Abort callback
// if it returns true, execution of llama_decode() will be aborted