tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-08-08 01:26:51 -04:00
Code Issues Packages Projects Releases Wiki Activity

graph : make FA compatible with MLA + add initial Metal kernels (#12953)

Browse Source 
* graph : make mla compatible with FA

* metal : add exp FA kernels for DeepSeek models

ggml-ci

* llama : minor naming updates

ggml-ci

* ggml : disable FA for DS head sizes

* tests : add FA tests for MLA shapes

ggml-ci
This commit is contained in:
Georgi Gerganov
2025-04-17 18:16:36 +03:00
committed by GitHub
parent 207c22ec2d
commit 2f74c354c0
8 changed files with 117 additions and 26 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
ggml/src/ggml-cuda/ggml-cuda.cu
View File

@@ -3237,6 +3237,10 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g
if (op->src[0]->ne[0] == 192) {
return false;
}
if (op->src[0]->ne[0] == 576) {
// DeepSeek MLA
return false;
}
if (op->src[0]->ne[3] != 1) {
return false;
}