tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-08-28 19:18:57 -04:00
Code Issues Packages Projects Releases Wiki Activity

ggml : remove ggml_flash_attn and ggml_flash_ff (#7463)

Browse Source 
ggml-ci
This commit is contained in:
Georgi Gerganov
2024-05-23 10:00:44 +03:00
committed by GitHub
parent e84b71c2c6
commit d48c88cbd5
5 changed files with 47 additions and 769 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
ggml.h
View File

@@ -481,9 +481,7 @@ extern "C" {
GGML_OP_ARGSORT,
GGML_OP_LEAKY_RELU,
GGML_OP_FLASH_ATTN,
GGML_OP_FLASH_ATTN_EXT,
GGML_OP_FLASH_FF,
GGML_OP_FLASH_ATTN_BACK,
GGML_OP_SSM_CONV,
GGML_OP_SSM_SCAN,
@@ -1761,13 +1759,6 @@ extern "C" {
struct ggml_tensor * a,
int k);
GGML_API struct ggml_tensor * ggml_flash_attn(
struct ggml_context * ctx,
struct ggml_tensor * q,
struct ggml_tensor * k,
struct ggml_tensor * v,
bool masked);
#define GGML_KQ_MASK_PAD 32
// q: [n_embd, n_batch, n_head, 1]
@@ -1788,6 +1779,7 @@ extern "C" {
struct ggml_tensor * a,
enum ggml_prec prec);
// TODO: needs to be adapted to ggml_flash_attn_ext
GGML_API struct ggml_tensor * ggml_flash_attn_back(
struct ggml_context * ctx,
struct ggml_tensor * q,
@@ -1796,14 +1788,6 @@ extern "C" {
struct ggml_tensor * d,
bool masked);
GGML_API struct ggml_tensor * ggml_flash_ff(
struct ggml_context * ctx,
struct ggml_tensor * a,
struct ggml_tensor * b0,
struct ggml_tensor * b1,
struct ggml_tensor * c0,
struct ggml_tensor * c1);
GGML_API struct ggml_tensor * ggml_ssm_conv(
struct ggml_context * ctx,
struct ggml_tensor * s,