Lumpiasty/llama.cpp
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

ggml: add GATED_DELTA_NET op (#19504)

Browse Source 
* ggml: add GATED_DELTA_NET op

* remove the transpose

* add KDA

* add qwen35 dense

* llama : check for fused gated delta net backend support

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Aman Gupta
2026-03-07 15:41:10 +08:00
committed by GitHub
co-authored by Georgi Gerganov
parent 6fce5c6a7d
commit c5a778891b
15 changed files with 627 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/src/ggml-cuda/ggml-cuda.cu
+5
View File
@@ -53,6 +53,7 @@
#include "ggml-cuda/upscale.cuh"
#include "ggml-cuda/wkv.cuh"
#include "ggml-cuda/gla.cuh"
#include "ggml-cuda/gated_delta_net.cuh"
#include "ggml-cuda/set.cuh"
#include "ggml-cuda/set-rows.cuh"
#include "ggml-cuda/pad_reflect_1d.cuh"
@@ -2733,6 +2734,9 @@ static bool ggml_cuda_compute_forward(ggml_backend_cuda_context & ctx, struct gg
case GGML_OP_GATED_LINEAR_ATTN:
ggml_cuda_op_gated_linear_attn(ctx, dst);
break;
case GGML_OP_GATED_DELTA_NET:
ggml_cuda_op_gated_delta_net(ctx, dst);
break;
case GGML_OP_RWKV_WKV7:
ggml_cuda_op_rwkv_wkv7(ctx, dst);
break;
@@ -4972,6 +4976,7 @@ static bool ggml_backend_cuda_device_supports_op(ggml_backend_dev_t dev, const g
case GGML_OP_LEAKY_RELU:
case GGML_OP_RWKV_WKV6:
case GGML_OP_GATED_LINEAR_ATTN:
case GGML_OP_GATED_DELTA_NET:
case GGML_OP_RWKV_WKV7:
return true;
case GGML_OP_FLASH_ATTN_EXT: