furyhawk/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-07-21 10:15:35 +00:00
Code Issues Packages Projects Releases Wiki Activity

Reuse shared inverse RoPE operation for DSV4

Browse Source
This commit is contained in:
samuel
2026-07-17 13:30:53 +00:00
committed by Kawrakow
parent 0508580097
commit 17114aa3a7
3 changed files with 1 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files
ggml/include/ggml.h
-15
View File
@@ -2147,21 +2147,6 @@ extern "C" {
float beta_fast,
float beta_slow);
GGML_API struct ggml_tensor * ggml_rope_ext_back(
struct ggml_context * ctx,
struct ggml_tensor * a,
struct ggml_tensor * b,
struct ggml_tensor * c,
int n_dims,
int mode,
int n_ctx_orig,
float freq_base,
float freq_scale,
float ext_factor,
float attn_factor,
float beta_fast,
float beta_slow);
GGML_API struct ggml_tensor * ggml_rope_cache(
struct ggml_context * ctx,
struct ggml_tensor * b,
ggml/src/ggml.c
-20
View File
@@ -9507,26 +9507,6 @@ struct ggml_tensor * ggml_rope_back(
return result;
}
struct ggml_tensor * ggml_rope_ext_back(
struct ggml_context * ctx,
struct ggml_tensor * a,
struct ggml_tensor * b,
struct ggml_tensor * c,
int n_dims,
int mode,
int n_ctx_orig,
float freq_base,
float freq_scale,
float ext_factor,
float attn_factor,
float beta_fast,
float beta_slow) {
struct ggml_tensor * result = ggml_rope_ext(
ctx, a, b, c, n_dims, mode, n_ctx_orig, freq_base, freq_scale, ext_factor, attn_factor, beta_fast, beta_slow);
result->op = GGML_OP_ROPE_BACK;
return result;
}
// ggml_clamp
struct ggml_tensor * ggml_clamp(
src/graphs/build_deepseek4.cpp
+1 -1
View File
@@ -1335,7 +1335,7 @@ ggml_cgraph * llm_build_context::build_deepseek4() {
ggml_row_size(attn->type, n_embd_head),
ggml_row_size(attn->type, n_embd_head) * n_head,
ggml_row_size(attn->type, n_embd_head_nope));
attn_pe = ggml_rope_ext_back(ctx0, attn_pe, inp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig_l,
attn_pe = ggml_rope_back(ctx0, attn_pe, inp_pos, nullptr, n_embd_head_rope, rope_type, n_ctx_orig_l,
freq_base_l, freq_scale_l, ext_factor_l, attn_factor_l, beta_fast_l, beta_slow_l);
cb(attn_pe, "attn_derope", il);
attn = ggml_concat(ctx0, attn_nope, attn_pe, 0);