tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-28 04:15:21 +00:00
Code Issues Packages Projects Releases Wiki Activity

llama : DeepSeek V2/V3 MLA implementation (#12801)

Browse Source 
* Merged using squash to remove all noise commit messages

* Force flash attention off for `LLM_ARCH_DEEPSEEK2` - embedding too large

* Removed 3 conts (2x RoPE and 1x RMS-norm)

* Changed to use `<cmath>` instead of `<math.h>`

* Reverted removal of the 3 conts

* Used `reshape` in `llm_graph_context::build_attn_mha()`

* Use `k_pe = ggml_reshape`

* Removed the 3 conts again

* Removed the 3D views of `wk_b` and `wv_b`, and just save and 3D in GGUF

* Removed MQA optimisation from `build_attn_mha()` as no gains now

* Simplified `is_mla` branch in `llm_build_deepseek2()`

* Removed `build_attn_mla` and added `nullptr` to all `build_atnn` calls

* Fixed call to `build_attn` in `llm_build_t5_enc`
This commit is contained in:
Juk Armstrong
2025-04-15 07:49:57 +01:00
committed by GitHub
parent eccc7a1602
commit daa422881a
13 changed files with 289 additions and 165 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
src/llama-graph.h
View File

@ -505,11 +505,12 @@ struct llm_graph_context {
ggml_tensor * build_attn_mha(
ggml_cgraph * gf,
ggml_tensor * q, // [n_embd_head_q, n_tokens, n_head_q]
ggml_tensor * k, // [n_embd_head_k, n_tokens, n_head_k]
ggml_tensor * v, // [n_embd_head_v, n_tokens, n_head_v] (v_trans == false)
ggml_tensor * q, // [n_embd_head_q, n_tokens, n_head_q]
ggml_tensor * k, // [n_embd_head_k, n_tokens, n_head_k]
ggml_tensor * v, // [n_embd_head_v, n_tokens, n_head_v] (v_trans == false)
ggml_tensor * kq_b,
ggml_tensor * kq_mask,
ggml_tensor * v_mla, // [n_embd_head_v_mla, n_embd_head_v, n_head_v]
bool v_trans,
float kq_scale) const;
@ -524,6 +525,7 @@ struct llm_graph_context {
ggml_tensor * k_cur, // [n_embd_head_k, n_head_k, n_tokens]
ggml_tensor * v_cur, // [n_embd_head_v, n_head_v, n_tokens]
ggml_tensor * kq_b,
ggml_tensor * v_mla, // [n_embd_head_v_mla, n_embd_head_v, n_head_v]
float kq_scale,
int il) const;
@ -538,6 +540,7 @@ struct llm_graph_context {
ggml_tensor * k_cur, // [n_embd_head_k, n_head_k, n_tokens]
ggml_tensor * v_cur, // [n_embd_head_v, n_head_v, n_tokens]
ggml_tensor * kq_b,
ggml_tensor * v_mla, // [n_embd_head_v_mla, n_embd_head_v, n_head_v]
float kq_scale,
int il) const;
@ -552,6 +555,7 @@ struct llm_graph_context {
ggml_tensor * k_cur, // [n_embd_head_k, n_head_k, n_tokens]
ggml_tensor * v_cur, // [n_embd_head_v, n_head_v, n_tokens]
ggml_tensor * kq_b,
ggml_tensor * v_mla, // [n_embd_head_v_mla, n_embd_head_v, n_head_v]
float kq_scale,
int il) const;