diff options
| author | Kawrakow <iwankawrakow@gmail.com> | 2025-03-21 07:24:22 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-03-21 07:24:22 +0100 |
| commit | ddc8eee10ee9216de57429167e6f74e618577d93 (patch) | |
| tree | 1dc5d251aa32e1bc6b1e3c4e68ea8619a89a5697 /src/llama.cpp | |
| parent | b8d1fac97b756968b86b470d44bb1026ded7157a (diff) | |
FlashMLA-3: the best of both worlds (CPU only) (#273)
* Repack a model with the quantize tool
* WIP
* Fixed various issues
As we don't have a way to tell if a repacked quant has been modified,
I had to remove the modification at the expense of a slight decrease
in performance. This affects q8_0_r8, q8_KV_r8, q8_k_r8 on Zen4, and
q4_0_r8 on ARM.
* Create wk_b and wv_b as Q8_0_R8 if the wkv_b type is interleaved
* Fix GCC 13.3 compilation error
* Another one
* Add missing include
* FlashMLA-3: the best of both worlds - CPU only
---------
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
Diffstat (limited to 'src/llama.cpp')
| -rw-r--r-- | src/llama.cpp | 4 |
1 files changed, 2 insertions, 2 deletions
diff --git a/src/llama.cpp b/src/llama.cpp index a459cb00..33b69389 100644 --- a/src/llama.cpp +++ b/src/llama.cpp @@ -13760,7 +13760,7 @@ struct llm_build_context { ggml_tensor * kqv; - if (lctx.cparams.mla_attn > 1 && lctx.cparams.flash_attn && (pp_opt || lctx.cparams.mla_attn > 2)) { + if (lctx.cparams.mla_attn > 1 && lctx.cparams.flash_attn && pp_opt) { // PP for mla=2,3 auto kv_cache_nope = ggml_view_2d(ctx0, kv_self.kv_l[il], kv_lora_rank, n_kv, kv_self.kv_l[il]->nb[1], 0); @@ -13869,7 +13869,7 @@ struct llm_build_context { ggml_tensor * q = ggml_concat(ctx0, q_nope2, ggml_permute(ctx0, q_rope, 0, 2, 1, 3), 0); cb(q, "q", il); - if (lctx.cparams.flash_attn && lctx.cparams.mla_attn == 1) { + if (lctx.cparams.flash_attn && lctx.cparams.mla_attn == 1 || lctx.cparams.mla_attn == 3) { ggml_tensor * kv_cache_lora = ggml_view_2d(ctx0, kv_self.kv_l[il], kv_lora_rank, n_kv, ggml_row_size(kv_self.kv_l[il]->type, kv_lora_rank + n_embd_head_qk_rope), 0); |