DarthAffe/stable-diffusion.cpp
1
0 Fork 0
mirror of https://github.com/leejet/stable-diffusion.cpp.git synced 2026-11-11 03:24:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

perf: cache CPU position embeddings across DiT runners (#2211)

Browse Source
This commit is contained in:
leejet authored and GitHub committed 2026-10-25 05:18:30 +08:00
1 parent a85db96aa5
commit ea209dbfcf
31 files changed
+855 -610
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

src/model/diffusion/hunyuan.hpp
+21 -18
View File
@@ -3,6 +3,7 @@
#include <cinttypes>
#include <memory>
#include "core/pe_cache.h"
#include "model/common/block.hpp"
#include "model/diffusion/flux.hpp"
@@ -540,10 +541,11 @@ namespace Hunyuan {
};
struct HunyuanVideoRunner : public DiffusionModelRunner {
sd::PECache pe_cache_;
public:
HunyuanVideoConfig config;
HunyuanVideoModel hunyuan_video;
std::vector<float> pe_vec;
SDVersion version;
HunyuanVideoRunner(ggml_backend_t backend,
@@ -605,23 +607,24 @@ namespace Hunyuan {
if (vision != nullptr) {
text_len += static_cast<int>(vision->ne[1]);
}
pe_vec = Rope::gen_hunyuan_video_pe(static_cast<int>(x->ne[2]),
static_cast<int>(x->ne[1]),
static_cast<int>(x->ne[0]),
std::get<0>(config.patch_size),
std::get<1>(config.patch_size),
std::get<2>(config.patch_size),
1,
text_len,
config.theta,
config.axes_dim);
int64_t pos_len = static_cast<int64_t>(pe_vec.size() / config.axes_dim_sum / 2);
// LOG_VERBOSE("pos_len %d", pos_len);
auto pe = ggml_new_tensor_4d(compute_ctx, GGML_TYPE_F32, 2, 2, config.axes_dim_sum / 2, pos_len);
// pe->data = pe_vec.data();
// print_ggml_tensor(pe, true, "pe");
// pe->data = nullptr;
set_backend_tensor_data(pe, pe_vec.data());
sd::PECacheKey pe_key("hunyuan.pe");
pe_key.add_shape(x).add_shape(context).add(config.theta).add(config.axes_dim).add(config.axes_dim_sum);
pe_key.add(circular_x_enabled).add(circular_y_enabled).add(std::get<0>(config.patch_size));
pe_key.add(std::get<1>(config.patch_size)).add(std::get<2>(config.patch_size)).add(text_len);
auto pe = make_input(pe_cache_.get_or_create(pe_key, [&]() {
auto pe_vec = Rope::gen_hunyuan_video_pe(static_cast<int>(x->ne[2]),
static_cast<int>(x->ne[1]),
static_cast<int>(x->ne[0]),
std::get<0>(config.patch_size),
std::get<1>(config.patch_size),
std::get<2>(config.patch_size),
1,
text_len,
config.theta,
config.axes_dim);
int64_t pos_len = static_cast<int64_t>(pe_vec.size() / config.axes_dim_sum / 2);
return sd::Tensor<float>({2, 2, config.axes_dim_sum / 2, pos_len}, std::move(pe_vec));
}));
auto runner_ctx = get_context();