DarthAffe/stable-diffusion.cpp
1
0 Fork 0
mirror of https://github.com/leejet/stable-diffusion.cpp.git synced 2026-11-11 11:26:34 +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/lens.hpp
+15 -10
View File
@@ -4,6 +4,7 @@
#include <cinttypes>
#include <memory>
#include <vector>
#include "core/pe_cache.h"
#include "model/common/block.hpp"
#include "model/common/rope.hpp"
@@ -352,9 +353,9 @@ namespace Lens {
};
struct LensRunner : public DiffusionModelRunner {
sd::PECache pe_cache_;
LensConfig config;
LensModel lens;
std::vector<float> pe_vec;
LensRunner(ggml_backend_t backend,
const String2TensorStorage& tensor_storage_map = {},
@@ -384,15 +385,19 @@ namespace Lens {
GGML_ASSERT(!context_tensor.empty());
ggml_tensor* context = make_input(context_tensor);
pe_vec = finish_rope_pe(Rope::gen_lens_pe(static_cast<int>(x->ne[1]),
static_cast<int>(x->ne[0]),
static_cast<int>(x->ne[3]),
static_cast<int>(context->ne[1]),
config.theta,
config.axes_dim));
int pos_len = static_cast<int>(pe_vec.size() / config.axes_dim_sum / 2);
auto pe = ggml_new_tensor_4d(compute_ctx, GGML_TYPE_F32, 2, 2, config.axes_dim_sum / 2, pos_len);
set_backend_tensor_data(pe, pe_vec.data());
sd::PECacheKey pe_key("lens.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);
auto pe = make_input(pe_cache_.get_or_create(pe_key, [&]() {
auto pe_vec = finish_rope_pe(Rope::gen_lens_pe(static_cast<int>(x->ne[1]),
static_cast<int>(x->ne[0]),
static_cast<int>(x->ne[3]),
static_cast<int>(context->ne[1]),
config.theta,
config.axes_dim));
int pos_len = static_cast<int>(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();
ggml_tensor* out = lens.forward(&runner_ctx, x, timesteps, context, pe);