voice_audio_multi/include/engine/community_models/soulx_flashhead/vae.h
Fabio c36510c21e
Some checks failed
Linux Build / Linux cpu (push) Has been cancelled
Linux Build / Linux cpu-clang (push) Has been cancelled
Linux Build / Linux vulkan (push) Has been cancelled
mac-build / macOS CPU (push) Has been cancelled
Nix Build / Nix (cpu) (push) Has been cancelled
Nix Build / Nix (vulkan) (push) Has been cancelled
windows-build / Windows CPU (push) Has been cancelled
Initial import: audio.cpp with Orange Pi documentation
2026-11-04 19:05:22 +01:00

25 lines
1.1 KiB
C++

#pragma once
#include "engine/community_models/soulx_flashhead/assets.h"
#include "engine/framework/core/execution_context.h"
#include "engine/framework/core/module.h"
namespace engine::community_models::soulx_flashhead {
class VideoVae {
public:
VideoVae(core::ExecutionContext & execution, const Assets & assets, bool cuda_cudnn = false);
~VideoVae();
void reset(uint64_t seed);
std::vector<float> encode(const std::vector<float> & pixels, int64_t frames, int64_t height, int64_t width);
// Valid until the next decode for this shape or destruction.
const std::vector<float> & decode(const std::vector<float> & latents, int64_t frames, int64_t height,
int64_t width);
core::TensorValue decode_device(const std::vector<float> & latents, int64_t frames,
int64_t height, int64_t width);
std::vector<float> encode_device(const core::TensorValue & pixels, int64_t frames,
int64_t height, int64_t width);
private:
struct State;
std::unique_ptr<State> state_;
};
} // namespace engine::community_models::soulx_flashhead