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
75 lines
2.2 KiB
C++
75 lines
2.2 KiB
C++
#pragma once
|
|
|
|
#include "engine/framework/core/backend.h"
|
|
#include "engine/framework/core/execution_context.h"
|
|
#include "engine/framework/core/module.h"
|
|
#include "engine/framework/modules/text_encoders/t5_gemma_encoder.h"
|
|
#include "engine/models/stable_audio/assets.h"
|
|
#include "engine/models/stable_audio/conditioner.h"
|
|
|
|
#include <ggml-backend.h>
|
|
|
|
#include <cstdint>
|
|
#include <memory>
|
|
#include <vector>
|
|
|
|
namespace engine::core {
|
|
class BackendWeightStore;
|
|
}
|
|
|
|
namespace engine::models::stable_audio {
|
|
|
|
struct StableAudioT5Weights {
|
|
std::shared_ptr<core::BackendWeightStore> store;
|
|
modules::T5GemmaEncoderWeights encoder;
|
|
};
|
|
|
|
StableAudioT5Weights load_stable_audio_t5_weights(
|
|
const StableAudioAssets & assets,
|
|
ggml_backend_t backend,
|
|
core::BackendType backend_type,
|
|
size_t weight_context_bytes,
|
|
assets::TensorStorageType weight_storage_type);
|
|
|
|
core::TensorValue build_stable_audio_t5_encoder(
|
|
core::ModuleBuildContext & ctx,
|
|
const core::TensorValue & input_ids,
|
|
const core::TensorValue & positions,
|
|
const core::TensorValue & additive_attention_mask,
|
|
const StableAudioT5Weights & weights,
|
|
const StableAudioConfig & config);
|
|
|
|
struct StableAudioT5Encoding {
|
|
int64_t batch = 0;
|
|
int64_t tokens = 0;
|
|
int64_t hidden_size = 0;
|
|
std::vector<float> values;
|
|
};
|
|
|
|
class StableAudioT5EncoderRuntime {
|
|
public:
|
|
StableAudioT5EncoderRuntime(
|
|
core::ExecutionContext & execution,
|
|
std::shared_ptr<const StableAudioAssets> assets,
|
|
assets::TensorStorageType weight_storage_type,
|
|
int64_t max_batch);
|
|
~StableAudioT5EncoderRuntime();
|
|
|
|
StableAudioT5EncoderRuntime(const StableAudioT5EncoderRuntime &) = delete;
|
|
StableAudioT5EncoderRuntime & operator=(const StableAudioT5EncoderRuntime &) = delete;
|
|
|
|
void prepare() const;
|
|
StableAudioT5Encoding encode(const StableAudioTokenBatch & tokens) const;
|
|
void release_runtime_graphs() const;
|
|
|
|
private:
|
|
class Graph;
|
|
|
|
core::ExecutionContext * execution_ = nullptr;
|
|
std::shared_ptr<const StableAudioAssets> assets_;
|
|
assets::TensorStorageType weight_storage_type_ = assets::TensorStorageType::Native;
|
|
int64_t max_batch_ = 1;
|
|
mutable std::unique_ptr<Graph> graph_;
|
|
};
|
|
|
|
} // namespace engine::models::stable_audio
|