voice_audio_multi/include/engine/models/pocket_tts/audio_decoder.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

52 lines
1.4 KiB
C++

#pragma once
#include "engine/models/pocket_tts/mimi_decoder.h"
#include "ggml-backend.h"
#include <cstddef>
#include <cstdint>
#include <vector>
namespace engine::models::pocket_tts {
struct PocketTTSAssets;
struct PocketTTSBackendWeights;
class AudioDecoder {
public:
explicit AudioDecoder(MimiDecoderConfig config = {});
std::vector<float> decode(
ggml_backend_t backend,
int threads,
const PocketTTSAssets & manifest,
const PocketTTSBackendWeights & weights,
const std::vector<float> & normalized_latents,
int64_t steps,
size_t conv_graph_context_bytes,
size_t transformer_graph_context_bytes,
size_t tail_graph_context_bytes,
int64_t full_chunk_frames,
int64_t stage2_chunk_frames,
bool use_full_sequence_path) const;
void reset_streaming_state() const;
std::vector<float> decode_streaming_step(
ggml_backend_t backend,
int threads,
const PocketTTSAssets & manifest,
const PocketTTSBackendWeights & weights,
const std::vector<float> & normalized_latent,
size_t conv_graph_context_bytes,
size_t transformer_graph_context_bytes,
size_t tail_graph_context_bytes) const;
void clear_runtime_cache() const noexcept;
private:
MimiDecoder decoder_;
};
} // namespace engine::models::pocket_tts