voice_audio_multi/include/engine/community_models/parakeet_tdt/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

99 lines
3.4 KiB
C++

#pragma once
#include "engine/framework/core/execution_context.h"
#include "engine/framework/runtime/session.h"
#include "engine/community_models/parakeet_tdt/assets.h"
#include "engine/community_models/parakeet_tdt/encoder.h"
#include "engine/community_models/parakeet_tdt/weights.h"
#include "engine/community_models/parakeet_tdt/hotwords.h"
#include <cstdint>
#include <array>
#include <functional>
#include <memory>
#include <string>
#include <vector>
namespace engine::community_models::parakeet_tdt {
struct ParakeetDecodeOptions {
int64_t max_tokens = 0;
bool keep_language_tags = false;
int64_t audio_end_sample = -1;
std::shared_ptr<const PhononHotwordAutomaton> hotwords;
};
struct ParakeetDecodedText {
std::string text;
std::vector<int32_t> token_ids;
std::vector<int32_t> token_frame_indices;
std::vector<int32_t> durations;
std::vector<runtime::WordTimestamp> word_timestamps;
};
class ParakeetTDTDecoderRuntime {
public:
ParakeetTDTDecoderRuntime(
std::shared_ptr<const ParakeetTDTAssets> assets,
std::shared_ptr<const ParakeetWeights> weights,
engine::core::ExecutionContext & execution_context,
size_t graph_arena_bytes);
~ParakeetTDTDecoderRuntime();
void prepare();
ParakeetDecodedText decode(
const ParakeetEncodedAudio & encoded,
const ParakeetDecodeOptions & options);
void reset_state();
ParakeetDecodedText decode_incremental(
const ParakeetEncodedAudio & encoded,
const ParakeetDecodeOptions & options,
int64_t frame_offset = 0);
ParakeetDecodedText format_tokens(
std::vector<int32_t> token_ids,
std::vector<int32_t> token_frame_indices,
std::vector<int32_t> durations,
const ParakeetDecodeOptions & options,
int64_t audio_end_frame) const;
int32_t run_step(int32_t input_token, const float * encoder_frame, bool decoder_cache_valid, int32_t * out_dur_id = nullptr);
int32_t run_joint_step(const float * encoder_frame, int32_t * out_dur_id = nullptr);
private:
struct StepGraph;
struct JointGraph;
std::shared_ptr<const ParakeetTDTAssets> assets_;
std::shared_ptr<const ParakeetWeights> weights_;
engine::core::ExecutionContext * execution_context_ = nullptr;
size_t graph_arena_bytes_ = 0;
std::array<int64_t, 2> state_widths_{};
std::unique_ptr<StepGraph> step_graph_;
std::unique_ptr<JointGraph> joint_graph_;
std::vector<float> logits_scratch_;
std::vector<float> hidden_scratch_;
std::vector<float> cell_scratch_;
std::vector<float> decoder_cache_scratch_;
std::vector<float> hidden_read_scratch_;
std::vector<float> cell_read_scratch_;
int32_t pending_input_token_ = 0;
bool predictor_cache_valid_ = false;
bool state_initialized_ = false;
std::shared_ptr<const PhononHotwordAutomaton> hotwords_;
int32_t hotword_state_ = 0;
int32_t select_token();
void ensure_step_graph();
void ensure_joint_graph();
std::string decode_text(const std::vector<int32_t> & token_ids, bool keep_language_tags) const;
std::vector<runtime::WordTimestamp> build_word_timestamps(
const std::vector<int32_t> & token_ids,
const std::vector<int32_t> & token_frame_indices,
const std::vector<int32_t> & durations,
int64_t audio_end_frame,
int64_t audio_end_sample) const;
};
} // namespace engine::community_models::parakeet_tdt