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
39 lines
1.2 KiB
C++
39 lines
1.2 KiB
C++
#pragma once
|
|
|
|
#include "engine/framework/assets/tensor_source.h"
|
|
#include "engine/framework/core/execution_context.h"
|
|
#include "engine/models/fun_asr_nano/assets.h"
|
|
#include "engine/models/fun_asr_nano/types.h"
|
|
|
|
#include <cstddef>
|
|
#include <functional>
|
|
#include <memory>
|
|
|
|
namespace engine::models::fun_asr_nano {
|
|
|
|
using FunAsrNanoTokenCallback =
|
|
std::function<void(const FunAsrNanoGeneratedTokens &)>;
|
|
|
|
class FunAsrNanoQwen3DecoderRuntime {
|
|
public:
|
|
struct Impl;
|
|
|
|
FunAsrNanoQwen3DecoderRuntime(std::shared_ptr<const FunAsrNanoAssets> assets,
|
|
core::ExecutionContext &execution,
|
|
size_t prefill_graph_arena_bytes,
|
|
size_t decode_graph_arena_bytes,
|
|
size_t weight_context_bytes,
|
|
assets::TensorStorageType weight_storage_type);
|
|
~FunAsrNanoQwen3DecoderRuntime();
|
|
|
|
FunAsrNanoGeneratedTokens
|
|
generate(const FunAsrNanoPrompt &prompt,
|
|
const FunAsrNanoAdaptorEmbeddings &audio_embeddings,
|
|
const FunAsrNanoGenerationOptions &options,
|
|
const FunAsrNanoTokenCallback &token_callback = {});
|
|
|
|
private:
|
|
std::unique_ptr<Impl> impl_;
|
|
};
|
|
|
|
} // namespace engine::models::fun_asr_nano
|