Qwen3ASROptions

Options & Configuration · AiDotNet.SpeechRecognition.AlibabaASR

Options for Qwen3-ASR-1.7B: an AuT audio encoder, a projector and the Qwen3-1.7B decoder (Qwen3-ASR technical report, 2026).

Inheritance: ModelOptions → Qwen3ASROptions

For Beginners

Defaults are the released Qwen3-ASR-1.7B configuration: 128-band Whisper log-mel features, an AuT encoder of 24 layers at width 1024 (16 heads, feed-forward 4096) that emits 12.5 tokens per second, and Qwen3-1.7B (2,048 wide, 28 layers, 16 query and 8 key-value heads of width 128, SwiGLU 6,144, 151,936 tokens, tied embeddings). Fine-tuning uses AdamW at 2e-5 with a linear schedule, as the release’s SFT script does.

Constructors

Constructor Summary
Qwen3ASROptions() Initializes a new instance with default values.
Qwen3ASROptions(Qwen3ASROptions other) Initializes a new instance by copying from another instance.

Properties

Property Summary
OnnxModelOptions OnnxOptions { get; set; } Gets or sets the ONNX Runtime settings used when the model is loaded from an ONNX file.
double LlmRmsEpsilon { get; set; } Gets or sets the RMS-normalization stability constant for the language-model.
double LlmRopeTheta { get; set; } Gets or sets the rotary positional-embedding frequency base for the language-model.
int ChunkFrames { get; set; } Spectrogram frames per convolution chunk (100, one second).
int DownsampleChannels { get; set; } The convolutional front end’s channels (480).
int EncoderDim { get; set; } The AuT encoder’s width (1024).
int EncoderFeedForwardDim { get; set; } The AuT encoder’s feed-forward width (4096).
int LlmHeadDim { get; set; } Qwen3-1.7B’s head width (128).
int LlmHeads { get; set; } Qwen3-1.7B’s query heads (16).
int LlmHiddenDim { get; set; } Qwen3-1.7B’s width (2,048).
int LlmIntermediateDim { get; set; } Qwen3-1.7B’s SwiGLU width (6,144).
int LlmKeyValueHeads { get; set; } Qwen3-1.7B’s key-value heads (8).
int MaxAudioLengthSeconds { get; set; } The longest recording transcribed (20 minutes, the report’s single-utterance limit).
int MaxTextLength { get; set; } The longest answer generated, in tokens.
int NumAttentionHeads { get; set; } The AuT encoder’s attention heads (16).
int NumEncoderLayers { get; set; } The AuT encoder’s layers (24).
int NumLlmLayers { get; set; } Qwen3-1.7B’s layers (28).
int NumMels { get; set; } Whisper log-mel bands (128).
int SampleRate { get; set; } Gets or sets the sample rate, in Hz, that input audio is expected at.
int VocabSize { get; set; } The decoder’s vocabulary (151,936).
int WarmupSteps { get; set; } Linear warmup steps.
int WindowChunks { get; set; } Chunks per encoder attention window (8).
string Context { get; set; } Context text (hotwords, background) placed in the system message; empty by default.
string Language { get; set; } The language to force (“English”, “Chinese”, …): the prompt then ends with language {Language}<asr_text> and the answer is the transcript alone.
string ModelPath { get; set; } Gets or sets the path to the model weights.
string[] Vocabulary { get; set; } Optional vocabulary mapping token ids to text when the vocabulary is not Qwen’s.