Decode back to waveform. Usage: 1) Pass the raw output of `encode(...)` directly (recommended). - 25Hz: expects f

Voice clone speech using the Base model. You can provide either: - (ref_audio, ref_text, x_vector_only_mode) and let this

Batch-encode audio into discrete codes (and optional conditioning, depending on 25Hz/12Hz). Args: audios (AudioInput):

Load a Qwen3 TTS model and its processor in HuggingFace `from_pretrained` style. This method: 1) Loads config via AutoConf

Build voice-clone prompt items from reference audio (and optionally reference text) using Base model. Modes: - x_vector_on

Rotates half the hidden dims of the input.

Merge user-provided generation arguments with defaults from `generate_config.json`. Rule: - If the user explicitly passes

Validate that requested languages are supported by the model. Args: languages (List[str]): Language names for each sampl

Generate speech with the CustomVoice model using a predefined speaker id, optionally controlled by instruction text. Args:

Generate speech with the VoiceDesign model using natural-language style instructions. Args: text: Text(s

Generates a 1D Kaiser-windowed sinc filter. Args: cutoff (float): Normalized cutoff frequency (0 to 0.5). half_width (float): Tra

Compute the log-Mel spectrogram of Parameters ---------- audio: Union[str, np.ndarray, torch.Tensor], shape = (*) The path t

Calculate the mel spectrogram of an input signal. This function uses slaney norm for the librosa mel filterbank (using librosa.filters.mel) a

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

This is the equivalent of torch.repeat_interleave(x, dim=1, repeats=n_rep). The hidden states go from (batch, num_key_value_heads, seqlen, he

Rotates half the hidden dims of the input.

x: torch.Tensor, shape = (T, D) q: torch.Tensor, shape = (T, D) i: torch.Tensor, shape = (T)

Creates a binary mask for each sequence. Reference: https://discuss.pytorch.org/t/how-to-generate-variable-length-mask/23397/3 Argum

Creates a binary mask for each sequence. Reference: https://discuss.pytorch.org/t/how-to-generate-variable-length-mask/23397/3 Argum

Normalize audio inputs into a list of (waveform, sr). Supported forms: - str: wav path / URL / base64 audio string

Normalize audio inputs into a list of (waveform, sr). Supported forms: - str: wav path / URL / base64 audio string

Normalize all supported input types into a list of 1-D numpy float32 waveforms at `self.feature_extractor.sampling_rate`. Ar

Validate that requested speakers are supported by the Instruct model. Args: speakers (List[Optional[str]]): Speaker name

Applies Rotary Position Embedding with Multimodal Sections to the query and key tensors (https://qwenlm.github.io/blog/qwen2-vl/). Explanation:

Applies Rotary Position Embedding to the query and key tensors. Args: q (`torch.Tensor`): The query tensor. k (`torch.Tensor`): T

Applies Rotary Position Embedding to the query and key tensors. Args: q (`torch.Tensor`): The query tensor. k (`torch.Tensor`): T

Applies Rotary Position Embedding to the query and key tensors. Args: q (`torch.Tensor`): The query tensor. k (`torch.Tensor`): T

Download model weights from Hugging Face Hub. Users can specify the allow_patterns to download only the necessary weights. Args: mode