Struct ConvNetModel 

pub struct ConvNetModel {
    pub blocks: Vec<ConvNetBlock>,
    pub head_scale: f32,
    pub receptive_field_size: usize,
    pub post_stack_head: Option<PostStackHead>,
    pub head_output_scratch: AlignedVec<f32>,
    pub prewarm_on_reset: bool,
    pub linear_head: Option<LinearHead>,
    /* private fields */
}

ConvNet feed-forward model.

Composed of a sequence of ConvNetBlocks chained sequentially, followed by an optional PostStackHead and a final head_scale gain.

Unlike WaveNet, ConvNet has no gating, no rechannel projections, no condition_dsp, and no dual-array architecture. Each block’s output is the input of the next block directly.

Fields

blocks: Vec<ConvNetBlock>

Sequential ConvNet blocks.

head_scale: f32

Final voltage compensation scale (Target Output Scale).

receptive_field_size: usize

Total receptive field (sum of all block RFs + head RF contribution).

post_stack_head: Option<PostStackHead>

Optional post-stack head sub-object (Conv1D + activation).

head_output_scratch: AlignedVec<f32>

Scratch buffer for post-stack head output.

prewarm_on_reset: bool

Whether to execute prewarm during reset(). Default: true.

linear_head: Option<LinearHead>

Optional C++ flat-format linear head weights (no activation). Used when post_stack_head is None but the model has a separate linear projection from in_ch → out_ch.

Implementations

impl ConvNetModel

pub fn in_channels(&self) -> usize

Returns the number of input channels expected by the first block.

pub fn out_channels(&self) -> usize

Returns the number of output channels produced by the model.

pub fn process(&mut self, input: &[f32], output: &mut [f32])

Resolves the full forward pass and produces waveform samples in zero allocation (DSP).

pub fn prewarm(&mut self)

Stabilizes the model by processing silence (Zero Input) for pre-warm.

Trait Implementations

impl NamModel for ConvNetModel

fn process(&mut self, input: &[f32], output: &mut [f32])

Invoked by the DSP RT-Thread to process acoustic sample blocks (Float32).

fn prewarm(&mut self, _num_samples: usize)

“Heats up” the virtual tubes of the neural engine (prewarm).

fn prewarm_samples(&self) -> usize

Returns the number of samples needed to fully stabilize the model’s internal state (receptive field / recurrent memory depth).

fn set_max_buffer_size(&mut self, _max_buf: usize) -> Result<()>

Reallocates internal buffers to support the given maximum block size.

fn prewarm_on_reset(&self) -> bool

Returns whether prewarm should be executed on reset().

fn set_prewarm_on_reset(&mut self, val: bool)

Sets whether prewarm should be executed on reset().

fn reset(&mut self, _sample_rate: u32, max_buffer_size: usize) -> Result<()>

Resets the model’s internal state with a new sample rate and max buffer size.

fn slimmable_breakpoints(&self) -> Vec<f64>

Returns the breakpoints at which slimmable quality transitions occur.