A high-performance inference engine for AI models
pub mod activation_transform;
pub(super) mod trellis;
pub const INT8_SYMMETRIC_QUANTIZATION_MAXIMUM: f32 = 127.0;
pub use activation_transform::quantize_transformed_row;
pub fn min_max_symmetric_divisor(values: &[f32]) -> f32 {
let (min, max) =
values.iter().fold((f32::INFINITY, f32::NEG_INFINITY), |(min, max), &value| (min.min(value), max.max(value)));
let magnitude = min.abs().max(max.abs());
if magnitude.is_finite() && magnitude > 0.0 {
magnitude / INT8_SYMMETRIC_QUANTIZATION_MAXIMUM
} else {
1.0
}
}
pub fn quantize_symmetric_i8(
value: f32,
divisor: f32,
) -> i8 {
(value / divisor).round().clamp(-INT8_SYMMETRIC_QUANTIZATION_MAXIMUM, INT8_SYMMETRIC_QUANTIZATION_MAXIMUM) as i8
}
pub fn hadamard_transform(values: &mut [f32]) {
let mut stride = 1;
while stride < values.len() {
for lane in 0..values.len() {
if lane & stride == 0 {
let a = values[lane];
let b = values[lane | stride];
values[lane] = a + b;
values[lane | stride] = a - b;
}
}
stride <<= 1;
}
let scale = 1.0 / (values.len() as f32).sqrt();
for v in values.iter_mut() {
*v *= scale;
}
}