#[repr(i32)]pub enum QuantizationType {
Q4_0 = 2,
Q4_1 = 3,
Q5_0 = 8,
Q5_1 = 9,
Q8_0 = 7,
Q2_K = 10,
Q3_K = 11,
Q4_K = 12,
Q5_K = 13,
Q6_K = 14,
}Expand description
Quantization types supported by whisper.cpp
Variants§
Q4_0 = 2
4-bit quantization (method 0) - ~3.5 GB for base model
Q4_1 = 3
4-bit quantization (method 1) - ~3.9 GB for base model
Q5_0 = 8
5-bit quantization (method 0) - ~4.3 GB for base model
Q5_1 = 9
5-bit quantization (method 1) - ~4.7 GB for base model
Q8_0 = 7
8-bit quantization - ~7.7 GB for base model
Q2_K = 10
2-bit k-quantization
Q3_K = 11
3-bit k-quantization
Q4_K = 12
4-bit k-quantization
Q5_K = 13
5-bit k-quantization
Q6_K = 14
6-bit k-quantization
Implementations§
Source§impl QuantizationType
impl QuantizationType
Sourcepub fn size_factor(&self) -> f32
pub fn size_factor(&self) -> f32
Estimate the size reduction factor for this quantization type. Returns the approximate size as a fraction of the original F32 model.
Sourcepub fn all() -> &'static [QuantizationType]
pub fn all() -> &'static [QuantizationType]
Get all available quantization types
Trait Implementations§
Source§impl Clone for QuantizationType
impl Clone for QuantizationType
Source§fn clone(&self) -> QuantizationType
fn clone(&self) -> QuantizationType
Returns a duplicate of the value. Read more
1.0.0 · Source§fn clone_from(&mut self, source: &Self)
fn clone_from(&mut self, source: &Self)
Performs copy-assignment from
source. Read moreSource§impl Debug for QuantizationType
impl Debug for QuantizationType
Source§impl Display for QuantizationType
impl Display for QuantizationType
Source§impl FromStr for QuantizationType
impl FromStr for QuantizationType
Source§impl PartialEq for QuantizationType
impl PartialEq for QuantizationType
impl Copy for QuantizationType
impl Eq for QuantizationType
impl StructuralPartialEq for QuantizationType
Auto Trait Implementations§
impl Freeze for QuantizationType
impl RefUnwindSafe for QuantizationType
impl Send for QuantizationType
impl Sync for QuantizationType
impl Unpin for QuantizationType
impl UnwindSafe for QuantizationType
Blanket Implementations§
Source§impl<T> BorrowMut<T> for Twhere
T: ?Sized,
impl<T> BorrowMut<T> for Twhere
T: ?Sized,
Source§fn borrow_mut(&mut self) -> &mut T
fn borrow_mut(&mut self) -> &mut T
Mutably borrows from an owned value. Read more