Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Make .init() method of LR schedulers return Result #2527

Merged
merged 1 commit into from
Nov 25, 2024
Merged

Make .init() method of LR schedulers return Result #2527

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions crates/burn-core/src/lr_scheduler/base.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,5 @@
pub(super) use alloc::string::String;

use burn_tensor::backend::Backend;

use crate::{record::Record, LearningRate};
Expand Down
100 changes: 67 additions & 33 deletions crates/burn-core/src/lr_scheduler/cosine.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
use super::LrScheduler;
use super::{LrScheduler, String};
use crate as burn;
use crate::{config::Config, LearningRate};
use burn_tensor::backend::Backend;
Expand All @@ -24,28 +24,34 @@ pub struct CosineAnnealingLrSchedulerConfig {
impl CosineAnnealingLrSchedulerConfig {
/// Initializes a [Cosine learning rate scheduler](CosineAnnealingLrScheduler).
///
/// # Panics
/// This function panics if `initial_lr` and `final_lr` are not between 0 and 1.
pub fn init(&self) -> CosineAnnealingLrScheduler {
assert!(
self.initial_lr > 0. && self.initial_lr <= 1.,
"Initial learning rate must be greater than 0 and at most 1"
);
assert!(
self.min_lr >= 0.0 && self.min_lr <= self.initial_lr,
"Minimum learning rate must be at least 0 and at most equal to the initial learning rate"
);
assert!(
self.num_iters > 0,
"Number of iterations must be at least 1"
);
/// # Errors
///
/// An error will be returned if any of the following conditions is true:
///
/// * `initial_lr` is out of range (0.0, 1.0]
/// * `min_lr` is out of range [0.0, `initial_lr`]
/// * `num_iters` is 0
pub fn init(&self) -> Result<CosineAnnealingLrScheduler, String> {
if self.initial_lr <= 0. || self.initial_lr > 1. {
return Err("Initial learning rate must be greater than 0 and at most 1".into());
}
if self.min_lr < 0.0 || self.min_lr > self.initial_lr {
return Err(
"Minimum learning rate must be at least 0 and at most equal to the initial \
learning rate"
.into(),
);
}
if self.num_iters == 0 {
return Err("Number of iterations must be at least 1".into());
}

CosineAnnealingLrScheduler {
Ok(CosineAnnealingLrScheduler {
min_lr: self.min_lr,
max_lr: self.initial_lr,
num_iters: self.num_iters,
current_iter: usize::MAX,
}
})
}
}

Expand Down Expand Up @@ -94,48 +100,75 @@ mod tests {
use super::*;

#[test]
#[should_panic = "Initial learning rate must be greater than 0 and at most 1"]
fn config_initial_lr_too_low() {
CosineAnnealingLrSchedulerConfig::new(0., 10).init();
let r = CosineAnnealingLrSchedulerConfig::new(0., 10).init();
assert!(r.is_err(), "Should return an error");
assert_eq!(
r.unwrap_err(),
"Initial learning rate must be greater than 0 and at most 1",
"Error messages should match",
);
}

#[test]
#[should_panic = "Initial learning rate must be greater than 0 and at most 1"]
fn config_initial_lr_too_high() {
CosineAnnealingLrSchedulerConfig::new(1.5, 10).init();
let r = CosineAnnealingLrSchedulerConfig::new(1.5, 10).init();
assert!(r.is_err(), "Should return an error");
assert_eq!(
r.unwrap_err(),
"Initial learning rate must be greater than 0 and at most 1",
"Error messages should match",
);
}

#[test]
#[should_panic = "Minimum learning rate must be at least 0 and at most equal to the initial learning rate"]
fn config_min_lr_too_low() {
CosineAnnealingLrSchedulerConfig::new(0.5, 10)
let r = CosineAnnealingLrSchedulerConfig::new(0.5, 10)
.with_min_lr(-0.1)
.init();
assert!(r.is_err(), "Should return an error");
assert_eq!(
r.unwrap_err(),
"Minimum learning rate must be at least 0 and at most equal to the initial learning \
rate",
"Error messages should match",
);
}

#[test]
#[should_panic = "Minimum learning rate must be at least 0 and at most equal to the initial learning rate"]
fn config_min_lr_too_high() {
CosineAnnealingLrSchedulerConfig::new(0.5, 10)
let r = CosineAnnealingLrSchedulerConfig::new(0.5, 10)
.with_min_lr(0.6)
.init();
assert!(r.is_err(), "Should return an error");
assert_eq!(
r.unwrap_err(),
"Minimum learning rate must be at least 0 and at most equal to the initial learning \
rate",
"Error messages should match",
);
}

#[test]
#[should_panic = "Number of iterations must be at least 1"]
fn config_num_iters_too_low() {
CosineAnnealingLrSchedulerConfig::new(0.5, 0).init();
let r = CosineAnnealingLrSchedulerConfig::new(0.5, 0).init();
assert!(r.is_err(), "Should return an error");
assert_eq!(
r.unwrap_err(),
"Number of iterations must be at least 1",
"Error messages should match",
);
}

#[test]
fn test_lr_change() {
const INITIAL_LR: LearningRate = 0.5;
const MIN_LR: LearningRate = 0.1;
const NUM_ITERS: usize = 2;

let scheduler = CosineAnnealingLrSchedulerConfig::new(INITIAL_LR, NUM_ITERS)
let scheduler = CosineAnnealingLrSchedulerConfig::new(INITIAL_LR, 2)
.with_min_lr(MIN_LR)
.init();
.init()
.unwrap();
let expected_lrs = [
INITIAL_LR, // cos(0)
(INITIAL_LR + MIN_LR) * 0.5, // cos(PI/2)
Expand All @@ -147,9 +180,10 @@ mod tests {

#[test]
fn test_save_and_load() {
const INITIAL_LR: LearningRate = 1.0;
const NUM_ITERS: usize = 9;
let scheduler = CosineAnnealingLrSchedulerConfig::new(INITIAL_LR, NUM_ITERS).init();
let scheduler = CosineAnnealingLrSchedulerConfig::new(1.0, NUM_ITERS)
.init()
.unwrap();
test_utils::check_save_load(scheduler, NUM_ITERS / 3 * 2);
}
}
77 changes: 48 additions & 29 deletions crates/burn-core/src/lr_scheduler/exponential.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
use super::LrScheduler;
use super::{LrScheduler, String};
use crate as burn;
use crate::{config::Config, LearningRate};
use burn_tensor::backend::Backend;
Expand All @@ -19,24 +19,26 @@ pub struct ExponentialLrSchedulerConfig {
impl ExponentialLrSchedulerConfig {
/// Initializes a [exponential learning rate scheduler](ExponentialLrScheduler).
///
/// # Panics
/// This function panics if `initial_lr` and `gamma` are not between 0 and 1.
pub fn init(&self) -> ExponentialLrScheduler {
assert!(
self.initial_lr > 0. && self.initial_lr <= 1.,
"Initial learning rate must be greater than 0 and at most 1"
);
assert!(
self.gamma > 0. && self.gamma <= 1.,
"Gamma must be greater than 0 and at most 1"
);
/// # Errors
///
/// An error will be returned if any of the following conditions is true:
///
/// * `initial_lr` is out of range (0.0, 1.0]
/// * `gamma` is out of range (0.0, 1.0]
pub fn init(&self) -> Result<ExponentialLrScheduler, String> {
if self.initial_lr <= 0. || self.initial_lr > 1. {
return Err("Initial learning rate must be greater than 0 and at most 1".into());
}
if self.gamma <= 0. || self.gamma > 1. {
return Err("Gamma must be greater than 0 and at most 1".into());
}

ExponentialLrScheduler {
Ok(ExponentialLrScheduler {
// Such an initial value eliminates the need for special-case handling of the first
// learning rate.
previous_lr: self.initial_lr / self.gamma,
gamma: self.gamma,
}
})
}
}

Expand Down Expand Up @@ -75,44 +77,61 @@ mod tests {
use super::*;

#[test]
#[should_panic = "Initial learning rate must be greater than 0 and at most 1"]
fn config_initial_lr_too_low() {
ExponentialLrSchedulerConfig::new(0., 0.5).init();
let r = ExponentialLrSchedulerConfig::new(0., 0.5).init();
assert!(r.is_err(), "Should return an error");
assert_eq!(
r.unwrap_err(),
"Initial learning rate must be greater than 0 and at most 1",
"Error messages should match",
);
}

#[test]
#[should_panic = "Initial learning rate must be greater than 0 and at most 1"]
fn config_initial_lr_too_high() {
ExponentialLrSchedulerConfig::new(1.5, 0.5).init();
let r = ExponentialLrSchedulerConfig::new(1.5, 0.5).init();
assert!(r.is_err(), "Should return an error");
assert_eq!(
r.unwrap_err(),
"Initial learning rate must be greater than 0 and at most 1",
"Error messages should match",
);
}

#[test]
#[should_panic = "Gamma must be greater than 0 and at most 1"]
fn config_gamma_too_low() {
ExponentialLrSchedulerConfig::new(0.5, 0.0).init();
let r = ExponentialLrSchedulerConfig::new(0.5, 0.0).init();
assert!(r.is_err(), "Should return an error");
assert_eq!(
r.unwrap_err(),
"Gamma must be greater than 0 and at most 1",
"Error messages should match",
);
}

#[test]
#[should_panic = "Gamma must be greater than 0 and at most 1"]
fn config_gamma_too_high() {
ExponentialLrSchedulerConfig::new(0.5, 1.5).init();
let r = ExponentialLrSchedulerConfig::new(0.5, 1.5).init();
assert!(r.is_err(), "Should return an error");
assert_eq!(
r.unwrap_err(),
"Gamma must be greater than 0 and at most 1",
"Error messages should match",
);
}

#[test]
fn test_lr_change() {
const INITIAL_LR: LearningRate = 0.8;
const GAMMA: f64 = 0.1;

let scheduler = ExponentialLrSchedulerConfig::new(INITIAL_LR, GAMMA).init();
let scheduler = ExponentialLrSchedulerConfig::new(0.8, 0.1).init().unwrap();
let expected_lrs = [0.8, 0.08, 0.008, 0.0008, 0.00008];
test_utils::check_lr_sequence(scheduler, expected_lrs);
}

#[test]
fn test_save_and_load() {
const INITIAL_LR: LearningRate = 0.083;
const GAMMA: f64 = 0.3;
let scheduler = ExponentialLrSchedulerConfig::new(INITIAL_LR, GAMMA).init();
let scheduler = ExponentialLrSchedulerConfig::new(0.083, 0.3)
.init()
.unwrap();
test_utils::check_save_load(scheduler, 7);
}
}
Loading