2023-01-01 22:54:49 +00:00
|
|
|
import numpy as np
|
|
|
|
|
|
|
|
|
|
|
|
class LambdaWarmUpCosineScheduler:
|
|
|
|
"""
|
|
|
|
note: use with a base_lr of 1.0.
|
|
|
|
"""
|
|
|
|
|
|
|
|
def __init__(
|
|
|
|
self,
|
|
|
|
warm_up_steps,
|
|
|
|
lr_min,
|
|
|
|
lr_max,
|
|
|
|
lr_start,
|
|
|
|
max_decay_steps,
|
|
|
|
verbosity_interval=0,
|
|
|
|
):
|
|
|
|
self.lr_warm_up_steps = warm_up_steps
|
|
|
|
self.lr_start = lr_start
|
|
|
|
self.lr_min = lr_min
|
|
|
|
self.lr_max = lr_max
|
|
|
|
self.lr_max_decay_steps = max_decay_steps
|
|
|
|
self.last_lr = 0.0
|
|
|
|
self.verbosity_interval = verbosity_interval
|
|
|
|
|
|
|
|
def schedule(self, n, **kwargs):
|
2023-09-29 08:13:50 +00:00
|
|
|
if self.verbosity_interval > 0 and n % self.verbosity_interval == 0:
|
|
|
|
print(f"current step: {n}, recent lr-multiplier: {self.last_lr}")
|
2023-01-01 22:54:49 +00:00
|
|
|
if n < self.lr_warm_up_steps:
|
|
|
|
lr = (
|
|
|
|
self.lr_max - self.lr_start
|
|
|
|
) / self.lr_warm_up_steps * n + self.lr_start
|
|
|
|
self.last_lr = lr
|
|
|
|
return lr
|
|
|
|
t = (n - self.lr_warm_up_steps) / (
|
|
|
|
self.lr_max_decay_steps - self.lr_warm_up_steps
|
|
|
|
)
|
|
|
|
t = min(t, 1.0)
|
|
|
|
lr = self.lr_min + 0.5 * (self.lr_max - self.lr_min) * (1 + np.cos(t * np.pi))
|
|
|
|
self.last_lr = lr
|
|
|
|
return lr
|
|
|
|
|
|
|
|
def __call__(self, n, **kwargs):
|
|
|
|
return self.schedule(n, **kwargs)
|
|
|
|
|
|
|
|
|
|
|
|
class LambdaWarmUpCosineScheduler2:
|
|
|
|
"""
|
|
|
|
supports repeated iterations, configurable via lists
|
|
|
|
note: use with a base_lr of 1.0.
|
|
|
|
"""
|
|
|
|
|
|
|
|
def __init__(
|
|
|
|
self, warm_up_steps, f_min, f_max, f_start, cycle_lengths, verbosity_interval=0
|
|
|
|
):
|
|
|
|
assert (
|
|
|
|
len(warm_up_steps)
|
|
|
|
== len(f_min)
|
|
|
|
== len(f_max)
|
|
|
|
== len(f_start)
|
|
|
|
== len(cycle_lengths)
|
|
|
|
)
|
|
|
|
self.lr_warm_up_steps = warm_up_steps
|
|
|
|
self.f_start = f_start
|
|
|
|
self.f_min = f_min
|
|
|
|
self.f_max = f_max
|
|
|
|
self.cycle_lengths = cycle_lengths
|
2023-09-29 08:13:50 +00:00
|
|
|
self.cum_cycles = np.cumsum([0, *list(self.cycle_lengths)])
|
2023-01-01 22:54:49 +00:00
|
|
|
self.last_f = 0.0
|
|
|
|
self.verbosity_interval = verbosity_interval
|
|
|
|
|
|
|
|
def find_in_interval(self, n):
|
|
|
|
interval = 0
|
|
|
|
for cl in self.cum_cycles[1:]:
|
|
|
|
if n <= cl:
|
|
|
|
return interval
|
|
|
|
interval += 1
|
|
|
|
return None
|
|
|
|
|
|
|
|
def schedule(self, n, **kwargs):
|
|
|
|
cycle = self.find_in_interval(n)
|
|
|
|
n = n - self.cum_cycles[cycle]
|
2023-09-29 08:13:50 +00:00
|
|
|
if self.verbosity_interval > 0 and n % self.verbosity_interval == 0:
|
|
|
|
print(
|
|
|
|
f"current step: {n}, recent lr-multiplier: {self.last_f}, "
|
|
|
|
f"current cycle {cycle}"
|
|
|
|
)
|
2023-01-01 22:54:49 +00:00
|
|
|
if n < self.lr_warm_up_steps[cycle]:
|
|
|
|
f = (self.f_max[cycle] - self.f_start[cycle]) / self.lr_warm_up_steps[
|
|
|
|
cycle
|
|
|
|
] * n + self.f_start[cycle]
|
|
|
|
self.last_f = f
|
|
|
|
return f
|
|
|
|
|
|
|
|
t = (n - self.lr_warm_up_steps[cycle]) / (
|
|
|
|
self.cycle_lengths[cycle] - self.lr_warm_up_steps[cycle]
|
|
|
|
)
|
|
|
|
t = min(t, 1.0)
|
|
|
|
f = self.f_min[cycle] + 0.5 * (self.f_max[cycle] - self.f_min[cycle]) * (
|
|
|
|
1 + np.cos(t * np.pi)
|
|
|
|
)
|
|
|
|
self.last_f = f
|
|
|
|
return f
|
|
|
|
|
|
|
|
def __call__(self, n, **kwargs):
|
|
|
|
return self.schedule(n, **kwargs)
|
|
|
|
|
|
|
|
|
|
|
|
class LambdaLinearScheduler(LambdaWarmUpCosineScheduler2):
|
|
|
|
def schedule(self, n, **kwargs):
|
|
|
|
cycle = self.find_in_interval(n)
|
|
|
|
n = n - self.cum_cycles[cycle]
|
2023-09-29 08:13:50 +00:00
|
|
|
if self.verbosity_interval > 0 and n % self.verbosity_interval == 0:
|
|
|
|
print(
|
|
|
|
f"current step: {n}, recent lr-multiplier: {self.last_f}, "
|
|
|
|
f"current cycle {cycle}"
|
|
|
|
)
|
2023-01-01 22:54:49 +00:00
|
|
|
|
|
|
|
if n < self.lr_warm_up_steps[cycle]:
|
|
|
|
f = (self.f_max[cycle] - self.f_start[cycle]) / self.lr_warm_up_steps[
|
|
|
|
cycle
|
|
|
|
] * n + self.f_start[cycle]
|
|
|
|
self.last_f = f
|
|
|
|
return f
|
|
|
|
|
|
|
|
f = self.f_min[cycle] + (self.f_max[cycle] - self.f_min[cycle]) * (
|
|
|
|
self.cycle_lengths[cycle] - n
|
|
|
|
) / (self.cycle_lengths[cycle])
|
|
|
|
self.last_f = f
|
|
|
|
return f
|