Files
pytorch-lightning/pytorch_lightning/callbacks/gradient_accumulation_scheduler.py
T
Jirka BorovecandWilliam Falcon ff1f8ef400 Test deprecated API for 0.8.0 and 0.9.0 (#1071)
* till 0.8

* refactor

* fix tests

* fix tests

* deprx till 0.9

* Update trainer.py

* Apply suggestions from code review

Co-authored-by: William Falcon <waf2107@columbia.edu>
2020-03-06 21:36:44 +01:00

61 lines
2.1 KiB
Python

r"""
Gradient Accumulator
====================
Change gradient accumulation factor according to scheduling.
"""
import warnings
from .base import Callback
class GradientAccumulationScheduler(Callback):
r"""
Change gradient accumulation factor according to scheduling.
Args:
scheduling: scheduling in format {epoch: accumulation_factor}
.. warning:: Epochs indexing starts from "1" until v0.6.x,
but will start from "0" in v0.8.0.
Example::
from pytorch_lightning import Trainer
from pytorch_lightning.callbacks import GradientAccumulationScheduler
# at epoch 5 start accumulating every 2 batches
accumulator = GradientAccumulationScheduler(scheduling: {5: 2})
Trainer(accumulate_grad_batches=accumulator)
"""
def __init__(self, scheduling: dict):
super().__init__()
if not scheduling: # empty dict error
raise TypeError("Empty dict cannot be interpreted correct")
for key in scheduling:
if not isinstance(key, int) or not isinstance(scheduling[key], int):
raise TypeError("All epoches and accumulation factor must be integers")
minimal_epoch = min(scheduling.keys())
warnings.warn('Epochs indexing of `scheduling` starts from "1" until v0.6.x,'
' but will start from "0" in v0.8.0.', DeprecationWarning)
if minimal_epoch < 1:
msg = f"Epochs indexing from 1, epoch {minimal_epoch} cannot be interpreted correct"
raise IndexError(msg)
if minimal_epoch != 1: # if user didnt define first epoch accumulation factor
scheduling.update({1: 1})
self.scheduling = scheduling
self.epochs = sorted(scheduling.keys())
def on_epoch_start(self, trainer, pl_module):
# indexing epochs from 1 (until v0.6.x)
# In v0.8.0, ` + 1` should be removed.
epoch = trainer.current_epoch + 1
for i in reversed(range(len(self.epochs))):
if epoch >= self.epochs[i]:
trainer.accumulate_grad_batches = self.scheduling.get(self.epochs[i])
break