Skip to content

Make tbptt imports Python 3.10 compatible #13973

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

Merged
merged 2 commits into from
Aug 2, 2022
Merged

Make tbptt imports Python 3.10 compatible #13973

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
bcfdbdf
Make tbptt imports Python 3.10 compatible
awaelchli Aug 2, 2022
e86b081
add chlog
awaelchli Aug 2, 2022
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
3 changes: 3 additions & 0 deletions src/pytorch_lightning/CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -402,6 +402,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
- Fixed default `amp_level` for `DeepSpeedPrecisionPlugin` to `O2` ([#13897](https://github.com/PyTorchLightning/pytorch-lightning/pull/13897))


- Fixed Python 3.10 compatibility for truncated back-propagation through time (TBPTT) ([#13973](https://github.com/Lightning-AI/lightning/pull/13973))



## [1.6.5] - 2022-07-13

Expand Down
8 changes: 4 additions & 4 deletions src/pytorch_lightning/core/module.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -13,7 +13,7 @@
# limitations under the License.
"""The LightningModule - an nn.Module with many additional features."""

import collections
import collections.abc
import inspect
import logging
import numbers
Expand Down Expand Up @@ -1712,7 +1712,7 @@ def tbptt_split_batch(self, batch, split_size):
for i, x in enumerate(batch):
if isinstance(x, torch.Tensor):
split_x = x[:, t:t + split_size]
elif isinstance(x, collections.Sequence):
elif isinstance(x, collections.abc.Sequence):
split_x = [None] * len(x)
for batch_idx in range(len(x)):
split_x[batch_idx] = x[batch_idx][t:t + split_size]
Expand All @@ -1726,7 +1726,7 @@ def tbptt_split_batch(self, batch, split_size):
if :paramref:`~pytorch_lightning.core.module.LightningModule.truncated_bptt_steps` > 0.
Each returned batch split is passed separately to :meth:`training_step`.
"""
time_dims = [len(x[0]) for x in batch if isinstance(x, (Tensor, collections.Sequence))]
time_dims = [len(x[0]) for x in batch if isinstance(x, (Tensor, collections.abc.Sequence))]
assert len(time_dims) >= 1, "Unable to determine batch time dimension"
assert all(x == time_dims[0] for x in time_dims), "Batch time dimension length is ambiguous"

Expand All @@ -1736,7 +1736,7 @@ def tbptt_split_batch(self, batch, split_size):
for i, x in enumerate(batch):
if isinstance(x, Tensor):
split_x = x[:, t : t + split_size]
elif isinstance(x, collections.Sequence):
elif isinstance(x, collections.abc.Sequence):
split_x = [None] * len(x)
for batch_idx in range(len(x)):
split_x[batch_idx] = x[batch_idx][t : t + split_size]
Expand Down