diff --git a/CHANGELOG.md b/CHANGELOG.md index f2c9bcb726a79..1491fa26b9128 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -121,9 +121,6 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/). - Add support for optimizers and learning rate schedulers to `LightningCLI` ([#8093](https://github.com/PyTorchLightning/pytorch-lightning/pull/8093)) -- Add torchelastic check when sanitizing GPUs ([#8095](https://github.com/PyTorchLightning/pytorch-lightning/pull/8095)) - - - Added XLA Profiler ([#8014](https://github.com/PyTorchLightning/pytorch-lightning/pull/8014)) @@ -232,9 +229,6 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/). - `Trainer(resume_from_checkpoint=...)` now restores the model directly after `LightningModule.setup()`, which is before `LightningModule.configure_sharded_model()` ([#7652](https://github.com/PyTorchLightning/pytorch-lightning/pull/7652)) -- Added a mechanism to detect `deadlock` for `DDP` when only 1 process trigger an `Exception`. The mechanism will `kill the processes` when it happens ([#8167](https://github.com/PyTorchLightning/pytorch-lightning/pull/8167)) - - ### Deprecated @@ -308,9 +302,6 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/). ### Fixed - -- Fixed SWA to also work with `IterableDataset` ([#8172](https://github.com/PyTorchLightning/pytorch-lightning/pull/8172)) - - Fixed `lr_scheduler` checkpointed state by calling `update_lr_schedulers` before saving checkpoints ([#7877](https://github.com/PyTorchLightning/pytorch-lightning/pull/7877)) @@ -338,25 +329,33 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/). - Pass the `stage` argument of `Callback.{setup,teardown}` as a keyword ([#7973](https://github.com/PyTorchLightning/pytorch-lightning/pull/7973)) -- Fixed a DDP info message that was never shown ([#8111](https://github.com/PyTorchLightning/pytorch-lightning/pull/8111)) +- Fixed metrics generated during `validation sanity checking` are cleaned on end ([#8171](https://github.com/PyTorchLightning/pytorch-lightning/pull/8171)) -- Fixed metrics generated during `validation sanity checking` are cleaned on end ([#8171](https://github.com/PyTorchLightning/pytorch-lightning/pull/8171)) +- Fixed `log_gpu_memory` metrics not being added to `logging` when nothing else is logged ([#8174](https://github.com/PyTorchLightning/pytorch-lightning/pull/8174)) -- Fixed a bug where an infinite recursion would be triggered when using the `BaseFinetuning` callback on a model that contains a `ModuleDict` ([#8170](https://github.com/PyTorchLightning/pytorch-lightning/pull/8170)) +- Fixed a bug where calling `log` with a `Metric` instance would raise an error if it was a nested attribute of the model ([#8181](https://github.com/PyTorchLightning/pytorch-lightning/pull/8181)) -- Fixed NCCL error when selecting non-consecutive device ids ([#8165](https://github.com/PyTorchLightning/pytorch-lightning/pull/8165)) +- Fixed a bug where using `precision=64` would cause buffers with complex dtype to be cast to real ([#8208](https://github.com/PyTorchLightning/pytorch-lightning/pull/8208)) -- Fixed `log_gpu_memory` metrics not being added to `logging` when nothing else is logged ([#8174](https://github.com/PyTorchLightning/pytorch-lightning/pull/8174)) +## [1.3.8] - 2021-07-01 -- Fixed a bug where calling `log` with a `Metric` instance would raise an error if it was a nested attribute of the model ([#8181](https://github.com/PyTorchLightning/pytorch-lightning/pull/8181)) +### Fixed +- Fixed a sync deadlock when checkpointing a `LightningModule` that uses a torchmetrics 0.4 `Metric` ([#8218](https://github.com/PyTorchLightning/pytorch-lightning/pull/8218)) +- Fixed compatibility TorchMetrics v0.4 ([#8206](https://github.com/PyTorchLightning/pytorch-lightning/pull/8206)) +- Added torchelastic check when sanitizing GPUs ([#8095](https://github.com/PyTorchLightning/pytorch-lightning/pull/8095)) +- Fixed a DDP info message that was never shown ([#8111](https://github.com/PyTorchLightning/pytorch-lightning/pull/8111)) +- Fixed metrics deprecation message at module import level ([#8163](https://github.com/PyTorchLightning/pytorch-lightning/pull/8163)) +- Fixed a bug where an infinite recursion would be triggered when using the `BaseFinetuning` callback on a model that contains a `ModuleDict` ([#8170](https://github.com/PyTorchLightning/pytorch-lightning/pull/8170)) +- Added a mechanism to detect `deadlock` for `DDP` when only 1 process trigger an `Exception`. The mechanism will `kill the processes` when it happens ([#8167](https://github.com/PyTorchLightning/pytorch-lightning/pull/8167)) +- Fixed NCCL error when selecting non-consecutive device ids ([#8165](https://github.com/PyTorchLightning/pytorch-lightning/pull/8165)) +- Fixed SWA to also work with `IterableDataset` ([#8172](https://github.com/PyTorchLightning/pytorch-lightning/pull/8172)) -- Fixed a bug where using `precision=64` would cause buffers with complex dtype to be cast to real ([#8208](https://github.com/PyTorchLightning/pytorch-lightning/pull/8208)) - Fixed a bug where `truncated_bptt_steps` would throw an AttributeError when the target RNN has multiple hidden states ([#8145](https://github.com/PyTorchLightning/pytorch-lightning/pull/8145)) @@ -364,6 +363,8 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/). ## [1.3.7] - 2021-06-22 +### Fixed + - Fixed a bug where skipping an optimizer while using amp causes amp to trigger an assertion error ([#7975](https://github.com/PyTorchLightning/pytorch-lightning/pull/7975)) - Fixed deprecation messages not showing due to incorrect stacklevel ([#8002](https://github.com/PyTorchLightning/pytorch-lightning/pull/8002), [#8005](https://github.com/PyTorchLightning/pytorch-lightning/pull/8005)) - Fixed setting a `DistributedSampler` when using a distributed plugin in a custom accelerator ([#7814](https://github.com/PyTorchLightning/pytorch-lightning/pull/7814)) @@ -379,8 +380,6 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/). - Fixed `DataModule.prepare_data` could only be called on the global rank 0 process ([#7945](https://github.com/PyTorchLightning/pytorch-lightning/pull/7945)) - Fixed setting `worker_init_fn` to seed dataloaders correctly when using DDP ([#7942](https://github.com/PyTorchLightning/pytorch-lightning/pull/7942)) - Fixed `BaseFinetuning` callback to properly handle parent modules w/ parameters ([#7931](https://github.com/PyTorchLightning/pytorch-lightning/pull/7931)) - - - Fixes access to `callback_metrics` in ddp_spawn ([#7916](https://github.com/PyTorchLightning/pytorch-lightning/pull/7916)) diff --git a/tests/checkpointing/test_legacy_checkpoints.py b/tests/checkpointing/test_legacy_checkpoints.py index abd6abfc036f3..13ae7300375e8 100644 --- a/tests/checkpointing/test_legacy_checkpoints.py +++ b/tests/checkpointing/test_legacy_checkpoints.py @@ -69,6 +69,8 @@ "1.3.4", "1.3.5", "1.3.6", + "1.3.7", + "1.3.8", ] ) def test_resume_legacy_checkpoints(tmpdir, pl_version: str):