Commit Graph

3687 Commits

Author SHA1 Message Date
Rohit Gupta 753cc4dfad
Fix default logging levels for train step specific hooks (#10756) 2021-11-29 19:51:17 +00:00
Carlos Mocholí 088818fbc6
Do not sanity check on reload (#10785) 2021-11-29 16:23:01 +01:00
Carlos Mocholí d3b7492bd0
[CLI] Add support for `--key.help=class` (#10767) 2021-11-29 14:12:53 +00:00
Adrian Wälchli 24fc54f07b
Fix typing in `pl.overrides.fairscale` (#10799)
* update typing in fairscale

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2021-11-29 11:28:28 +00:00
Adrian Wälchli bd3fb2e66e
Fix typing in `pl.core.mixins.hparams_mixin` (#10800)
* fix typing in hparams mixin

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* unused import

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2021-11-29 12:03:47 +01:00
Adrian Wälchli 97e52619ea
Fix typing in `pl.overrides.data_parallel` (#10796)
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2021-11-29 10:58:23 +01:00
Carlos Mocholí 724a92b065
Mark outputs as protected in the evaluation loops (#10781)
Co-authored-by: Rohit Gupta <rohitgr1998@gmail.com>
2021-11-28 20:09:30 +00:00
Adrian Wälchli c752060712
Consolidate state when retrieving sharded state dict in Lite (#10746)
Co-authored-by: thomas chaton <thomas@grid.ai>
2021-11-27 04:54:45 +00:00
Adrian Wälchli 038c151b6e
Improve typing for plugins (#10742)
Co-authored-by: Carlos Mocholi <carlossmocholi@gmail.com>
2021-11-26 20:14:58 +00:00
Adrian Wälchli 81a0a44d8f
Improve typing for Lite (#10743)
* improve typing in pytorch_lightning/lite

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* include lite again

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2021-11-26 20:14:11 +00:00
thomas chaton e94aff1c5b
Fault Tolerant: Add support for fault tolerant dataloader validator (#10465) 2021-11-26 19:33:47 +00:00
Carlos Mocholí 88930725dd
Add a custom `PossibleUserWarning` category (#10675) 2021-11-26 19:29:42 +00:00
Carlos Mocholí 3089dc3829
Improve typing for loops (#10749)
* Improve typing for loops

* Free memory
2021-11-26 18:39:09 +00:00
Carlos Mocholí 78face65e8
Improve typing for logging (#10748)
Co-authored-by: Justus Schock <12886177+justusschock@users.noreply.github.com>
2021-11-26 18:12:21 +00:00
Carlos Mocholí 31bb6e69ca
Avoid optional instances in Loops (#10735)
* Avoid optional instances in Loops

* More cleanup
2021-11-26 18:00:18 +00:00
Carlos Mocholí ae53562c97
Remove dead code in `TrainingEpochLoop` (#10750) 2021-11-26 17:49:00 +00:00
thomas chaton 6fe6e9e414
Delete TensorBoardLogger experiment before spawning the processes. (#10777) 2021-11-26 17:07:57 +00:00
thomas chaton 412d507a73
Fault Tolerant: move signal to SIGTERM (#10605) 2021-11-26 13:37:27 +00:00
Kaushik B e507bc9027
Fix compare version for packages (#10762) 2021-11-26 09:15:22 +00:00
thomas chaton 3d6262b7a9
Fault Tolerant Manual: Add support for DDP (#10638) 2021-11-25 18:31:53 +01:00
Kaushik B e0b4bb2ea3
Deprecate `DeviceType` in favor of `_AcceleratorType` (#10503)
Co-authored-by: Carlos Mocholí <carlossmocholi@gmail.com>
2021-11-25 16:41:03 +01:00
Danielle Pintz 85d7c4dce4
Configure mypy to install dependencies in CI and update pyproject.toml (#10682)
* mypy install deps

* fix deps

* add examples

* fix type errors

* fix type error

* fix

* fix

* update pyproject.toml
2021-11-25 00:19:30 +00:00
Carlos Mocholí f8b2d5b128
Improve error message on `TypeError` during `DataLoader` reconstruction (#10719) 2021-11-24 21:51:11 +00:00
thomas chaton 0066ff0129
Fault Tolerant Manual: Enable the feature (#10707) 2021-11-24 17:36:08 +00:00
Adrian Wälchli 30ec4815cb
Support re-instantiation for custom DataLoader in Lightning (#10680)
Co-authored-by: Justus Schock <12886177+justusschock@users.noreply.github.com>
2021-11-24 15:58:51 +01:00
thomas chaton e51a8ee7a3
Fault Tolerant Manual: utilities cleanup (#10703)
Co-authored-by: Kaushik B <45285388+kaushikb11@users.noreply.github.com>
Co-authored-by: Carlos Mocholi <carlossmocholi@gmail.com>
2021-11-24 15:01:55 +01:00
Rohit Gupta f36b395c4e
Update `LightningDataModule` docs (#10678) 2021-11-24 11:31:03 +00:00
Adrian Wälchli 89d0064b33
Use `PrecisionType` enum instead of checking raw values (#10704)
* use precision type

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2021-11-23 18:23:36 +01:00
thomas chaton b28ab34ff5
Fault Tolerant Manual: Add loading to reload the states (#10699)
Co-authored-by: Kaushik B <45285388+kaushikb11@users.noreply.github.com>
Co-authored-by: Carlos Mocholi <carlossmocholi@gmail.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2021-11-23 17:18:36 +00:00
Adrian Wälchli dca1776870
LiteDataLoader wrapper improvements (#10297) 2021-11-23 16:35:07 +01:00
thomas chaton 7cf6374bd0
Fault Tolerant Manual: Add support for collecting states across processes (#10639) 2021-11-23 14:27:33 +00:00
Adrian Wälchli ee9f7c0421
Update DeepSpeed precision handling after moving PrecisionPlugin (#10657) 2021-11-23 13:51:41 +00:00
thomas chaton 1702036c14
Fault Tolerant Manual: Add stateful dataloader iter (#10674) 2021-11-23 12:30:50 +00:00
Kaushik B 48cf1adfd3
Move Colab setup to ProgressBar (#10542)
Co-authored-by: Carlos Mocholí <carlossmocholi@gmail.com>
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
2021-11-23 06:16:31 +00:00
thomas chaton 2036dfb5df
Fault Tolerant Manual: Add _rotate_worker_indices utility (#10647) 2021-11-22 19:52:04 +00:00
Rohit Gupta 823bfa6f8a
Update `LightningModule` docs (#10637) 2021-11-23 01:02:04 +05:30
thomas chaton 6acfef680f
Fault Tolerant Manual: Add is_obj_stateful utility (#10646) 2021-11-22 18:48:32 +00:00
Adrian Wälchli 15305c459c
Update DDPShardedPlugin precision handling after moving PrecisionPlugin (#10658) 2021-11-22 17:45:34 +00:00
Andres Algaba 6fc7c54c3a
refactor slurm_job_id (#10622)
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: thomas chaton <thomas@grid.ai>
Co-authored-by: Adrian Wälchli <aedu.waelchli@gmail.com>
Co-authored-by: ananthsub <ananth.subramaniam@gmail.com>
2021-11-22 17:41:08 +00:00
Carlos Mocholí 338f3cf636
Use `Set` operations in `Environment.detect` (#10673) 2021-11-22 23:03:47 +05:30
Rohit Gupta d431ce14a1
Raise an error if batch_size cannot be inferred from current batch (#10541)
Co-authored-by: Carlos Mocholí <carlossmocholi@gmail.com>
2021-11-22 16:55:19 +00:00
Danielle Pintz 6810c40fc9
Small improvements to `_init_debugging_flags` (#10620) 2021-11-22 11:38:09 -05:00
Carlos Mocholí a6dedcf492
Fix `move_metrics_to_cpu` with evaluation (#10631) 2021-11-22 15:58:21 +00:00
thomas chaton 991cd895c6
1/n Add `FaultTolerantMode` (#10645) 2021-11-22 14:58:23 +00:00
Carlos Mocholí 48cb38ac5d
Fix docs filterwarnings snippet (#10671) 2021-11-22 14:52:21 +00:00
Kaushik B ce0a977742
Moved `env_vars_connector._defaults_from_env_vars` to `utilities.argsparse._defaults_from_env_vars` (#10501)
Co-authored-by: Carlos Mocholí <carlossmocholi@gmail.com>
2021-11-22 08:06:35 +00:00
Adrian Wälchli 8ea39d2c8f
LiteDataLoader code improvements and docs (#10625)
Co-authored-by: Carlos Mocholí <carlossmocholi@gmail.com>
2021-11-21 02:33:13 +01:00
puhuk af0bb96f0f
Remove the "_precision" suffix from some precision plugin files (#10052) 2021-11-19 17:37:39 +00:00
ananthsub a18b6409d1
Check torch.distributed availability before sharded tensor state dict hook registration (#10621)
Co-authored-by: Carlos Mocholí <carlossmocholi@gmail.com>
2021-11-19 17:34:23 +00:00
Adam Reeve 5fe0dac119
Fix misleading ModelCheckpoint documentation on every_n_epochs parameter (#10421) 2021-11-19 17:26:50 +00:00