karlinjfandJosh Karlin 41b6cbb3ca Don't copy the batch when training on a single gpu (#1576) ...

* fix

* whitespace

Co-authored-by: Josh Karlin <karlinjf@gmail.com>

2020-04-23 14:28:20 -04:00

..

__init__.py

Add ddp_cpu backend for testing ddp without GPUs (#1158)

2020-04-15 23:17:31 -04:00

auto_mix_precision.py

Set precision=16 when use_amp is passed as True (#1145)

2020-04-06 08:13:24 -04:00

callback_config.py

Default value for ModelCheckpoint filepath (#1548)

2020-04-23 11:50:58 -04:00

callback_hook.py

cleaning imports (#1032)

2020-03-12 12:41:37 -04:00

data_loading.py

Added Horovod distributed backend (#1529)

2020-04-22 17:39:08 -04:00

deprecated_api.py

fix deprecated default_save_path (#1449)

2020-04-10 14:32:56 -04:00

distrib_data_parallel.py

Added Horovod distributed backend (#1529)

2020-04-22 17:39:08 -04:00

distrib_parts.py

Added Horovod distributed backend (#1529)

2020-04-22 17:39:08 -04:00

evaluation_loop.py

Added Horovod distributed backend (#1529)

2020-04-22 17:39:08 -04:00

ignored_warnings.py

Dim 0 warning (#256)

2019-09-26 13:20:54 -04:00

logging.py

fixed memory leak from opt return (#1528)

2020-04-19 16:41:54 -04:00

lr_finder.py

Learning Rate finder (#1347)

2020-04-10 14:34:23 -04:00

model_hooks.py

Remove error when test dataloader used in test (#1495)

2020-04-15 22:16:40 -04:00

optimizers.py

Fixed configure optimizer from dict without "scheduler" key (#1443)

2020-04-10 11:43:06 -04:00

supporters.py

Fixed configure optimizer from dict without "scheduler" key (#1443)

2020-04-10 11:43:06 -04:00

trainer.py

fix boolean argparse (#1571)

2020-04-23 11:44:18 -04:00

training_io.py

Fix weights path (#1445)

2020-04-10 12:02:59 -04:00

training_loop.py

Don't copy the batch when training on a single gpu (#1576)

2020-04-23 14:28:20 -04:00

training_tricks.py

Removed redundant computations in clip_gradients that slowed down the gradient clipping. (#1523)

2020-04-18 23:07:15 -04:00