Taeksang Kim 7f29a46d44 Add gradient accumulation option to train.py ...

option: iters-to-accum(iterations to accmulate)

Gradient accumulation improves training performance(samples/s).
It can reduce the number of parameter sharing between each node.
This option can be helpful when network is bottleneck.

Signed-off-by: Taeksang Kim <voidbag@puzzle-ai.com>

2023-02-06 09:24:48 +09:00

..

data

Cleanup resolve data config fns, add 'model' variant that takes model as first arg, make 'args' arg optional in original fn

2023-01-20 14:47:55 -08:00

layers

Merge remote-tracking branch 'origin/main' into levit_efficientformer_redux

2023-02-03 14:47:01 -08:00

loss

Remove inplace operators when calculating the loss

2022-10-17 10:41:21 +02:00

models

Update head metadata for effformerv2

2023-02-04 23:11:51 -08:00

optim

Add Adan to __init__.py

2022-12-15 11:37:29 +08:00

scheduler

Scheduler update, add v2 factory method, support scheduling on updates instead of just epochs. Add LR to summary csv. Add lr_base scaling calculations to train script. Fix #1168

2022-10-07 10:43:04 -07:00

utils

Add gradient accumulation option to train.py

2023-02-06 09:24:48 +09:00

__init__.py

Major module / path restructure, timm.models.layers -> timm.layers, add _ prefix to all non model modules in timm.models

2022-12-06 15:00:06 -08:00

version.py

Update version.py

2023-02-05 10:06:15 -08:00