Ross Wightman d4386219c6 Improve type handling for arange & rel pos embeds, keep calculations in float32 until application (may change to apply in float32 in future). Prevent arange type hijacking by DeepSpeed Zero

2024-01-26 16:35:51 -08:00

..

data

Pass train-crop-mode to create_loader/transforms from train.py args

2024-01-24 16:19:02 -08:00

layers

Improve type handling for arange & rel pos embeds, keep calculations in float32 until application (may change to apply in float32 in future). Prevent arange type hijacking by DeepSpeed Zero

2024-01-26 16:35:51 -08:00

loss

…

models

Improve type handling for arange & rel pos embeds, keep calculations in float32 until application (may change to apply in float32 in future). Prevent arange type hijacking by DeepSpeed Zero

2024-01-26 16:35:51 -08:00

optim

…

scheduler

…

utils

…

__init__.py

…

version.py

Update version.py

2024-01-08 09:39:22 -08:00