Open dxoigmn opened 1 year ago
For some reason, setting trainer.devices=2 does not also set datamodule.world_size to 2 even though this line exists: https://github.com/IntelLabs/MART/blob/eced15bdad18b6683190997590e2500a332b03e7/mart/configs/datamodule/default.yaml#L11
trainer.devices=2
datamodule.world_size
I think this is because many of the experiments modify datamodule.world_size: https://github.com/IntelLabs/MART/blob/eced15bdad18b6683190997590e2500a332b03e7/mart/configs/experiment/CIFAR10_CNN.yaml#L27 https://github.com/IntelLabs/MART/blob/eced15bdad18b6683190997590e2500a332b03e7/mart/configs/experiment/COCO_TorchvisionRetinaNet.yaml#L27 https://github.com/IntelLabs/MART/blob/eced15bdad18b6683190997590e2500a332b03e7/mart/configs/experiment/COCO_TorchvisionFasterRCNN.yaml#L28 https://github.com/IntelLabs/MART/blob/eced15bdad18b6683190997590e2500a332b03e7/mart/configs/experiment/ImageNet_Timm.yaml#L25
For some reason, setting
trainer.devices=2
does not also setdatamodule.world_size
to 2 even though this line exists: https://github.com/IntelLabs/MART/blob/eced15bdad18b6683190997590e2500a332b03e7/mart/configs/datamodule/default.yaml#L11I think this is because many of the experiments modify
datamodule.world_size
: https://github.com/IntelLabs/MART/blob/eced15bdad18b6683190997590e2500a332b03e7/mart/configs/experiment/CIFAR10_CNN.yaml#L27 https://github.com/IntelLabs/MART/blob/eced15bdad18b6683190997590e2500a332b03e7/mart/configs/experiment/COCO_TorchvisionRetinaNet.yaml#L27 https://github.com/IntelLabs/MART/blob/eced15bdad18b6683190997590e2500a332b03e7/mart/configs/experiment/COCO_TorchvisionFasterRCNN.yaml#L28 https://github.com/IntelLabs/MART/blob/eced15bdad18b6683190997590e2500a332b03e7/mart/configs/experiment/ImageNet_Timm.yaml#L25