Is resize needed for training a classification model?

Please provide the following information when requesting support.

• Hardware (T4/V100/Xavier/Nano/etc)
RTX3090 on ubuntu 20
• Network Type (Detectnet_v2/Faster_rcnn/Yolo_v4/LPRnet/Mask_rcnn/Classification/etc)
• TLT Version (Please run “tlt info --verbose” and share “docker_tag” here)
tollkit_version: 3.21.11
• Training spec file(If have, please share here)

arch: "resnet",
n_layers: 18
input_image_size: "3,224,224"
train_config {
enable_random_crop: True
enable_center_crop: True

• How to reproduce the issue ? (This is for errors. Please share the command line and the detailed log here.)
I have a dataset with 3 classes of images, the images are in varies of resolutions(300x300, 400x400, 80x100…).

Do I need to use offline tools to resize the images all to 224x224 before start training ?

OK, i already see it in doc:

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.