I try to train Yolo_v4 with resnet18 backbone, on Coco dataset. In augmentation config, it has 3 lines
output_width: 1248 output_height: 384 output_channel: 3
and it’s not clear from documentation how its working. Tha coco17 dataset have images, with folowing sizes 640x480, or 480x640, so I need to know if it makes rescale, crop, or what.
Also why you didnt use original kitti resolution?
There is no update from you for a period, assuming this is not an issue anymore.
Hence we are closing this topic. If need further support, please open a new one.
The output_width is the base output image width of augmentation pipeline.
The output_height is the base output image height of augmentation pipeline.
You can use different resolution of training images. The augmentation pipeline will do mosaic/jitter/resize/random-crop/etc.
You can set other output_width/output_height. But need to make sure it is multiple of 32.
- Input size: C * W * H (where C = 1 or 3, W >= 128, H >= 128, W, H are multiples of 32)
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.