How to train your own dataset? What is the structure and form of the composition of the dataset？

THU-MIG / RepViT

RepViT: Revisiting Mobile CNN From ViT Perspective [CVPR 2024] and RepViT-SAM: Towards Real-Time Segmenting Anything

https://arxiv.org/abs/2307.09283

Apache License 2.0

730 stars 55 forks source link

Closed lucy11111111 closed 11 months ago

jameslahm commented 11 months ago

Thanks. You can arrange your dataset like ImageNet, and then set --data-path to your own dataset.

/path/to/imagenet-1k/
  train/
    class1/
      img1.jpeg
    class2/
      img2.jpeg
  val/
    class1/
      img3.jpeg
    class2/
      img4.jpeg

lucy11111111 commented 11 months ago

Okay, I got it. Thanks.