VisionHOPE-T

Use it as a visual backbone for image classification, object detection, instance segmentation, or semantic segmentation.

Availability

Pretrained weights available for download

Pricing model

MIT license

Deployment

Linux with an NVIDIA GPU and CUDA toolkit

Privacy

Not established

Licence

MIT

• VisionHOPE-T is a visual backbone based on self-referential nested learning.
• The installation instructions require Linux, an NVIDIA GPU, and the CUDA toolkit.
• VisionHOPE-T pretrained checkpoints are downloadable from the Hugging Face repository. ^[2]
• VisionHOPE-T has an ImageNet-1K classification checkpoint. ^{[1] [2]}

  This repository contains the hierarchical VisionHOPE-T/S/B checkpoints for ImageNet-1K classification, COCO object detection and instance segmentation, and ADE20K semantic segmentation.

• VisionHOPE-T has COCO Mask R-CNN 1× and 3× checkpoints and an ADE20K UPerNet checkpoint. ^[2]

  This repository contains the hierarchical VisionHOPE-T/S/B checkpoints for ImageNet-1K classification, COCO object detection and instance segmentation, and ADE20K semantic segmentation.