TF-Vision Model Garden

Table of Contents

Introduction

Image Classification

ResNet models trained with vanilla settings

ResNet-RS models trained with various settings

Vision Transformer (ViT)

Object Detection and Instance Segmentation

Common Settings and Notes

COCO Object Detection Baselines

RetinaNet (ImageNet pretrained)

RetinaNet (Trained from scratch)

Mobile-size RetinaNet (Trained from scratch):

Instance Segmentation Baselines

Mask R-CNN (Trained from scratch)

Cascade RCNN-RS (Trained from scratch)

Semantic Segmentation

PASCAL-VOC

CITYSCAPES

Video Classification

Common Settings and Notes

Kinetics-400 Action Recognition Baselines

Kinetics-600 Action Recognition Baselines