The latest machine learning models are being aggregated in the library How to use Detectron2
The latest machine learning often uses Pytorch’s wrapper library.
Models for Object Detection tasks are often used through pytorch wrappers such as mmdetection and detectron2.
For example, FaceBook Research’s model is announced through FaceBook Reseach’s easy-to-use detectron2.
The following models are models that can be used with detectron2.DensePose: Dense Human Pose Optimization In The Wild
Scale-Aware Trident Networks for Object Detection
TensorMask: A Foundation for Dense Object Segmentation
Mesh R-CNN
PointRend: Image Segmentation as Rendering
Momentum Contrast for Unsupervised Visual Representation Learning
DETR: End-to-End Object Detection with Transformers
Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation
D2Go (Detectron2Go), an end-to-end production system for training and deployment for mobile platforms.
Pointly-Supervised Instance Segmentation
Unbiased Teacher for Semi-Supervised Object Detection
Rethinking “Batch” in BatchNorm
Per-Pixel Classification is Not All You Need for Semantic Segmentation
The latest model is easy to use if you know how to use the wrapper library
To use the latest high-precision models, it is useful to familiarize yourself with the basic usage of these wrapper libraries.
Does it seem like it’s a pain to learn?
It’s easy
In the first place, the library itself is wrapped so that you can easily make inferences, so it is easy to use.
How to use (Detectron2)
Basically,
1. Pass the Config file and Weight of the model you want to use to the API to build the model
2. Inference
It is a procedure.
Great value to learn
It’s basically a library made for ease of use, so it’s useful to learn.
🐣
I’m a freelance engineer.
Work consultation
Please feel free to contact us with a brief development description.
rockyshikoku@gmail.com
I am making an app that uses Core ML and ARKit.
We send machine learning / AR related information.
