Blog
Blog
No items found.
Detectron2 is a computer vision model zoo of its own written in PyTorch by the FAIR Facebook AI Research group. Detectron2 includes all the models that were available in the original Detectron, such as Faster R-CNN, Mask R-CNN, RetinaNet, and DensePose as well as some newer models including Cascade R-CNN, Panoptic FPN, and TensorMask. You can use Detectron2 to do key point detection, object detection, and semantic segmentation. Detectron2 registers datasets in COCO JSON format.
Machine Learning Framework: The original detection was written in Caffe2 whereas Detectron2 has made a switch to PyTorch. This allows for developers to take a far more intuitive approach to test and change the model.
Modular Design: Detectron2's modular design allows for the maximum customizability possible; developers can write custom implementations for any part of the object detection model, allowing for greater flexibility within projects.
New Models: Detectron2 includes new models such as Cascade R-CNN, Panoptic FPN, and TensorMask.
Bringing Models to Production: Detectron2 now contains an extra layer called Detectron2go. This makes it easier to bring state-of-the-art models to production through features including model conversions for the cloud and mobile devices, network quantization, and standard training workflows.
Speed: Detectron2's entire training pipeline has been moved to GPUs, making Detectron2 much faster.
Training Detectron2: https://blog.roboflow.com/how-to-train-detectron2/
Detectron2
is licensed under a
Apache-2.0
license.
You can use Roboflow Inference to deploy a
Detectron2
API on your hardware. You can deploy the model on CPU (i.e. Raspberry Pi, AI PCs) and GPU devices (i.e. NVIDIA Jetson, NVIDIA T4).
Below are instructions on how to deploy your own model API.
