A new Docker image of the Keras image segmentation frameworks is available, now using TensorFlow 2.4.1 and Keras 2.4.3.
More information on the Docker image is available from Github:
github.com/waikato-datamining/tensorflow/tree/master/image-segmentation-keras
Docker images are now available for the TensorFlow 2 image classification, using the 2.9.1 release of TensorFlow:
For training the models, the make_image_classifier Python library is employed.
A tutorial on how to use these Docker images is available as well:
https://www.data-mining.co.nz/applied-deep-learning/image_classification/tf2_make_image_classifier/
A new resources for learning how to use our Docker images is now available (also available from the menu):
https://www.data-mining.co.nz/applied-deep-learning/
In these tutorials, you learn how to prepare the data, adapt configuration files, train a model and make predictions with a model.
At this stage, it covers the following domains:
image classification
image segmentation
instance segmentation
object detection
New Docker images are available for the MMDetection object detection framework, using the 2.24.1 release of MMDetection (code base as of 2022-05-01):
Docker images for building (and using) image classification models using the PyTorch-based framework MMClassification (version 0.23.1) are now available:
More information on the Docker images is available from Github:
Docker images for building (and using) image segmentation models using the PyTorch-based framework MMSegmentation (version 0.25.0) are now available:
More information on the Docker images is available from Github:
A Docker image for building image segmentation models using the PyTorch-based framework Segmentation Models (version 0.2.1) is available:
More information on the Docker image is available from Github:
github.com/waikato-datamining/pytorch/tree/master/segmentation_models/0.2.1
A new Docker image for building models with version 0.6 of Detectron2 is available:
More information on the Docker image is available from Github:
github.com/waikato-datamining/pytorch/tree/master/detectron2/0.6
Docker images are now available for Yolov5, a family of object detection architectures and models pretrained on the COCO dataset. Yolov5 performs very fast bounding box predictions.
The code used by the docker images is available from here:
github.com/waikato-datamining/pytorch/tree/master/yolov5
The tags for the images are as follows:
In-house registry: public.aml-repo.cms.waikato.ac.nz:443/pytorch/pytorch-yolov5:2022-01-21_cuda11.1
Docker hub: waikatodatamining/pytorch-yolov5:2022-01-21_cuda11.1