torchvision in R improvements

Background

torch is a popular package for machine learning, and specifically torchvision provides direct access to computer-vision models and datasets, but R support is missing some features compared to python.

Related work

keras/tensorflow in R interface with python, and so are even more difficult to install, compared to torch (uses C/C++ code).

Details of your coding project

• Implement all of the torchvision data sets: https://github.com/mlverse/torchvision/issues/104
• Implement torchvision models covering 5 new computer-vision tasks:
  • Object Detection model like FasterRCNN : https://github.com/mlverse/torchvision/issues/54#issuecomment-885201975 and SSD
  • Instance Segmentation model like Mask R-CNN
  • Keypoint Detection model like Keypoint R-CNN
  • Semantic segmentation model like FCN
  • Quantized models providing low footprint image embedding like Quantized ResNet

Expected impact

Torch and torchvision are extremely popular so this project could have a large impact.

Mentors

Contributors, please contact mentors below after completing at least one of the tests below.

• EVALUATING MENTOR: Christophe Regouby [email protected] is the maintainer of one of the packages in the torch ecosystem, and contributor to many R packages.
• Toby Hocking [email protected] is the author of numerous R packages, and has been mentor/admin for R-GSOC since 2013.

Tests

Contributors, please do one or more of the following tests before contacting the mentors above.

TODO MENTORS: write several tests that potential contributors can do to demonstrate their capabilities for this particular project. Ask some hard questions that will give you insight about how the contributors write code to solve problems. You'll see that the harder the questions that you ask, the easier it will be for you to choose between the contributors that apply for your project! Please modify the suggestions below to make them specific for your project.

• Easy: something that any useR should be able to do, e.g. download some existing package listed in the Related Work, and run it on some example data.
• Medium: something a bit more complicated. You can encourage contributors to write a script or some functions that show their R coding abilities.
• Hard: Can the contributor write a package with Rd files, tests, and vignettes? If your package interfaces with non-R code, can the contributor write in that other language?

Solutions of tests

Contributors, please post a link to your test results here.

• EXAMPLE CONTRIBUTOR 1 NAME, LINK TO GITHUB PROFILE, LINK TO TEST RESULTS.