Installation
You can install OCLF either in a development setup using poetry or as a dependency in your own project via pip. As OCLF relies heavily on configuration files for running experiments it is easiest to start with a development setup as this allows configuration files to be inspected and edited in place for rapid prototyping. In contrast, if you want to keep your configurations separate from OCLF a installation of OCLF as a dependency might be the better way to go for you.
Development installation
Installing OCLF requires at least python3.8. Installation can be done using
poetry. After installing
poetry, check out the repo and setup a development environment:
git clone https://github.com/amazon-science/object-centric-learning-framework.git
cd object-centric-learning-framework
poetry install # Optionally add -E <extra> for each extra that should be installed
Valid extras are timm for access to timm models for feature extraction and
clip for access to OpenAI's clip model. For instance poetry install -E timm
-E clip installs both.
Poetry will create a separate virtual environment where the projects
dependencies are installed. It can be accessed using poetry shell or poetry
run. Please see the poetry docs for further
information on using poetry.
Installation as a dependency
It is also possible to install OCLF as a dependency for your project via pip for this simply run
this might take a while as pip tries to resolve the dependencies specified in
the OCLF pyproject.toml file whereas poetry directly uses locked dependencies
which were determined in a previous run and added to the repository.
With OCLF installed as a dependency you cannot directly explore configurations that are part of OCLF or edit them. Nevertheless, you can access OCLF components via the python API and run experiments using by adding your own configurations. For further information check out Usage/Separate codebase.