A reference implementation of an AI lab.
TensorSpace is a reference implementation of an artificial intelligence lab.
I was tired of setting up ad hoc environments for various research experiments. I wanted a solution that can turn any “computer” into a research environment that I can use right away.
- Downloads and normalizes datasets. Currently only COCO, but more coming. Saves everything into a nice organized directory structure.
- Single annotation schema for all datasets. You don’t need to research with just one dataset at a time anymore. You do queries like “give me all images with bounding boxes from all datasets”.
- Automatically preprocess vectors or other intermidate datasets.
- Coming soon: Demos and models that use the data.
- Coming soon: GraphQL API for running models
- Coming soon: Multiple deployment targets. This will include Kubernetes.
pip install tensorspace
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size tensorspace-0.0.3.tar.gz (2.2 kB)||File type Source||Python version None||Upload date||Hashes View|