Deepkit

The collaborative and analytical training suite for insightful, fast, and reproducible modern machine learning. All in one cross-platform desktop app for you alone, corporate or open-source teams.

Track your experiments, debug your machine learning models, and manage your computation servers. It’s made for you as a single developer working completely offline and teams with real-time collaboration tools out of the box.

FEATURES

- Experiment execution on your workstation directly or in Docker
- Unified experiment definition using YAML
- Automatic versioning of your experiment: configs, files, outputs & more
- Analytical data of your experiment in real-time
- Hardware monitoring of CPUs, memory, GPUs, & more
- Tensorflow and Pytorch debugger
- Execute your experiments on any Linux server
- Issue tracker
- Notes

FEATURES EXPLAINED

- Execute experiments on your workstation in Docker, automatically provisioned.
- Automatically track every execution.
- Attach custom analytical data (metrics, files, images, logs, numpy arrays) to experiments using the free Python SDK.
- Tensorflow and Pytorch model debugger, for debugging the model graph + visualize the output of each layer including histograms of activations, weights, and biases.
- Connect any Linux machine via ssh credentials and execute your experiments on team with a simple click or CLI argument.
- Mange your project using the integrated issue tracker

DO IT IN REAL-TIME WITH FRIENDS

Create an account at deepkit.ai (in the app) to share your experiments in real-time with your friend and colleagues. You can switch between your local environment and the deepkit.ai server anytime directly in the app.