Hi, I'm one of the creators of jovian.ml . I transitioned from software development into machine learning, and realized that the data science & ML ecosystem needs better tools. It's just so hard to keep track of all the experiments you're working on, manage dozens of python environment & Jupyter notebooks for every project, and replicate or build upon prior work.
That's why we created Jovian, a platform built to fit in with a data scientist or ML engineer's workflow. It's a tiny python library that you include within your Jupyter notebook, and with just a couple of lines of code, it captures all the important information related to your ML experiments:
- Jupyter notebook (inputs & outputs)
- Python environment ( libraries with versions)
- Datasets, hyperparameters & metrics
- Model checkpoints, weights etc.
We want it to make it seamless & powerful, yet easy. Try out it at jovian.ml or read the docs at https://jovian-py.readthedocs.io