experiment tracking, checkpoints and GPUs
Train on your cloud. Lose a machine, not the run.
Resume from the last checkpoint, fork from any other: weights, optimizer and data come back. GPUs run in your cloud only while there is work. Free: you pay your cloud, not us.
train.py
run resumes from step 48,000
pip install "ravex[moonclip]"import ravex
@ravex.train_loop()
def train():
... # unchangedpython train.pythe console
Every run, in one place.

resume and fork
Any checkpoint is a new beginning.
Checkpoints as often as every step, only what changed. Run the script again and it carries on from the newest; fork any of them with new hyperparameters.
your cloud
Your GPUs. Your bill. Your bucket.
Your providers, on your own accounts, the cheapest offer first; a machine is given back, by default, about 5 minutes after its last job. Checkpoints in your own bucket, or in one we keep for you until you connect yours.
open source
The parts that touch your training are open
Runtime, checkpoint engine, node agent: Apache-2.0, at github.com/gpuzero-dev. Read them, run them without us, see exactly what runs beside your model.
Measured on a 50M-parameter model on a 4-core CPU machine, a Moonclip save holds the training loop for 226 ms where torch.save holds it for 2.7 s (its README, with the script to measure your own).
PyTorch runtime
ravex
One decorator, and your training loop saves, resumes and forks.
Checkpoint engine
moonclip
Saves only what changed, compressed, straight to S3 or R2: fast enough to save often.
On the GPU machine
smith
Runs only the scripts you gave it, never commands sent to it. Needs nothing but Python.