Experiment tracking for fastai-trained models.
- Log, organize, visualize, and compare ML experiments in a single place
- Monitor model training live
- Version and query production-ready models, and associated metadata (e.g. datasets)
- Collaborate with the team and across the organization
- Hyperparameters
- Losses and metrics
- Training code (Python scripts or Jupyter notebooks) and Git information
- Dataset version
- Model configuration, architecture, and weights
- Other metadata
Example dashboard with train-valid metrics and selected parameters
- Documentation
- Code example on GitHub
- Example dashboard in the Neptune app
- Run example in Google Colab
On the command line:
pip install neptune-fastai
In Python:
import neptune
# Start a run
run = neptune.init_run(
project="common/fastai-integration",
api_token=neptune.ANONYMOUS_API_TOKEN,
)
# Log a single training phase
learn = learner(...)
learn.fit(..., cbs = NeptuneCallback(run=run))
# Log all training phases of the learner
learn = cnn_learner(..., cbs=NeptuneCallback(run=run))
learn.fit(...)
learn.fit(...)
# Stop the run
run.stop()
If you got stuck or simply want to talk to us, here are your options:
- Check our FAQ page
- You can submit bug reports, feature requests, or contributions directly to the repository.
- Chat! When in the Neptune application click on the blue message icon in the bottom-right corner and send a message. A real person will talk to you ASAP (typically very ASAP),
- You can just shoot us an email at support@neptune.ai