FathomNet
AI & ML interests
FathomNet is an open-source, expert annotated image database that can be used to train, test, and validate state-of-the-art artificial intelligence algorithms to help us understand our ocean and its inhabitants.
Welcome to the FathomNet Model Zoo!
The FathomNet Model Zoo (FMZ) is a collection of pretrained models that make use of FathomNet data, populated by members of the community and maintained by the FathomNet team. We encourage submitting any and all models that are trained on FathomNet data to the FMZ, in order to help the community share and build on each other's work.
How to contribute
In order to contribute to the FMZ, you must first request to join the FathomNet Model Zoo organization. To do this, create an account or log in and click the "Request to join" button at the top of this page. Once your request is approved, you will receive an email notification and be able to contribute models to the FMZ.
To contribute a model, click the "New" button at the top of the page and select "Model". Give your repository unique name and mark it as public. At a minimum, you should include your model weights, training metrics, and a model card.
Model card
The model card should detail the following information:
- Intended Use: What are the intended use cases of the model? What are the limitations that should be considered by people using the model?
- Factors: What are the primary factors that contribute to the model's performance? How was the model trained and evaluated?
- Metrics: What are the metrics used to evaluate the model's performance? How does the model perform on these metrics on relevant validation/test sets?
- Training and Evaluation Data: What data was used to train the model? How was the data collected and processed?
- Deployment: What are the steps required to deploy the model?
See the FathomNet/MBARI-midwater-supercategory-detector for an example of a model card. For more information on writing a model card, see the model card documentation.
For more information on how to upload these files, see the model upload documentation.