The AgML Community is a public space for identifying and collaboratively addressing the data, evaluation, and infrastructure gaps limiting agricultural machine learning.
We welcome machine learning researchers, agronomists, plant scientists, data stewards, engineers, students, growers, extension professionals, and anyone working at the intersection of agriculture and AI.
You do not need to write code to contribute. Domain knowledge, datasets, metadata expertise, evaluation ideas, field use cases, and constructive questions are all valuable.
Visit Project-AgML Discussions to:
- Introduce yourself and share your work. Tell us what agricultural problem you are working on, which data modalities you use, what you can contribute, and where you need help.
- Discuss datasets, metadata, and standards. Help make agricultural data more discoverable, interoperable, and useful.
- Share models, benchmarks, and VLM experiments. Contribute results, evaluation protocols, and reproducible findings.
- Ask a question or propose an idea. Early-stage questions belong here; no polished proposal is required.
- Vote on community priorities. Help determine which gaps and experiments the community should address next.
- Discovering and standardizing public agricultural datasets
- Improving metadata, provenance, licensing, and interoperability
- Assessing dataset quality, diversity, leakage, and annotation reliability
- Benchmarking foundation models and vision-language models
- Expanding beyond classification to detection, segmentation, counting, and reasoning
- Connecting imagery with sensor, weather, soil, management, harvest, spatial, and temporal data
- Measuring geographic, temporal, and real-world generalization
- A community member raises a question, use case, or proposal in Discussions.
- The community clarifies the need, evidence, and possible approaches.
- When the work is actionable, maintainers summarize the discussion and open an issue or proposal in the appropriate AgML repository.
- Contributors develop datasets, metadata, documentation, benchmark results, or code through the relevant repository or Hugging Face collection.
- Outcomes are shared publicly and contributors are credited.
Discussions are for open questions, collaboration, and early proposals. Bugs and scoped implementation tasks belong in Issues in the relevant repository. Dataset- or model-specific changes may also be proposed through the corresponding Hugging Face repository.
Be curious, specific, constructive, and respectful. Explain relevant context, avoid overstating results, and distinguish tested findings from proposals or preliminary work.
Participation is governed by our Code of Conduct. See Contributing for practical guidance on choosing the right contribution pathway.
- AgML website
- AgML datasets and models on Hugging Face
- AgML model leaderboard
- AgML Python library
- Contributing leaderboard results
Open a discussion in Questions & Help. If you are unsure where an idea belongs, post it in Introductions & Projects and a maintainer will help route it.