model
model copied to clipboard
The Clay Foundation Model (in development)
Extended docs on sampling to describe v1 strategy.
The purpose of this PR is to create a notebook demonstrating how to run inference on new NAIP scenes. It may also serve the purpose of demonstrating sim search with...
174 Sentinel-2 scenes captured between 2015-2022, with approximately 1.5M pixel-level annotations (_cl), confidence levels (_conf) and reports (_rep). These derive from 3 field-expert annotators. https://marine-pollution.github.io/ License is open: Creative Commons...
Hi I understand the dataset can be streamed from S3, following the example in the docs I get an error, and assume access must be granted? ``` > aws s3...
- [ ] Basic documentation to train/inference - [ ] Model card - [ ] Data card. - [ ] GH release commit. - [ ] Deploy buttons working (the...
This needs another sanity check, and the actual band stats are in computation right now. Should update tomorrow morning with the stats from the calculations on batch.
Has anyone has assessed the spatial autocorrelation error of Clay vs standard models in downstream prediction/fine tuning tasks? Here's an example assessment vs Bayesian models: https://www.mdpi.com/1660-4601/18/13/6856 I'm considering generating embeddings...
Relates to https://github.com/orgs/Clay-foundation/discussions/291
We need to generate embeddings, like on v0.1, on all the training set. Since we are currently doing random cropping on the training set on loading, we also need to...
Make the model easy to use by converting it into a torchscript utility.