Bruno Sánchez-Andrade Nuño
Bruno Sánchez-Andrade Nuño
The current bottleneck of the Unet architecture are patch embeddings, for each section of the image. When we create the embedding of the image, we use [an average](https://github.com/Clay-foundation/model/blob/98a39b730a6aa0b145ad65fb0060df84083ad286/src/model_clay.py#L869) of all...
This PR changes the learning method (**we do not change the architecture or outputs**) from using the MAE (Masked Autoencoder) to the DINO (Distillation with No Labels) approach. Background on...
The main practical use case of Clay as of now, and the center of the upcoming app, is the ability to find similar features. Think: 1) Click on a pool,...
In the process of generating embeddings from a trained deep learning model, we perform inference on an input image and save the resulting embedding, which captures the semantic reconstructions of...
We need to release the full list of input images for the model v0.2. [Just like we have for v1](https://clay-foundation.github.io/model/specification.html#data-card). (Stanford transparency index suggests an explicit list of input files)....
We currently have model v0, are trainning v0.2 and working on v1 models. We should retain v0 documentation for those who will still use that model checkpoint, but add the...
To support the similarity search in our testbed of Bali, we are trying to improve the algorithm on #186 but I think it would help if we can take v0.2...
I am trying to run some more training loops for a specific region, using [this notebook](https://github.com/Clay-foundation/model/blob/region/docs/run_region.ipynb). I was not happy with the clustering: So I wanted to run a few...
Some of the inputs we use need a Planetary Computer token. on `datacube.py` We should ensure user goes to [PC developer portal](https://planetarycomputer.developer.azure-api.net/) and sets it with e.g. ```bash planetarycomputer configure...