Lack of toy data in visualize-embeddings.ipynb

[enemni]

The notebook https://github.com/Clay-foundation/model/blob/compile/docs/tutorials/visualize-embeddings.ipynb cannot be reproduced end-to-end as it reads from DATA_DIR = "/home/ubuntu/data" and there is no clear instruction on how to download the required toy dataset.

[santos-naudiyal]

try this
https://gofile.io/d/yMHabA
select the desired compiler in the installer

[srmsoumya]

Thanks for bringing this up, @enemni. We do need to provide some sample data for users to test the tutorial notebooks. I'll add that to my to-do list.

In the meantime, you can use the stachip library, which we use to create the training dataset, to pull some sample data. Here's a helpful tutorial to guide you through the process.

cc @yellowcap

[enemni]

On a similar note: I was working on reproducing the code from regression notebook and noticed that there is no mechanism in place to download the data. Cloning the dataset from HF is returning errors and there are already issues open here and here. Probably this is on nascetti-a/BioMassters but it is currently blocking the reproducibility of the notebook.