Add dataset card, paper link, and sample usage for Earth-Bench
#2
by
nielsr HF Staff - opened
Hi! I'm Niels from the community science team at Hugging Face. I've updated the dataset card for Earth-Bench to include:
- Relevant metadata such as the
image-text-to-texttask category. - Links to the original research paper, project website, and GitHub repository.
- A brief description of the dataset (248 tasks, 13,729 images across three modalities).
- Sample usage instructions for downloading the dataset using the CLI and Python API, as documented in the official repository.
- The dataset structure and citation information.
This will help researchers better find and utilize your benchmark on the Hugging Face Hub.
It appears that the result files generated by the ChangeOS tool are missing from the dataset. Since ChangeOS retrieves output paths via table lookups, the absence of these files prevents the Agent from proceeding to the subsequent steps in the task trajectory.