---
pipeline_tag: zero-shot-image-classification
---

This repository contains the checkpoints of the paper [Self-Adapting Large Visual-Language Models to Edge Devices across Visual Modalities](https://huggingface.co/papers/2403.04908).

To download one checkpoint, one can do:

```python
!pip install huggingface_hub
from huggingface_hub import hf_hub_download

filepath = hf_hub_download(repo_id="ramfais/edgevl_weights", filename="datt_eurosat/wandb/latest-run/files/src/best_model.pth")
```