|
|
--- |
|
|
library_name: pytorch |
|
|
license: mit |
|
|
tags: |
|
|
- clip |
|
|
- vision-language |
|
|
- negation |
|
|
- checkpoint |
|
|
datasets: |
|
|
- coco |
|
|
pipeline_tag: feature-extraction |
|
|
model_name: NegationCLIP (CLIP fine-tuned for negation) |
|
|
--- |
|
|
|
|
|
# NegationCLIP |
|
|
- This repository hosts the **model checkpoint** only. |
|
|
- Usage example (PyTorch): |
|
|
|
|
|
```python |
|
|
|
|
|
import torch |
|
|
import clip |
|
|
|
|
|
device = "cuda" if torch.cuda.is_available() else "cpu" |
|
|
model, preprocess = clip.load("ViT-B/32", device=device) |
|
|
model.load_state_dict(torch.load("negationclip_ViT-B32.pth", map_location=device)) |
|
|
|
