README.md

---
license: apache-2.0
pipeline_tag: zero-shot-image-classification
library_name: openclip
---

# LongCLIP model

This repository contains the weights of the LongCLIP model.

Paper: https://huggingface.co/papers/2403.15378

Github repository: https://github.com/beichenzbc/long-clip

## Installation

```bash
git clone https://github.com/beichenzbc/Long-CLIP.git
cd Long-CLIP
```

## Usage

```
from model import longclip
import torch
from PIL import Image
from huggingface_hub import hf_hub_download

device = "cuda" if torch.cuda.is_available() else "cpu"
filepath = hf_hub_download(repo_id="BeichenZhang/LongCLIP-L-336px", filename="longclip-L@336px.pt")
model, preprocess = longclip.load(filepath, device=device)

text = longclip.tokenize(["A man is crossing the street with a red car parked nearby.", "A man is driving a car in an urban scene."]).to(device)
image = preprocess(Image.open("./img/demo.png")).unsqueeze(0).to(device)

with torch.no_grad():
    image_features = model.encode_image(image)
    text_features = model.encode_text(text)
    
    logits_per_image = image_features @ text_features.T
    probs = logits_per_image.softmax(dim=-1).cpu().numpy()

print("Label probs:", probs)
```