SanghyukChun/LongProLIP-ViT-B-16-S24M

## Official implementation of ViT-B/16 LongProLIP on ShareGPT4V

- This LongProLIP weight fine-tuned on ShareGPT4V 24M samples
- Pre-training dataset
    - ShareGPT4V / Seen samples 24M

### Overview
- LongProLIP Paper: https://arxiv.org/abs/2503.08048
- ProLIP Paper: https://arxiv.org/abs/2410.18857
- GitHub: https://github.com/naver-ai/prolip
- More models are available at https://huggingface.co/collections/SanghyukChun/prolip-6712595dfc87fd8597350291

### Performance overview (main results)
- Zero-shot ImageNet-1k top-1 accuracy: 70.76% (before fine-tuning: 74.6%)
- Zero-shot ImageNet distribution shifts: 59.08% (before fine-tuning: 63.0%)
- Zero-shot VTAB performance: 60.94% (before fine-tuning: 63.7%)
- Zero-shot retrieval performance: 61.31% (before fine-tuning: 59.6%)
- Average zero-shot performance on 38 tasks: 60.48% (before fine-tuning: 63.3%)

### Performance overview (additional results)
- Urban-1k: 87.2% (before fine-tuning: 65.4%)
- ECCV mAP@R: 35.7% (before fine-tuning: 34.1%)

```python
import requests
from PIL import Image

import torch
from prolip.model import ProLIPHF
from transformers import CLIPProcessor
from prolip.tokenizer import HFTokenizer

import warnings
warnings.simplefilter(action='ignore', category=FutureWarning)

processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch16")
model = ProLIPHF.from_pretrained("SanghyukChun/LongProLIP-ViT-B-16-S24M")
tokenizer = HFTokenizer("timm/ViT-B-16-SigLIP", context_length=64, clean="canonicalize")

url = "http://images.cocodataset.org/val2017/000000039769.jpg"
image = Image.open(requests.get(url, stream=True).raw)
inputs = processor(images=image, return_tensors="pt", padding=True)
texts = ["A couple of cats laying on top of a pink blanket.", "A man walks through a flooded road during a rainstorm", "photo"]
texts = tokenizer(texts)

outputs = model(image=inputs["pixel_values"], text=texts)

l2_logit = outputs["image_features"]["mean"] @ outputs["text_features"]["mean"].T
i_unc = torch.exp(outputs["image_features"]["std"]).sum(dim=-1)
t_unc = torch.exp(outputs["text_features"]["std"]).sum(dim=-1)
csd_logit = l2_logit - 0.5 * t_unc
csd_logit2 = l2_logit.T - 0.5 * i_unc
print("Mean-only image-to-text logits (by L2 distance):", l2_logit)
print("Uncertainty-aware image-to-text logits (by CSD):", csd_logit)
print("Uncertainty-aware text-to-image logits (by CSD):", csd_logit2.T)
print("Image uncertainty: ", i_unc)
print("Text uncertainty: ", t_unc)

@article{chun2025longprolip,
  title={LongProLIP: A Probabilistic Vision-Language Model with Long Context Text},
  author={Chun, Sanghyuk and Yun, Sangdoo},
  journal={arXiv preprint arXiv:2503.08048},
  year={2025}
}

@article{chun2024prolip,
  title={Probabilistic Language-Image Pre-Training},
  author={Chun, Sanghyuk and Kim, Wonjae and Park, Song and Yun, Sangdoo},
  journal={arXiv preprint arXiv:2410.18857},
  year={2024}
}