Vit L 14 CLIPA Datacomp1b

🚀 Model card for ViT-L-14-CLIPA-datacomp1B

This is a CLIPA - v2 model designed for zero - shot image classification, leveraging contrastive image - text techniques.

🚀 Quick Start

The model can be used for zero - shot image classification tasks. You can utilize it through the OpenCLIP library.

✨ Features

• Model Type: Contrastive Image - Text, Zero - Shot Image Classification.
• Original Source: https://github.com/UCSC-VLAA/CLIPA
• Dataset: mlfoundations/datacomp_1b
• Papers:
  • CLIPA - v2: Scaling CLIP Training with 81.1% Zero - shot ImageNet Accuracy within a $10,000 Budget; An Extra $4,000 Unlocks 81.8% Accuracy: https://arxiv.org/abs/2306.15658
  • An Inverse Scaling Law for CLIP Training: https://arxiv.org/abs/2305.07017

📦 Installation

No specific installation steps are provided in the original document.

💻 Usage Examples

Basic Usage

import torch
import torch.nn.functional as F
from urllib.request import urlopen
from PIL import Image
from open_clip import create_model_from_pretrained, get_tokenizer

model, preprocess = create_model_from_pretrained('hf - hub:ViT - L - 14 - CLIPA')
tokenizer = get_tokenizer('hf - hub:ViT - L - 14 - CLIPA')

image = Image.open(urlopen(
    'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))
image = preprocess(image).unsqueeze(0)

text = tokenizer(["a diagram", "a dog", "a cat", "a beignet"], context_length=model.context_length)

with torch.no_grad(), torch.cuda.amp.autocast():
    image_features = model.encode_image(image)
    text_features = model.encode_text(text)
    image_features = F.normalize(image_features, dim=-1)
    text_features = F.normalize(text_features, dim=-1)

    text_probs = (100.0 * image_features @ text_features.T).softmax(dim=-1)

print("Label probs:", text_probs)  # prints: [[0., 0., 0., 1.0]]

📚 Documentation

The model details and usage examples are provided above. You can refer to the original papers for more in - depth information:

• CLIPA - v2: Scaling CLIP Training with 81.1% Zero - shot ImageNet Accuracy within a $10,000 Budget; An Extra $4,000 Unlocks 81.8% Accuracy: https://arxiv.org/abs/2306.15658
• An Inverse Scaling Law for CLIP Training: https://arxiv.org/abs/2305.07017

📄 License

The model is licensed under the Apache - 2.0 license.

📚 Citation

@article{li2023clipav2,
      title={CLIPA - v2: Scaling CLIP Training with 81.1% Zero - shot ImageNet Accuracy within a $10,000 Budget; An Extra $4,000 Unlocks 81.8% Accuracy}, 
      author={Xianhang Li and Zeyu Wang and Cihang Xie},
      journal={arXiv preprint arXiv:2306.15658},
      year={2023},
}

@inproceedings{li2023clipa,
      title={An Inverse Scaling Law for CLIP Training}, 
      author={Xianhang Li and Zeyu Wang and Cihang Xie},
      booktitle={NeurIPS},
      year={2023},
}