RADIO

🚀 AM-RADIO: Reduce All Domains Into One

This model performs visual feature extraction, generating image embeddings for downstream image classification tasks. It's for research and development only.

🚀 Quick Start

To pull the model from HuggingFace, you first need to log in:

huggingface-cli login

Then you can pull the model from a Python script:

from transformers import AutoModel
model = AutoModel.from_pretrained("nvidia/RADIO", trust_remote_code=True)

Alternatively, you can specify an access token:

access_token = "<YOUR ACCESS TOKEN"
model = AutoModel.from_pretrained("nvidia/RADIO", trust_remote_code=True, token=access_token)

✨ Features

• Performs visual feature extraction, generating image embeddings for downstream models to classify images.
• Flexible in input dimension, supporting inputs with width and height in the range $[14, 1008]$ as long as both axes are divisible by 14.

📦 Installation

No specific installation steps other than the HuggingFace model pulling process are provided in the original document.

💻 Usage Examples

Basic Usage

RADIO will return a tuple with two tensors. The summary is similar to the cls_token in ViT and is meant to represent the general concept of the entire image. It has shape $(B,C)$ with $B$ being the batch dimension, and $C$ being some number of channels. The spatial_features represent more localized content which should be suitable for dense tasks such as semantic segmentation, or for integration into an LLM. It has shape $(B,T,D)$ with $T$ being the flattened spatial tokens, and $D$ being the channels for spatial features. Note that $C \neq D$ in general.

Converting to a spatial tensor format can be done using the downsampling size of the model, combined with the input tensor shape. For 'radio_v1', the patch size is 14.

from einops import rearrange
spatial_features = rearrange(spatial_features, 'b (h w) d -> b d h w', h=x.shape[-2] // patch_size, w=x.shape[-1] // patch_size)

The resulting tensor will have shape $(B,D,H,W)$, as is typically seen with computer vision models.

Advanced Usage

We have trained this model to be flexible in input dimension. It supports inputs with both width and height in the range $[14, 1008]$ as long as both axes are divisible by 14. We have found that summarization tokens work best at $H=W=378$ (although the range $[192, 448]$ works well). For spatial tasks, we used $H=W=518$ to perform linear probing for semantic segmentation, and may perform better for more high - resolution tasks. Going up to $1008$, the model may need additional fine tuning at that resolution for best results. It is not required that $H=W$ although we have not specifically trained or testing the model in this setting.

📚 Documentation

Model Overview

Authors: Mike Ranzinger, Greg Heinrich, Jan Kautz, Pavlo Molchanov

This model performs visual feature extraction. For instance, RADIO generates image embeddings that can be used by a downstream model to classify images. This model is for research and development only. NVIDIA Research

References

• Paper
• PHI - S Paper
• BibTex
• GitHub examples

Model Architecture

Input

Output

Software Integration

Runtime Engine(s):

• TAO - 24.10

Supported Hardware Microarchitecture Compatibility:

• NVIDIA Ampere
• NVIDIA Blackwell
• NVIDIA Jetson
• NVIDIA Hopper
• NVIDIA Lovelace
• NVIDIA Pascal
• NVIDIA Turing
• NVIDIA Volta

[Preferred/Supported] Operating System(s):

• Linux
• Linux 4 Tegra
• QNX
• Windows

Pretrained Models

Refer to model_results.csv for model versions and their metrics. Link: https://huggingface.co/collections/nvidia/radio - 669f77f1dd6b153f007dd1c6

Training, Testing, and Evaluation Datasets

Training Dataset

Evaluation Dataset

Inference

Citing RADIO

If you find this repository useful, please consider giving a star and citation:

@InProceedings{Ranzinger_2024_CVPR,
    author    = {Ranzinger, Mike and Heinrich, Greg and Kautz, Jan and Molchanov, Pavlo},
    title     = {AM - RADIO: Agglomerative Vision Foundation Model Reduce All Domains Into One},
    booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
    month     = {June},
    year      = {2024},
    pages     = {12490 - 12500}
}

@misc{ranzinger2024phisdistributionbalancinglabelfree,
      title={PHI - S: Distribution Balancing for Label - Free Multi - Teacher Distillation}, 
      author={Mike Ranzinger and Jon Barker and Greg Heinrich and Pavlo Molchanov and Bryan Catanzaro and Andrew Tao},
      year={2024},
      eprint={2410.01680},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2410.01680}, 
}

Ethical Considerations (For NVIDIA Models Only)

NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. When downloaded or used in accordance with our terms of service, developers should work with their internal model team to ensure this model meets requirements for the relevant industry and use case and addresses unforeseen product misuse.

📄 License

RADIO code and weights are released under the NSCLv1 License.