C-RADIO-SO

• Shape-Optimized (431M parameters).
• Huge (653M parameters).

C-RADIOv4 was trained using an updated set of teach models:

• SigLIP2-g
• DINOv3-7B
• SAM3

This model is ready for commercial/non-commercial use.

License/Terms of Use

GOVERNING TERMS: Use of this model is governed by the NVIDIA Open Model License Agreement.

Deployment Geography

Global

Use Case

The embeddings generated by this model are expected to be used by a downstream application. For example:

• Image-level understanding (image classification, curation, etc.).
• Dense processing (semantic segmentation, depth estimation, etc.).
• Integration into a Vision-Language Model.

Release Date

Hugging Face: 01/27/2026 via RADIO Collection of Models.

References

• AM-RADIO: Agglomerative Vision Foundation Model -- Reduce All Domains Into One
• PHI-S: Distribution Balancing for Label-Free Multi-Teacher Distillation
• RADIOv2.5: Improved Baselines for Agglomerative Vision Foundation Models
• FeatSharp: Your Vision Model Features, Sharper
• C-RADIOv4 (Tech Report)

Model Architecture

Architecture Type: Neural Network Network Architecture: Vision Transformer Number of model parameters: -SO400M size: 431M, -H size: 653M

Input

Input Type(s): Image Input Format(s): Red, Green, Blue (RGB) Input Parameters: Two Dimensional (2D) Other Properties Related to Input: Image resolutions up to 2048x2028 in increments of 16 pixels

Output

Output Type(s): Embeddings Output Format: Tensor Output Parameters: Two Dimensional 2D Other Properties Related to Output: Downstream model required to leverage image features. Our AI models are designed and/or optimized to run on NVIDIA GPU-accelerated systems. By leveraging NVIDIA’s hardware (e.g. GPU cores) and software frameworks (e.g., CUDA libraries), the model achieves faster training and inference times compared to CPU-only solutions.

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.

import torch
from PIL import Image
from transformers import AutoModel, CLIPImageProcessor

hf_repo = "nvidia/C-RADIOv4-SO400M"

image_processor = CLIPImageProcessor.from_pretrained(hf_repo)
model = AutoModel.from_pretrained(hf_repo, trust_remote_code=True)
model.eval().cuda()

image = Image.open('./assets/radio.png').convert('RGB')
pixel_values = image_processor(images=image, return_tensors='pt', do_resize=True).pixel_values
pixel_values = pixel_values.cuda()

summary, features = model(pixel_values)

Spatial features have shape (B,T,D) with T being the flattened spatial tokens, and D being the channels for spatial features. Note that C!=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, the patch size is 16.

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.

Software Integration

Runtime Engine(s):

• [TAO-6.1]

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

The integration of foundation and fine-tuned models into AI systems requires additional testing using use-case-specific data to ensure safe and effective deployment. Following the V-model methodology, iterative testing and validation at both unit and system levels are essential to mitigate risks, meet technical and functional requirements, and ensure compliance with safety and ethical standards before deployment.

This AI model can be embedded as an Application Programming Interface (API) call into the software environment described above.

Model Version(s)

• C-RADIOv4-SO400M (400M parameters).
• C-RADIOv4-H (653M parameters).

Links:

• https://huggingface.co/nvidia/C-RADIOv4-SO400M
• https://huggingface.co/nvidia/C-RADIOv4-H

Training and Evaluation Datasets

Training Dataset

NV-CC-Img-Text-Dataset

Data Modality: Image Image Training Data Size: 1 Million to 1 Billion Images Data Collection Method by dataset: Automated Labeling Method by dataset: Not Applicable (no labels are needed) Properties: 700 Million Images

Evaluation Datasets

ImageNet

Link: ImageNet Data Collection: Automated Labeling Method: Human Training Images: 1,281,167 Validation Images: 50,000 Test Images: 100,000

To perform the semantic segmentation evaluation, we use training sets from ADE20K and PascalVOC to train a linear layer, and subseque