See our collection for all versions of DINOv2.

Run DINOv2 with Keras 3: JAX, PyTorch, or TensorFlow

kerasformers/dinov2_vits14

Paper: DINOv2: Learning Robust Visual Features without Supervision (arXiv:2304.07193) · HF Papers

DINOv2 scales self-supervised ViT pretraining for strong transferable visual features without labels. These checkpoints are backbones that return patch tokens for downstream heads.

For more details on the model, please go to the upstream model card.

Pure-Keras 3 conversion of facebook/dinov2-small for kerasformers. One implementation runs unmodified on TensorFlow / Torch / JAX.

This is a self-supervised backbone (DinoV2Model), not a task head.

✨ Quick start

import os
os.environ["KERAS_BACKEND"] = "torch"  # or "jax" / "tensorflow"

import numpy as np
from PIL import Image
from kerasformers.models.dino_v2 import DinoV2Model

model = DinoV2Model.from_weights("kerasformers/dinov2_vits14", image_size=448)
image = Image.open("your_image.jpg").convert("RGB")
x = np.asarray(image.resize((448, 448)))[None].astype("float32")
tokens = model(x, training=False)
print(tokens.shape)

Load any DINOv2 variant the same way with from_weights("kerasformers/<variant>"):

Variant	Hub	Backbone
`dinov2_vits14`	`kerasformers/dinov2_vits14`	ViT-S/14
`dinov2_vitb14`	`kerasformers/dinov2_vitb14`	ViT-B/14
`dinov2_vitl14`	`kerasformers/dinov2_vitl14`	ViT-L/14

Tips

Set KERAS_BACKEND before importing Keras / kerasformers.
Feed raw [0, 255] pixels; normalization happens inside by default.
See DINOv2 docs and Loading Weights.
Community / upstream weights: DinoV2Model.from_weights("hf:facebook/dinov2-small").