onnx-community

/

dinov3-vitl16-pretrain-lvd1689m-ONNX

like 2

Follow

ONNX Community 1.82k

# pip install torchao
import torch
from transformers import TorchAoConfig, AutoImageProcessor, AutoModel
from torchao.quantization import Int4WeightOnlyConfig
from transformers.image_utils import load_image


url = "http://images.cocodataset.org/val2017/000000039769.jpg"
image = load_image(url)

processor = AutoImageProcessor.from_pretrained("facebook/dinov3-vitsplus-pretrain-lvd1689m")

quant_type = Int4WeightOnlyConfig(group_size=128)
quantization_config = TorchAoConfig(quant_type=quant_type)

model = AutoModel.from_pretrained(
    "facebook/dinov3-vit7b16-pretrain-lvd1689m",
    dtype=torch.bfloat16,
    device_map="auto",
    quantization_config=quantization_config
)

inputs = processor(images=image, return_tensors="pt").to(model.device)
with torch.inference_mode():
    outputs = model(**inputs)

pooled_output = outputs.pooler_output
print("Pooled output shape:", pooled_output.shape)