nvidia

/

C-RADIOv3-g

like 6

Follow

NVIDIA 36.2k

use the code that present in the Model card, but encountered the error below

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

hf_repo = "nvidia/C-RADIOv3-g"

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)

It looks like it lack some code from another repo https://github.com/NVlabs/RADIO/tree/main/radio , but after i copyed it to the current folder, it will report another error.

vit_patch_generator.py line 302 
assert (src_patch_size ** 2) * 3 == chk_weight.shape[1], 'Unable to interpolate non-square patch size'