featup入坑笔记

一、新建环境

在conda中建立一个虚拟环境featup，

conda create -n featup python=3.9

二、开始配置：

 我是先下载了FeatUp，之后

pip install -e . -i https://mirrors.aliyun.com/pypi/simple/

 但是，突然出错了，说无法安装torch，我就自己动手安装的pytorch，速度嗷嗷快

然后再重新执行 

pip install -e . -i https://mirrors.aliyun.com/pypi/simple/

 哈哈哈，配置环境完成

三、运行程序

import torch
import torchvision.transforms as T
from PIL import Image

from featup.util import norm, unnorm
from featup.plotting import plot_feats, plot_lang_heatmaps



input_size = 224
image_path = "c:/1.jpg"
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
use_norm = True

transform = T.Compose([
    T.Resize(input_size),
    T.CenterCrop((input_size, input_size)),
    T.ToTensor(),
    norm
])

image_tensor = transform(Image.open(image_path).convert("RGB")).unsqueeze(0).to(device)



upsampler = torch.hub.load("mhamilton723/FeatUp", 'dinov2', use_norm=use_norm).to(device)
hr_feats = upsampler(image_tensor)
lr_feats = upsampler.model(image_tensor)
plot_feats(unnorm(image_tensor)[0], lr_feats[0], hr_feats[0])

注意这里image_path = "c:/1.jpg"

结果提示：

后来还发现缺很多库：
pip install ftfy

pip install regex

四、成功