Jee_King的博客

因为处理数据需求，要对tensor实现 cv2.warpAffine( )的功能原图： size = 360 × 480首先cv2.warpAffine( )，import cv2import numpy as npimg_path = r'C:\Users\CSDN\Desktop\00000001.jpg'image = cv2.imread(img_path)mapping2 = np.array([[ 1.23704604, 0., -256], [ 0.,

一. 批量生成事件（通过视频序列）打开终端。$: ssim$: roscd esim_ros运行脚本 gene_event.py 生成事件 .bag文件。import osroot_path="/media/zhaokai/Seagate\ Expansion\ Drive/OTB50/"fileList = os.listdir('/media/zhaokai/Seagate Expansion Drive/OTB50/')print(fileList)# os.system

#!/usr/bin/env pythonimport rosbagimport osfrom tqdm import tqdmfileList = os.listdir('./')bag_list = [n for n in fileList if 'bag' in n ]for bag_name in bag_list: bag=rosbag.Bag(bag_name) with open(bag_name[:-3]+'txt','w') as f: for

import cv2path=r'C:\Users\Qing\Desktop\eval_box_video\eval\box\video\seq_00.mp4'cap = cv2.VideoCapture(path)suc = cap.isOpened() frame_count = 0while suc: frame_count += 1 suc, frame = cap.read() params = [] params.append(2) cv2

apt-get updateapt-get upgradepip install -i https://mirrors.aliyun.com/pypi/simple/ opencv-pythonimport cv2 可能报一下错误：若报错：libgthread-2.0.so.0: cannot open shared object file则：apt-get install lib...

版本信息：python 3.6pytorch 0.4.1tensorflow-gpu 1.4.0tensorboard 1.6.0 [tensorflow 与 tensorboard版本匹配]from tensorboard_logger import Loggerlogger = Logger(logdir="./tensorboard_logs", flush_secs=10)...

脑子不好使，记录方便查阅...optimizer = optim.SGD(para_optim, lr=0.001, momentum=0.9, weight_decay=0.0002)def adjust_learning_rate(optimizer, epoch): lr = optimizer.param_groups[0]['lr'] * 0.1 #学习率每个epoch...

需求：打印梯度，检查网络学习情况net = your_network().cuda()def train(): ... outputs = net(inputs) loss = criterion(outputs, targets) loss.backward() for name, parms in net.named_parameters(): print('-...

转自: https://blog.csdn.net/github_36923418/article/details/84567849如何根据自己需求设定，可学习参数，并进行初始化。#比如cnn输出4个东西，你又不想concate到到一起，你想用权重加法，权重又不想自己设定，想让网络自己学#requires_grad=True这个很重要#设置前置网络及 可学习参数self.cnn=cnn_...

模式“RGB”转换为“YCbCr”的公式如下：Y= 0.257R+0.504G+0.098B+16Cb = -0.148R-0.291G+0.439B+128Cr = 0.439R-0.368G-0.071*B+128python 中的PIL Image 本身是带有rgb2ycbcr()函数的，但是其MATLAB实现上有差异。python 自带的函数实现：from PIL import...

iccd@iccd-937:~$ python>>> import torch>>> print(torch.__version__) (两个下划线) 

1.通用序列操作 （包括列表、元组、Unicode字符、buffer对象和xrange对象）索引 ：greeting = 'Hello'  greeting[0] greeting[-1] ； 'Hello'[1]分片：分片需要两个索引作为边界。第二个索引元素不包括在分片内。         number[7:10];  number[-3:];  number[:3];  number[:]  ...